Determination of Oil and Gas Reserves Petroleum Society Monograph No.1
THE PETROLEUM SOCIETY OF THE CANADIAN INSTITUTE OF MINING, METALLURGY AND PETROLEUM
Determination of Oil and Gas Reserves Petroleum Society Monograph No.1
© 1994 by The Petroleum Society of the Canadian Institute of Mining, Metallurgy and Petroleum, Calgary Section. All rights reserved. First edition published 1994. Printed in Canada. 10 9 8 7 6 5 4 3 2 Permission is granted for individuals to make single copies for their personal use in research, study, or teaching and to use figures, tables and short quotes from this monograph for republication in scientific books and journals. There is no charge for any of these uses. The publisher requests that the source be cited appropriately.
Canadian Cataloguing in Publication Data Main entry under title: Determination of oil and gas reserves.
(Petroleum Society monograph; no. I) Includes bibliographical references and index. ISBN 0-9697990-0-4 I. Petroleum reserves. I. Petroleum Society of CIM. II. Series. TN871.D47 1994 622'.1828 C94-910092-7
Edited by Virginia MacKay. Cover design by Guy Parsons. Typesetting and graphic design by lA. (Sandy) Irvine, By Design Services. Printed and bound in Canada by D.W. Friesen Ltd., Altona, ME.
CONTENTS
Figures
xiv
Tables
xvii
Foreword
xix
Preface
xxi
Acknowledgements Authors
xxiii .'
xxiv
PART ONE: DEFINITIONS AND GUIDELINES FOR CLASSIFICATION OF OIL AND GAS RESERVES 1.
OVERVIEW OF PART ONE
3
2.
DEFINITIONS 2.1 Introduction 2.2 Resources 2.2.1 Discovered Resources or Initial Volumes in Place 2.2.2 Undiscovered Resources or Future Initial Volumes in Place 2.3 Remaining Reserves 2.3.1 Remaining Proved Reserves 2.3.2 Probable Reserves 2.3.3 Possible Reserves 2.3.4 Development and Production Status 2.4 Cumulative Production 2.4.1 Sales 2.4.2 Inventory 2.5 Reserves Ownership 2.6 Specified Economic Conditions 2.7 Reporting of Reserves Estimates 2.7.1 Risk-Weighting of Reserves Estimates 2.7.2 Aggregation of Reserves Estimates 2.7.3 Barrels of Oil Equivalent
4 4 4 5 5 5 5 5 5 6 7 7 7 7 8 8 8 8 9
3.
GUIDELINES FOR ESTIMATION OF OIL AND GAS RESERVES 3.1 Introduction 3.2 Methods ofCaiculating Reserves 3.2.1 Deterministic Procedure 3.2.2 Probabilistic Procedure 3.3 Guidelines for Specific Methods 3.3.1 Volumetric Method 3.3.2 Material Balance Method 3.3.3 Decline Curve Analysis 3.3.4 Reservoir Simulation Method 3.3.5 Reserves from Improved Recovery Projects 3.3.6 Related Products
10 10 10 10 II 12 12 17 18 22 22 22 v
PART TWO: DETERMINATION OF IN-PLACE RESOURCES 4.
5.
vi
OVERVIEWOF PART TWO 4.1 Introduction 4.2 Resource Estimates 4.2.1 Volumetric Estimates 4.2.2 Material Balance Estimates 4.3 Procedures for EstimatingIn-Place Resources 4.4 Sources and Reliability of Data 4.5 Interrelationship of Parameters 4.6 Uses of Resource Estimates 4.7 Backgroundand Experience of Evaluators ESTIMATION OF VOLUMES OF HYDROCARBONS IN PLACE 5.1 Reservoir Area and Volume 5.1.1 Introduction 5.1.2 Acquisition of Data 5.1.3 Data Analysis 5.1.4 Mapping 5.1.5 Refinementof Volumetric Estimates 5.2 Thickness 5.2.1 Introduction 5.2.2 Defining Net Pay 5.2.3 Data Acquisition Programs 5.2.4 Data Interpretation 5.2.5 Factors Affecting Data Quality 5.3 Permeability 5.3.1 Introduction 5.3.2 Permeabilityfrom Core 5.3.3 Relative Permeability Measurement 5.4 Porosity 5.4.1 Introduction 5.4.2 Sources and Acquisition of Data 5.4.3 Analysis of Data 5.4.4 Factors Affecting Data Quality 5.5 Hydrocarbon Saturation 5.5.1 Introduction 5.5.2 Saturation Determination From Core 5.5.3 Saturation Determination From Logs 5.5.4 Flow Test Procedures for Gas and Oil Saturation 5.5.5 Factors Affecting Data Quality 5.6 Testing and Sampling 5.6.1 Introduction 5.6.2 DrillstemTests 5.6.3 Production Tests 5.6.4 Sampling 5.7 Reservoir Temperature 5.7.1 Introduction 5.7.2 Data Sources 5.7.3 Data Analysis 5.7.4 Data Analysis on a Regional Basis
27 27 27
27 30 30 31 31 31 34 35 35 35 35 36 38 43 44 44 45 46 48 49 53 53 53 54 55 55 55 58 63 65 65 65 69 70 72
75 75 75 75 77 81 81 81 82 82
5.8
5.9
5.10
5.11
5.7.5 Data Quality Reservoir Pressure 5.8.1 Introduction 5.8.2 Data Sources 5.8.3 Data Analysis Gas Formation Volume Factor 5.9.1 Introduction 5.9.2 Ideal Gas Law 5.9.3 Gas Compressibility Factor 5.9.4 Sour Gas 5.9.5 Derivation of Gas Formation Volume Factor Oil Formation Volume Factor 5.10.1 Introduction 5.10.2 Data Sources 5.10.3 Data Acquisition 5.10.4 Data Analysis 5.10.5 Data Adjustment 5.10.6 Summary Quality and Reliabilityof Data and Results 5.11.1 Introduction 5.11.2 Permeabilityfrom Cores 5.11.3 Porosity from Cores 5.11.4 Saturations from Cores 5.11.5 Effective Porous Zone and Net Pay from Cores 5.11.6 Porosity from Well Logs 5.11.7 Water Saturations from Well Logs '" 5.11.8 Effective Porous Zone and Net Pay from Well Logs 5.11.9 Drillstem Tests 5.11.10 Production Tests 5.11.11 Reservoir Fluid Samples 5.11.12 Reservoir Temperature 5.11.13 Reservoir Pressure 5.11.14 GasCompressibilityFactor 5.11.15 Formation Volume Factor 5.11.16 Material Balance 5.11.17 Interrelationships
'"
'"
85 86 86 86 86 91 91 91 91 92 94 96 96 96 96 96 98 100 101 101 101 101 102 102 103 103 103 104 104 104 104 104 105 105 105 105
6.
PROBABILITYANALYSIS FOR ESTIMATES OF HYDROCARBONS IN PLACE 6.1 Introduction 6.2 Warren Method Theory 6.3 Application 6.4 Typical Situation: Conventional Gas
106 106 107 108 110
7.
MATERIAL BALANCE DETERMINATION OF HYDROCARBONS IN PLACE 120 7.1 Introduction 120 7.2 Underlying Assumptions 120 121 7.3 Explanation of Terms 7.4 General Material Balance Equation .......................•.............. 122 7.5 Special Cases of the Material Balance Equation 122 7.5.1 Undersaturated Oil Reservoirs 122 7.5.2 Saturated Oil Reservoirs 123 7.5.3 Gas Reservoirs 123 vii
7.6 7.7
7.8 7.9
Limitations of Material Balance Methods Supplemental Calculations 7.7.1 Gas Caps and Aquifers 7.7.2 Water Influx Measurements 7.7.3 Analytical Water Influx Models Multiple Unknown Material Balance Situations Computer Solutions
123 124 124 124 124 125 127
PART THREE: ESTIMATION OF RECOVERY FACTORS AND FORECASTING OF RECOVERABLE HYDROCARBONS 8.
OVERVIEW OF PART THREE 8.1 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 8.2 Purpose of Depletion Strategy 8.3 Techniques for Reserves and Production Forecasting
131 131 131 132
9.
NATURAL DEPLETION MECHANISMS FOR OIL RESERVOIRS 9.1 Introduction 9.1.1 Fluid Expansion 9.1.2 Solution Gas Drive 9.1.3 WaterDrive 9.1.4 Gas Cap Drive , 9.1.5 Compaction Drive 9.1.6 CombinationDrive 9.2 Forecasting of Recoverable Oil 9.2.1 Solution Gas Drive 9.2.2 Water Drive 9.2.3 Gas Cap Drive 9.2.4 CombinationDrive 9.3 Factors Affecting Oil Recovery 9.3.1 Production Rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.3.2 Oil Quality 9.3.3 Reservoir Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 9.3.4 Reservoir Geometry 9.3.5 Effects of Economic Limit
133 133 133 133 134 134 134 135 135 137 137 140 140 140 140 141 141 141 142
10. DEPLETION MECHANISMS FOR NATURAL GAS RESERVOIRS 10.1 Introduction 10.2 Characteristics of Natural Gas 10.3 Definition of Reservoir Types from Phase Diagrams 10.4 Gas Recovery 10.5 Gas Reserves 10.5.1 Nonassociated Gas Reserves Determination .. , 10.5.2 Solution Gas Reserves Determination 10.5.3 Associated Gas Reserves Determination 10.6 Pipeline Gas Reserves 10.7 Reserves of Related Products 10.7.1 Natural Gas Liquids 10.7.2 Sulphur 10.8 Gas Deliverability Forecasting 10.9 Well Spacing 10.10 Cycling of Gas Condensate Reservoirswith Dry Gas viii
145 145 145 146 147 148 148 150 150 150 151 151 151 151 152 152
10.11 Secondary Recovery of Gas 10.12 EnhancedGasRecovery
153 153
II. ENHANCED RECOVERY BY WATERFLOODING 11.1 Introduction 11.2 Displacement Process 11.2.1 Mobility Ratio 11.2.2 Interfacial Tension 11.2.3 Fractional Flow 11.3 Types of Waterfloods 11.4 Analysis Methods and When to Apply Them 11.4.1 Pool Discovery 11.4.2 Delineated Pool: Immature Depletion 11.4.3 Post-Injection Startup 11.4.4 Post-Watertlood Response 11.4.5 Mature Watertlood U.s Volumetric Analysis 11.5.1 Overview of Method 11.5.2 Parameters and Factors Affecting Analysis 11.5.3 Reliability of Results 11.6 Decline Performance Analysis 11.6.1 Overview of Method 11.6.2 Factors Affecting Analysis 11.6.3 Reliability of Results 11.7 Comparison to Analogous Pools 11.7.1 Overview of Method 11.7.2 Procedure and Factors Affecting Analysis 11.7.3 Reliability of Results 11.8 Analytical Performance Prediction 11.8.1 Overview of Methods 11.8.2 Reliability of Results 11.9 Numerical Simulation 11.9.1 Overview of Method 11.9.2 Parameters and Factors Affecting Analysis 11.9.3 Reliability of Results 11.10 Waterflooding Variations 11.10.1 Naturally Fractured Reservoirs 11.10.2 Polymer Flooding 11.10.3 Micellar Flooding 11.11 Statistical Watertlood Analysis Survey 11.11.1 Overview of Database 11.11.2 Discussion of Results
154 154 154 154 154 155 156 156 157 157 158 158 158 158 158 158 162 162 162 162 163 163 163 163 164 164 164 164 166 166 166 166 167 167 168 168 168 168 168
12. ENHANCED RECOVERY BY HYDROCARBON MISCIBLE FLOODING 12.1 Introduction 12.2 Types of Hydrocarbon Miscible Floods 12.2.1 Vertical Miscible Floods 12.2.2 Horizontal Miscible Floods 12.3 Methods of Achieving Miscibility 12.3.1 First-Contact Miscible Process 12.3.2 MUltiple-Contact Miscible Process 12.3.3 Vapourizing Multiple-Contact Miscibility
171 171 171 171 172 172 172 172 173 ix
12.4
12.5
12.6
x
Experimental Methods to Determine Miscibility 12.4.1 P-X Diagram 12.4.2 Multi-Contact Ternary Diagram 12.4,3 Slim Tube Test 12.4.4 Rising Bubble Apparatus Screening and Feasibility Studies 12.5.1 Volumetric Method 12.5.2 Break-Through Ratio Method 12.5.3 Geological Model 12.5.4 Simulation Studies 12.5.5 Estimation of Uncertainties 12.5.6 Determination of Solvent and Chase Gas Slug Size 12.5.7 Field Performance of Miscible Floods Classification of Miscible Hydrocarbon Reserves 12.6.1 Possible Reserves 12.6.2 Probable Reserves 12.6,3 Proved Reserves
173 173 174 174 174 174 175 177 177 177 178 178 179 179 179 180 180
13. ENHANCED RECOVERY BY IMMISCIBLE GAS INJECTION 13.1 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 13.2 Types of Floods 13,3 Performance Prediction 13.3.1 External Injection Schemes 13,3.2 Dispersed Gas Injection Schemes
183 183 183 184 185 185
14. ENHANCED RECOVERY BY THERMAL STIMULATION 14.1 Introduction 14.2 Cyclic Steam Stimulation 14.2.1 Process Variation 14.2.2 Field Examples 14.2.3 Recovery Mechanisms 14.2.4 Design Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 14.3 Steam Flooding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 14.3.1 Process Variation 14,3.2 Design Considerations 14.4 Causes of Failure for Cyclic Steam Stimulation and Steam Flood Processes 14.5 Forecasting Models 14.5.1 Marx and Langenheim Model 14.5.2 Myhill and Stegeimeier Model 14.5,3. Vogel Model 14.5.4 ButierModel 14.6 In Situ Combustion Processes 14.6.1 Recovery Mechanisms 14.6.2 Process Variations 14.6.3 Design Considerations 14.6.4 Causes of Failure , , 14.7 Electromagnetic Heating
187 187 187 187 188 188 188 189 189 189 190 191 191 193 194 194 194 195 195 195 196 196
15. ENHANCED RECOVERY BY CARBON DIOXIDE FLOODING 15.1 Introduction , 15.2 Process Review , 15.3 Recovery Mechanisms
200 200 200 201
15.4
15.5 15.6
Design Considerations 15.4.1 Phase Behaviour 15.4.2 Displacement Efficiency 15.4.3 Volumetric Sweep Efficiency 15.4.4 Slug Sizing Reserve Evaluation Field Applications
201 201 201 202 202 202 203
16. RESERVES ESTIMATION FOR HORIZONTAL WELLS 16.1 Introduction 16.2 Reserves Determination Techniques 16.2.1 Performance Projection 16.2.2 Volumetric Method 16.2.3 Role of Heterogeneities ; 16.2.4 Importance of Channelling in Reserves Performance 16.2.5 Recovery Factors 16.3 Determination of Reserves 16.3.1 Determination of Reserves Parameters 16.3.2 Key Elements 16.3.3 Steps Involved in Reserves Determinations
205 205 206 206 209 209 209 210 211 211 211 211
17. NUMERICAL SIMULATION 17.1 Introduction 17.2 Types of Reservoir Simulators 17.3 Mathematical Formulation 17.4 Anatomy of Reservoir Simulation 17.5 Data Requirements 17.5.1 ReservoirGeometry 17.5.2 Rock and Fluid Properties 17.5.3 ProductionandWellData 17.6 Reservoir Model Grid Design 17.7 Reservoir Model Initialization 17.8 Model Sensitivity Analysis 17.9 History Matching 17.10 Forecasting Reservoir Performance 17.11 Use and Misuse of Reservoir Simulation 17.12 Summary
214 214 214 215 216 216 216 216 216 217 218 218 219 219 220 220
18. DECLINE CURVE METHODS 18.1 Introduction 18.2 Source and Accuracy of Production Data 18.3 Terminology 18.4 Single-Well vs. Aggregated-WellMethods 18.5 Decline Curve Methods for a Single Well 18.5.1 Exponential Decline 18.5.2 Hyperbolic Decline 18.5.3 Harmonic Decline 18.5.4 Dimensionless Solutions and Type-Curve Matching 18.6 Decline Curve Methods for a Group of Wells 18.6.1 Statistical Method 18.6.2 Theoretical Methods 18.7 Summary
222 222 222 223 223 224 225 226 229 230 231 231 234 235 Xl
19. RECOVERY FACTOR STATISTICS , 19.1 Introduction 19.2 Data Source and Reliability 19.3 Conventional Crude Oil 19.3.1 Natural or Primary Drive Mechanisms 19.3.2 Oil Recovery Factor Distributions 19.3.3 Average Recovery Factors 19.3.4 Pool Size 19.3.5 Fluid Type: Light and Medium vs. Heavy 19.3.6 Lithology: Clastics vs. Carbonates 19.3.7 Geological Period 19.3.8 Geological Play 19.3.9 Recovery vs, Common Reservoir Parameters 19.4 Conventional Gas 19.5 Using Recovery Factor Statistics
"
237 237 237 238 238 239 240 240 241 242 243 '" . 243 247 247 249
PART FOUR: PRICES, ECONOMICS, AND MARKETS
xii
20. OVERVIEW OF PART FOUR
253
21. CASH FLOW ANALySIS 21.1 Introduction 21.2 Mineral Rights Ownership 21.3 Principal Sources and Uses of Cash 21.4 Royalties and Mineral Tax 21.5 Federal Corporate Income Tax 21.6 Financial Statements 21.7 Finance and Economic Considerations
254 254 254 255 257 261 263 264
22. UNCERTAINTY AND RISK IN RESERVES EVALUATION 22.1 Introduction 22.2 Concepts 22.2.1 Definition of Risk and Uncertainty 22.2.2 Describing Uncertainty 22.2.3 Areas of Uncertainty 22.2.4 Causes of Uncertainty 22.2.5 Magnitude of Uncertainty 22.2.6 Use of Uncertainty 22.3 Estimation of Uncertainty 22.3.1· Parameters to be Estimated 22.3.2 Empirical Classification 22.3.3 Quantifying Subjective Estimates 22.3.4 Quantitative Estimation 22.4 Methods of Analysis 22.4.1 Carrying Out a Stochastic Evaluation 22.4.2 Decision Matrices 22.4.3 Decision Trees 22.4.4 Probabilistic Simulation 22.5 Evaluation of Undeveloped Lands
266 266 266 266 266 266 268 271 271 273 273 273 274 274 275 275 276 277 277 278
23. THE REGULATORY ENVIRONMENT 23.1 Introduction 23.2 Resource Assessments 23.3 Mineral Ownership 23.4 Economic Development Policies 23.5 Conservation Controls 23.5.1 Field Development and Production Conservation 23.5.2 Consumer Demand Conservation 23.6 Development, Operating, and Environmental Regulations 23.7 Domestic Supply Assurance 23.8 Fiscal Policies 23.9 Business Regulations 23.10 International Policies
281 281 281 282 282 283 283 283 283 284 285 285 285
24. CRUDE OIL MARKETS 24.1 Introduction 24.2 Transportation Network 24.3 Major Markets 24.4 North American Pricing 24.5 Price Risk Management 24.5.1 Futures 24.5.2 Options 24.5.3 Swaps 24.6 Outlook and Challenges
287 287 288 290 291 294 294 295 295 295
25. NATURAL GAS MARKETS 25.1 Introduction 25.2 The Market Environment 25.2.1 Review of Pre-Deregulation Era 25.2.2 Review of Current Era 25.2.3 Preview of Future Era 25.3 Market Mechanisms and Market Forces 25.3.1 Market Types and Market Mechanisms 25.3.2 Market Demand Forces 25.3.3 Production Forecasting 25.4 The Role of Reserves 25.5 Conclusions
297 297 297 297 298 300 300 300 302 304 304 305
26. USES OF RESERVES EVALUATIONS 26.1 Introduction 26.2 Users of Reserves Volumes and Production Forecasts 26.2.1 Producers 26.2.2 Transporters 26.2.3 Governments 26.2.4 Gas Marketers 26.2.5 Other Users 26.3 Developing Values from Reserves Estimates 26.3.1 Profitability Indices 26.3.2 Incremental Economics 26.3.3 Acceleration Projects 26.4 Uses of the Values Derived from Reserves Estimates 26.4.1 Valuing Oil and Gas Companies
306 306 306 306 306 306 307 307 307 307 310 310 311 311
xiii
26.4.2 26.4.3 26.4.4 26.4.5 26.4.6 26.4.7 26.4.8 26.4.9 26.4.10
Sale of ResourceProperties Evaluation of UnexploredLands and ExplorationWells Lending and Borrowing Auditing Evaluations Securities Reporting Accounting Requirements EstablishingFinding and Replacement Costs Estimating Barrels of Oil Equivalent EstimatingNet-Back Calculations
312 313 314 314 315 316 317 318 320
Biographies ofAuthors
32i
Acronyms
329
Glossary
333
Bibliography
345
Author index
349
Subject index
353
FIGURES 2.1-1 2.1-2 2.5-1 3.3-1 3.3-2 3.3-3 3.3-4 3.3-5 3.3-6 3.3-7 3.3-8 3.3-9 3.3-10 3.3-11 5.1-1 5.1-2 5.1-3 5.1-4 5.2-1 5.2-2 5.2-3 5.2-4 5.2-5 5.4-1 5.4-2 5.4-3
xiv
Resources Reserves Reserves Ownership Single Well Oil Pool with Good Geological Control Conventional Gas Pool, Zero Limit of Net Pay Map Conventional Gas Pool, Zero Limit of Net Pay Map with Individual Well Assignments Conventional Gas Pool, Zero Limit of Net Pay Map with Area of Proved Reserves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. Conventional Gas Pool, Zero Limit of Net Pay Map with Area of Proved Plus ProbableReserves Material Balance (Gas Reservoir) Material Balance (Scattered Data) Material Balance (ReservoirDrive and Depletion Mechanism) Decline Curve, Proved Reserves ., Decline Curve, Cumulative Gas Production Decline Curve, Cumulative Oil Production Pressure-Depth Plot for Free Water Level Determination Cross Contouring Series of Related Maps (zero edge from seismic, computer-contoured) (ZYCOR Software) Examples of Mechanical and Interpretive Mapping Reservoir IntervalTerminology Air Permeabilityvs. Porosity Flow Chart for a Core Analysis Program Hydrocarbon Fluid ContactIdentification from Pressure Gradients Sand Unit Shape Diagram Porosity of Cubic-Packed Spheres Typical Core Analysis Report Porosity vs. Horizontal Permeability
4
6 7 13 14
15 15 16 18 19 19 20 21 21 38 40 41 42 44 46 47 49 51 55 59 60
5.4-4 5.4-5 5.4-6 5.4-7 5.4-8 5.4-9 5.5-1 5.5-2 5.5-3 5.5-4 5.5-5 5.5-6 5.6-1 5.6-2 5.7-1 5.7-2 5.7-3 5.7-4 5.8-1 5.8-2 5.8-3 5.8-4 5.9-1 5.10-1 6.3-1 6.4-1 6.4-2 6.4-3 6.4-4 6.4-5 7.7-1 7.7-2 9.1-1 9.1-2 9.1-3 9.2-1 9.3-1 9.3-2 9.3-3 9.3-4 10.2-1 10.2-2 10.3-1 10.3-2 10.5-1 10.5-2 10.8-1 10.8-2
Core Analysis Report: Analytical Summary Sheet Porosityfrom Formation Density Log Porosityfrom Sonic Log Neutron PorosityEquivalence Curves Porosityand Lithology Determination from Neutron-Density Log Impact of Clay on Log and Core Measurements Porosityvs. Formation Factor Formation Resistivity Index Air Brine Capillary Pressure Test Log Interpretation Flow Chart Dual Water Model Shaly Sand Interpretation Process DrillstemTest Tool (UnsetPosition) DrillstemTest Tool (Set Position) : Representative Homer Plots from Wellsin the Utah-Wyoming Thrust Belt Relief Map for Southern Alberta ContourPlot of Spreadfor BHTValues in Southern Alberta Examples of Temperature vs. Depth Plots from Two Areas in Southern Alberta Static Gradient Pressure vs. Time Homer Plot PorosityVolume Map Compressibility Factors for Natural Gases Comparison of Formation Volume Factor by Differential and Flash Liberation Estimation of Reef Volume Typical Situation: Gas Pool Map Conversion of Base Area to Average Pool Area Typical Situation: Gas-in-Place Distribution Typical Situation: Reserve Distribution Typical Situation: Discounted Net Profit Before Investment StraightLine Plot for Oil Zone and Gas Cap Case StraightLine Plot for Oil Zone and Water Influx Case SolutionGas DriveReservoir Comparison of Solution Gas Drive and Water Drive Reservoirs Gas Cap Drive Reservoir Recommended Methods for the Stages of Exploitation Relationship Between Production Rate and Reserves Relationship Between Well Spacing and Abandonment Pressure Optimum Well Spacing Effectsof FacilityConstraints on Economic Limit Classification of Gas Based on Source in Reservoir Occurrence of Oil and Gas Pressure-Temperature Phase Diagram of a Reservoir Fluid Phase Diagram of a Cap Gas and Oil Zone Fluid Plot ofP/Z vs. Cumulative Gas Production Effect of Water Drive on Pressure Decline Back Pressure Plot Gas Deliverability Plot
60 61 61 62 62 64 67 68 70 71 72 73 76 ; 76 83 83 83 84 87 87 88 89 93 96 110
III 113 116 118 119
126 127 133 134 135 135 141 143 143 143 145 146 147 147 150 150 152 152
xv
11.2-1 11.2-2 11.3-1 11.3-2 11.3-3 11.5-1 12.3-1 12.3-2 12.3-3 12.5-1 13.2-1 14.5-1 14.5-2 16.2-1 17.2-1 17.3-1 17.6-1 17.6-2 17.6-3 17.6-4 18.3-1 18.3-2 18.5-1 18.5-2 18.5-3 18.5-4 18.5-5 18.5-6 18.6-1 18.6-2 18.6-3 18.6-4 18.6-5 19.3-1 19.3-2 19.3-3 19.3-4 19.3-5 19.3-6 19.3-7 19.3-8 19.3-9 19.3-10 19.3-11 19.3-12 19.3-13 19.3-14 xvi
Effect of Oil Viscosity on Fractional Flow Curve, Strongly Water-Wet Rock Effect of Oil Viscosity on Fractional Flow Curve, Strongly Oil-Wet Rock Cross Section for Vertical Waterflood Plan View for Horizontal Waterflood Flood Patterns for Horizontal Flood Schemes Effect of Mobility Ratio on Oil Production for the Five-Spot Pattern Pseudo-Ternary Diagram Indicating First-Contact Miscibility Development of Multiple-Contact Miscibility Condensing Process . . . . . . Development of Multiple-Contact Miscibility Vapourizing Process Reserves Distribution Gas Injection Types of Analytical Gravity Drainage Models Thermal Efficiency of Steam Zone as a Function of the Dimensionless Time Parameter Schematic of Horizontal and Vertical Well Drainage Areas Schematic Diagram of Matrix-Fracture Connectivity Mass Balance on Reservoir Element 2D Areal Model 2D Vertical Model 2D Radial Model .......................................... 3D Model Reservoir Performance Chart Production Performance Chart Exponential Decline Chart Decline Curve Analysis Chart Relating Production Rate to Time Decline Curve Analysis Chart Relating Production Rate to Cumulative Production Hyperbolic Curve Overlay Production Performance Graphs Composite of Analytical and Empirical Type Curves Production Performance Graph Rate-Cumulative Production Graph Distribution of Well Rates, Pembina Cardium Pool Rate-Ratio-Cumulative Graph, Pembina Cardium POOl Production Performance Graphs, Pembina Cardium Pool Oil Pools Distribution of Primary Oil Recovery Factors Large Mature Oil Pools Light and Medium Oil Pools Heavy Oil Pools Clastic Oil Pools Carbonate Oil Pools Upper Cretaceous Oil Pools Lower Cretaceous Oil Pools Jurassic Oil Pools Triassic Oil Pools Permian Oil Pools Mississippian Oil Pools Upper Devonian Oil Pools
155 155 156 156 157 159 172 173 173 178 184 192 193 208 215 215 217 217 217 218 224 224 226 227 227 228 229 230 232 232 233 234 234 239 240 241 241 242 242 242 243 243 244 244 244 244 245
19.3-15 19.3-16(a) 19.3-16(b) 19.3-17(a) 19.3-17(b) 19.3-18(a) 19.3-18(b) 19.3-19 19.3-20 22.2-1 22.2-2 22.2-3 22.2-4 22.2-5 24.2-1 24.2-2 24.4-1 24.4-2 25.3-1 25.3-2 25.3-3
Middle Devonian Oil Pools Oil Recovery vs. Porosity Porosity Distribution Oil Recovery vs. Net Pay Net Pay Distribution Oil Recovery vs. Water Saturation Water Saturation Distribution Gas Pools (Producing) Large Gas Pools (Producing) Risk and Uncertainty Level of Uncertainty in Reserves Estimates during the Life of a Producing Property The Effect of Error and Bias on a Reserve Estimate Expectation Curves: Comparison of Results Expectation Curve: Reconciliation of Different Views of Hydrocarbon Volumes and Values Major Alberta Pipeline Systems Major Crude Oil Pipelines and Refining Areas NYMEX WTI Prices at Cushing Alberta Crude Oil Pricing, Chicago Market (July 1992) Commercial and Regulatory Mechanisms for Ex-Alberta Markets Gas Marketing Options Reserves Connection to Markets
245 247 247 248 248 248 248 249 249 267 269 270 271 272 288 289 293 293 301 302 303
TABLES 4.2-1 4.4-1 5.4-1 5.5-1 5.10-1 5.10-2 5.10-3 6.1-1 6.4-1 6.4-2 6.4-3 6.4-4 6.4-5 6.4-6 7.2-1 7.2-2 9.2-1 9.2-2 10.7-1 11.8-1 11.11-1
In-Place Volumes of Related Products Sources of Data Comparison of Techniques of Determining Porosity Wettability and Interfacial Tension Pressure Volume Relations Separator Tests of Reservoir Fluid Sample Differential Vapourization In-Place Volumetric Estimation Techniques Gas-in-Place Distribution for Most Likely Area of384 Hectares Gas-in-Place Distribution for Most Likely Area of 576 Hectares Gas-in-Place Distribution for Most Likely Area of 704 Hectares Gas-in-Place Distribution for Most Likely Area of 576 Hectares, Variable Temperature and Gas Deviation Factor Reserve Distribution for Most Likely Area of 576 Hectares Discounted Net Profit Before Investment Distribution for Most Likely Area of 576 Hectares ReservoirVoidage Terms Reservoir Expansion Terms Recommended Reserves Forecasting Methods Decline Analysis Plots Used after Water Break-through Recoveries of Related Products Classification of 33 Waterflood Prediction Methods Summary of Recovery Factors: A Sampling of Western Canadian Waterfloods
30 32 56 69 98 99 99 107 114 115 115 117 118 119 121 122 136 139 151 165 169 xvii
7
11.11-2 13.3-1 18.5-1 18.6-1 19.2-1 19.3-1 19.3-2 19.3-3 19.3-4 21.4-1 21.4-2 21.4-3 21.5-1 24.3-1 26.4-1
XV1l1
Reserve Analysis Technique Distribution . . . . . . . . . . . . . . . . . . . . . . . . .. Recommended Performance Prediction Methods DeclineCurve Equations Statistical Parameters for Pembina Cardium Pool Public Data Available for Reserve Studies Primary Oil Recovery by DriveMechanism AverageOil Recoveries Recovery Factors for Upper Devonian Zones Recovery Factors for Geological Plays in WesternCanada Summaryof AlbertaNatural Gas Royalty Changes Summaryof Equations for Basic Royalty Summaryof AlbertaCrude Oil RoyaltyRate Changes Cash Flow and Income Tax Summary Importers of Canadian HeavyCrude Conversion Rates
169 185 225 233 237 238 241 245 246 258 259 259 262 292 318
FOREWORD
The estimating and reporting of reserves of oil and gas and related substances are of fundamental importance to the oil and gas industry. Reserves estimates form the basis for most development and operational decisions and are of critical importance in financing and other commercial arrangements that allow oil and gas developments to proceed in an orderly and efficient manner. Reserves estimates also playa key part in relevant planning and policy decisions by governments and others. The role of reserves estimates in operational, financial and policy decisions emphasizes the need for the estimates to be as accurate and current as possible. The use ofconsistent terminology and estimation procedures is also essential. This monograph, Determination ofOil and Gas Reserves, has been developed to assist in achieving the objectives of accuracy and consistency in estimating reserves. The idea ofdeveloping such a monographwas conceivedby Dr. Roberto Aguilera who, as Chairman of the Reserves Monograph Advisory Committee,has co-ordinatedthe preparation of the document. The project was sponsored by the Calgary Section of the Petroleum Society ofthe Canadian Institute of Mining, Metallurgy and Petroleum. The first organizational meeting of the committee took place in the spring of 1990. Since that time, members of the committee, on their own and with the support of their employers, have contributed substantial time and expertise to the project and enlisted the help of many industry experts in the preparation and critique of specific chapters. The objective was to develop a reference that would be of substantial value to geologists, engineers and other technical persons involved in estimating reserves, as well as to others who use such estimatesfor particular purposes. With the publication ofthe monograph in the spring of 1994, the committee will have achieved that objective. A total of over fifty people have been involved in the planning, the writing and review of the chapters, the drafting of figures, and the editing and preparation of the final copy for the printing of the monograph. All those involved in estimatingoil and gas reserves, or who use such estimates, owe them a vote of thanks. I am confident that the monograph will become a standard reference for all practitioners of the science of estimating oil and gas reserves. It will also serve as an excellent training tool for persons who have only a basic understanding of reserves estimation methods and who wish to advance their knowledge of the subject. G. 1. DeSorcy, P.Eng. Calgary, January 1994
xix
7
PREFACE
The estimation of reserves of oil, gas, and related substances has been a hot topic since the very beginning of the oil industry. Over the ensuing years, the concept of reserves has meant different things to different people within this industry, with each evaluator, oil and gas company, financial agency, securities commission, and government department using its own version of the definitions. The monograph represents our effort to find definitions and guidelines for the classification of reserves that will be acceptable to all ofthese users. When the concept of this monograph was first discussed, we wrestled with the question: "Should we ask one or two professionals to prepare the whole monograph or should we ask a variety of specialists to contribute to it?" In the end we concluded that we would not find one or two people with expertise in all the topics concerned with oil and gas reserves, so we should use a number of knowledgeable authors. We ended up with forty contributing authors and a group of reviewers who helped to polish the thirty-seven topics covered in the twenty-six chapters ofthe monograph. The topics have fallen into four major divisions that we have called "parts" in the monograph. Part One presents the definitions and guidelines for the classification of oil and gas reserves. These have been prepared by the Standing Committee on Reserves Definitions of the Petroleum Society of the Canadian Institute of Mining, Metallurgy and Petroleum. Part Two discusses the volumetric and material balance methods for estimating volumes of oil and gas in place, various sources of data, and the interpretation of the data. Part Two also deals with probabilistic methods for estimating the volumes of oil and gas contained in reservoirs, in addition to the more common deterministic methods. Part Three considers the estimation of recovery factors for oil and gas reservoirs, with particular emphasis on volumes recoverable by enhanced recovery methods. Secondary and tertiary recovery methods are discussed, as well as primary methods and the use of horizontal wells. Part Three also addresses decline curve analysis and reservoir modelling by numerical simulation. Part Four covers the other factors that must be considered in estimating reserves: cash flow analysis, the assessment of uncertainty, the role of markets, and potential regulatory impacts that must be recognized by evaluators. Part Four ends with a discussion of the uses that are made of reserves estimates. This part proved to be very challenging to write as the diverse nature of the applications of recovery estimates in economic evaluations led to some animated discussions between the engineering and financial groups. But in the end, I think we put together some information that will be useful to all the professionals who deal with economic evaluations.
(cont'd)
xxi
z
For simplicity, the nomenclature and units of measurement are defined following each equation. We have used the metric system (SI), with Imperial units shown as well in some cases. Following the text, we have included brief biographies of the authors and several lists for the convenience of readers: Acronyms, Glossary, Bibliography, Author Index, and Subject Index. It is our sincere hope that this monograph, Determination of Oil and Gas Reserves, will help to simplify and standardize the science and art of estimating oil and gas reserves throughout the world.
Roberto Aguilera, P. Eng. Calgary, January 1994
xxii
ACKNOWLEDGEMENTS
Associated with the publication of the monograph was the time-consuming and challenging task of co-ordinating the material produced by forty authors with forty different backgrounds and forty different writing styles. The Reserves Monograph Advisory Committee did a superb job ofco-ordinating the four parts of the monograph. As Chairman, I wish to thank the members of the committee for the many hours they devoted to planning the work, meeting with the authors, and reviewing the drafts. The following are the members of the committee with their company affiliations. We are grateful to the employers for supporting the members in this endeavour. N. Guy Berndtsson Keith D. Brown CAS. (Charlie) Bulmer R.V. (Bob) Etcheverry John Hewitt R. V. (Bob) Lang W.V. (Bill) Mandolidis Michael E. McCormack r. Glenn Robinson Roberto Aguilera, Chairman
Energy Resources Conservation Board Royal Bank of Canada Sproule Associates Limited CN Exploration Inc. Martin Petroleum and Associates Energy Consultant Saskatchewan Oil and Gas Corp. Fractical Solutions Inc. Sproule Associates Limited Servipetrol Ltd.
The work on the monograph involved authors and reviewers with backgrounds in government regulations, banks, stock brokers, securities commissions, consultants, the University of Calgary, and major, mid- and small-sized exploration and production companies. On the following pages are listed the names and company affiliations of the authors of the various chapters and sections of the monograph. These are the people who supplied the "meat" of the document through many volunteer hours of labour-writing, revising, and consulting with others-on the material they were responsible for. In addition, we would like to thank the Petroleum Society ofCIM, Canadian Well Logging Society, Society of Petroleum Engineers, Society of Professional Well Log Analysts, American Association of Petroleum Geologists, and Alberta Energy Resources Conservation Board, as well as Western Atlas International Inc., Schlumberger, Gulf Publishing Co., PanCanadian Petroleum Ltd., Chevron Canada Resources, and PennWell Publishing Co. for permission to use material from their publications. We also express our gratitude to all of the various authors and organizations that have published material on reserves estimation and thereby added to the body of knowledge on this subject. Virginia MacKay, P.Eng., the professional editor for this monograph, undertook the daunting task of editing the material written by the forty different authors and assembling it all into one coherent document. She was assisted very conscientiously by lA. (Sandy) Irvine, P.Geol., who entered the text and figures on the computer. Together they prepared the camera-ready copy for the printer. Mike McCormack checked the nomenclature throughout the monograph and also contributed to the compilation ofthe Subject Index. Our sincere thanks to Virginia, Sandy, Mike, and all the authors, reviewers and co-ordinators for their dedication to the quality of the monograph. Roberto Aguilera, P. Eng. Calgary, January 1994 XXlll
7
AUTHORS
Part One Standing Committee on Reserves Definitions GJ. (Gerry) DeSorcy Energy Consultant
Chairman
George A. Warne Energy Consultant
Secretary
R. V. (Bob) Lang Energy Consultant
Co-ordinator
J. Glenn Robinson Sproule Associates Limited
Co-ordinator
Barry R. Ashton Ashton Jenkins and Associates Ltd. Graham R. Campbell National Energy Board David R. Collyer Shell Canada Limited John Drury Consultant (Ontario Securities Commission) W.O. (Bill) Robertson Price Waterhouse David W. Tutt Bank of Montreal
Note: All committee members contributed to the writing of Part One.
xxiv
AUTHORS (cant'd)
Part Two N. Guy Berndtsson Energy Resources Conservation Board
Co-ordinator
CAS. (Charlie) Bulmer Sproule Associates Limited
Co-ordinator
Brent Austin PanCanadian Petroleum Limited
Co-Author of Sections 5.2, 5.3,5.4,5.5
Robin G. Bertram Talisman Energy Inc.
Co-Author of Section 5.6 and Author of Sections 5.8, 5.9
Mike J. Brusset Brusset Consultants Ltd.
Co-Author of Section 5.6 and Author of Section 5.11
Merlin B. (Mel) Field Consultant
Author of Chapter 7
J.D. (Joe) Giegerich Chevron Canada Resources
Author of Sections 5.7, 5.10
DJ. (Dave) Hemphill Shell Canada Limited
Author of Section 5.1
Craig F. Lamb Lonach Consulting Ltd.
Co-Author of Sections 5.2, 5.3, 5.4, 5.5
Raymond A. Mireault Gulf Canada Resources Limited
Author of Chapter 6
xxv R
AUTHORS (cont'd)
Part Three
XXVI
R.V. (Bob) Etcheverry CN Exploration Inc.
Co-ordinator and Author of Sections 8.1, 8.2
John M. Hewitt Martin Petroleum & Associates
Co-ordinator and Author of Section 8.3
Soheil Asgarpour Gulf Canada Resources Limited
Author of Chapter 12
Anthony D. Au Servipetrol Ltd.
Author of Chapter 17
Keith M. Braaten Coles Gilbert Associates Ltd.
Co-Author of Chapter II
RonM. Fish Imperial Oil Limited, Resources Division
Author of Chapter 13
Mam Chand Gupta GM International Oil and Gas Consulting Corp
Author of Chapter 10
William E. Kerr Joss Energy
Co-Author of Chapter 15
Gobi Kular Advanced Petroleum Technologies
Co-Author of Chapter 14
Dana B. Laustsen Coles Gilbert Associates Ltd.
Co-Author of Chapter II
Margaret Nielsen Petro-Canada
Co-Author of Chapter 9
David C. Poon Consultant, D.C. Poon Consulting Inc.
Co-Author of Chapter 14
Ross A. Purvis Energy Resources Conservation Board
Author of Chapter 18
Darlene A. Sheldon Petro-Canada
Co-Author of Chapter 9
Phillip M. Sigmund BRTR Petroleum Consultants Ltd.
Co-Author of Chapter 15
Ashok K. Singhal Petroleum Recovery Institute
Author of Chapter 16
Andy Warren Energy Resources Conservation Board
Author of Chapter 19
AUTHORS (cont'd)
Part Four Keith D. Brown Royal Bank of Canada
Co-ordinator and Authorof Chapters20, 21
Janusz Bielecki National Energy Board
Authorof Chapter 24
Noel A. Cleland Sproule Associates Limited
Author of Chapter 26
David C. Elliott Geosgil Consulting
Author of Chapter 22
Harold R. Keushnig Energy Resources Conservation Board
Authorof Chapter 23
Tim J. Reimer Pan-Alberta Gas Ltd.
Author of Chapter 25
xxvii
PART ONE DEFINITIONS AND GUIDELINES FOR CLASSIFICATION OF OIL AND GAS RESERVES
7
Chapter 1
OVERVIEW OF PART ONE
There are almost as many definitions for reserves of oil and gas and related substances as there are evaluators, oil and gas companies, financial agencies, securities commissions, and government departments. Each uses its own version of the definitions for its own purposes. In addition, because of today's unstable economic conditions in the oil and gas industry, the lower quality of the reservoirs being discovered, and the new recovery methods being developed, it is becoming increasingly difficult to estimate the reserves that will be produced. All ofthese factors have made it imperative to develop a universal set of definitions for reserves that will meet the needs of all users. Part One of the monograph contains the definitions of key terms, the system of reserves classification, and guidelines to illustrate the application ofthe definitions and the classification system. The task of writing the definitions was undertaken by the Standing Committee on Reserves Definitions ofthe Petroleum Society of the Canadian Institute of Mining, Metallurgy and Petroleum, and Part One of the monograph has been published as a separate document comprising the committee's 1993 report. The committee includes representatives of oil and gas companies, geological and petroleum engineering consulting firms, Canadian industry associations, financial and accounting organizations, regulatory agencies, and government. The definitions ofkey terms and reserves classifications presented in Chapter 2 are similar to those currently in use by the oil and gas industry, particularly in North America. They have been reviewed by users in the industry and representatives from regulatory agencies, government departments, industry associations, and technical and professional organizations. Chapter 3 presents the guidelines that illustrate the application of the definitions and the classification system. These are intended to complement the detailed guidelines on reserves estimation methods and procedures that follow in subsequent chapters of the monograph.
The Standing Committee believes that the recommended definitions and guidelines are suitable for use with respect to all types of oil and gas and related substances, including offshore reserves and oil sands. Although those segments of the industry have used somewhat different terms and definitions, the principles reflected in the definitions recommended here are applicable. The fundamental principle is that those quantities that are known to exist and to be economically recoverable are reserves. The total quantities, whether or not they have been discovered, are resources. Reserves and resources are further categorized depending on the level of certainty that they will be recovered.
It is the view of the Standing Committee that current reserves estimation methods and categories, in general, match the recommended definitions and guidelines. The committee, therefore, does not expect that major changes to reserves estimates would result from adoption of the definitions, although it recognizes that for some specific reserves estimates (generally for small pools) changes could be significant. The committee hopes that, over time, reserves evaluators will increasingly conform to the recommendations presented in this monograph and thus contribute to the overall quality and consistency of reserves estimates. The Standing Committee received assistance from many individuals and organizations in the form of comments as it formulated the definitions and guidelines. The committee will continue to communicate with interested parties to ensure that its intent with respect to the recommended definitions is fully understood. The committee welcomes comments on its recommendations as well as any other aspects of reserves definitions and their application. Since comments are being sought from those that use the recommendations, it is reasonable to expect that the definitions may change with time. If they do, the revisions will be available from the Petroleum Society.
3
?
Chapter 2
DEFINITIONS
2.1
INTRODUCTION
The terminology recommended for the classification of estimated quantities ofoil and gas and related substances, at a particular time, is presented in Figures 2.1-1 and 2.1-2. Each term is defined in this chapter. Figure 2.1-1 and its related definitions set the framework for Figure 2.1-2 and its related definitions. The major classifications identified in this chapter are resources, remaining reserves, and cumulative production, each of which can be further divided into
Figure 2.1-1
4
Resources
sub-classifications. Reserves ownership is also discussed in this chapter.
2.2
RESOURCES
Resources are the total quantities of oil and gas and related substances that are estimated, at a particular time, to be contained in, or that have been produced from, known accumulations, plus those estimated quantities in accumulations yet to be discovered.
-I DEFINITIONS
2.2.1
Discovered Resources or Initial Volumes in Place
Discovered resources, which may also be referred to as initial volumes in place (Figure 2.1-1), are those quantities of oil and gas and related substances that are estimated, at a particular time, to be initially contained in known accumulations that have been penetrated by a wellbore. They comprise those quantities that are recoverable from known accumulations and those that will remain in known accumulations, based on known technology under specified economic conditions that are generally accepted as being a reasonable outlook for the future.
Initial Reserves
Future Initial Reserves Future initial reserves are those quantities of oil and gas and related substances that are estimated, at a particular time, to be recoverable from accumulations yet to be discovered by known technology under specified economic conditions that are generally accepted as being a reasonable outlook for the future.
Future Unrecoverable Volumes Future unrecoverable volumes are those quantities of oil and gas and related substances that are estimated, at a particular time, to remain in accumulations yet to be discovered because they are not recoverable by known technology under specified economic conditions that are generally accepted as being a reasonable outlook for the future.
Initial reserves are those quantities of oil and gas and related substances that are estimated, at a particular time, to be recoverable from known accumulations. They include cumulative production plus those quantities that are estimated to be recoverable in the future by known technology under specified economic conditions that are generally accepted as being a reasonable outlook for the future. (Figure 2.1-2 shows how initial reserves are classified.)
Remaining reserves (Figure 2.1-2) are estimated quantities of oil and natural gas and related substances anticipated to be recoverable from known accumulations, from a given date forward, by known technology under specified economic conditions that are generally accepted as being a reasonable outlook for the future.
Unrecoverable Volumes
2.3.1
Unrecoverable volumes (Figure 2.1-1) are those quantities of oil and gas and related substances that are estimated, at a particular time, to remain in known accumulations because they are not recoverable by known technology under specified economic conditions that are generally accepted as being a reasonable outlook for the future.
Remainingproved reserves are those remaining reserves that can be estimated with a high degree of certainty, which for purposes ofreserves classification means that there is generally an 80 percent or greater probability that at least the estimated quantity will be recovered. These reserves may be divided into proved developed and proved undeveloped to identify the status of development. The proved developed may be further divided into producing and nonproducing categories.
Unrecoverable volumes may be further divided into currently uneconomic volumes, which are those quantities that are currently estimated to be technically recoverable, but that are not economically recoverable under the specified economic conditions, and residual unrecoverable volumes, which are those quantities that are unrecoverable by known technologies.
2.2.2
Undiscovered Resources or Future Initial Volumes in Place
Undiscovered resources, which may also be referred to as future initial volumes in place (Figure 2.1-1), are those in-place quantities of oil and gas and related substances that are estimated, at a particular time, to exist in accumulations yet to be discovered.
2.3
2.3.2
REMAINING RESERVES
Remaining Proved Reserves
Probable Reserves
Probable reserves are those remaining reserves that are less certain to be recovered than proved reserves, which for purposes of reserves classification means that generally there is a 40 to 80 percent probability that the estimated quantity will be recovered. Both the estimated quantity and the risk-weighted portion reflecting the respective probability should be reported. These reserves can be divided into probable developed and probable undeveloped to identify the status of development.
2.3.3
Possible Reserves
Possible reserves are those remaining reserves that are less certain to be recovered than probable reserves, which for purposes of reserves classification means that
5
7
DETERMINATION OFOIL AND GAS RESERVES
generally there is a 10 to 40 percent probability that the estimated quantity will be recovered. Both the estimated quantity and the risk-weighted portion reflecting the probability should be reported. These reserves can be divided into possible developed and possible undeveloped to identify the status of development.
2.3.4
Development and Production Status
Each of the three reserves classifications, remaining proved, probable and possible, may be divided into developed and undeveloped categories (Figure 2.1-2). The developed category for proved reserves is often divided into producing and nonproducing.
Developed Reserves Developed reserves are those reserves that are expected to be recovered from existing wells and installed facilities or, if facilities have not been installed, that would involve a low expenditure to put the reserves on production (i.e., when compared to the cost of drilling a well).
Developed Producing Reserves Developed producing reserves are those reserves that are expected to be recovered from completion intervals
Figure 2.1-2
6
Reserves
open at the time of the estimate. These reserves may be currently producing or, if shut in, they must have previously been on production, and the date ofresumption of production must be known with reasonable certainty. Developed Nonproducing Reserves Developed nonproducing reserves are those reserves that either have not been on production, or have previously been on production, but are shut in, and the date of resumption of production is unknown.
Undeveloped Reserves Undeveloped reserves are those reserves expected to be recovered from known accumulations where a significant expenditure (i.e., when compared to the cost of drilling a well) is required to render them capable of production.
In multi-well pools, it may be appropriate to allocate the total reserves for the pool between the developed and undeveloped categories or to subdivide the developed reserves for the pool between developed producing and developed nonproducing. This allocation should be based on the evaluator's assessment as to the reserves that will be recovered from specific wells, the facilities
DEFINITIONS
and completion intervals in the pool, and their respective development and production status.
2.4
CUMULATIVE PRODUCTION
Cumulative production (Figure 2.1-2) comprises those marketable quantities of oil and gas and related substances that have been recovered to date from known accumulations.
2.4.1
Sales
Sales are produced quantities of oil and gas and related substances that have been sold to date.
2.4.2
Inventory
Inventory consists of quantities of oil and gas and related substances that have been produced and are available for future use.
2.5
RESERVES OWNERSHIP
The terminology that is recommended for reporting the ownership of quantities of oil and gas and related substances is presented in Figure 2.5-1. The terms are defined as follows:
Gross remaining reserves are the total remaining reserves associated with the property in which an owner has an interest. Company* gross remaining reserves are the company's lessor royalty, overriding royalty and working interest share ofthe gross remaining reserves, before deduction of any Crown, freehold, and overriding royalties payable to others. Company* net remaining reserves are the company's lessor royalty, overriding royalty, and working interest
Other Owner Interest Reserves
• Lessor Royalty Interests Payable Overriding Royalty Interests Payable
Figure 2.5-1
Reserves Ownership
* The word "Company"may be replaced by moresuitable adjectives to better depictthe ownership of reserves, e.g., ABC Oil and Gas, 9367 LimitedPartnership, John Doe, etc.
7
7
DETERMINATION OFOIL AND GAS RESERVES
share of the gross remaining reserves, less all Crown, freehold, and overriding royalties payable to others.
2.6
SPECIFIED ECONOMIC CONDITIONS
In order for oil and gas and related substances to be classified as reserves, they must be economic to recover at specified economic conditions. The estimator should use, as the specified economic conditions, a price forecast and other economic parameters that are generally accepted as being a reasonable outlook for the future. The revenue, appropriately discounted, must be sufficient to cover the future capital and operating costs that would be required to produce, process, and transport the products to the marketplace. A more detailed discussion of discounting future cash flow is presented in Chapter 21, Cash Flow Analysis, and in Chapter 26,
Uses ofReservesEvaluations. Ifrequired by securities commissions or other agencies, current prices and costs may also be used. In either case, the economic conditions used in the evaluations should be clearly stated. Occasionally, the estimator also may wish to determine the impact of higher or lower price forecasts on estimates of reserves as compared to the most reasonable forecast. These cases (current, higher or lower prices) should not be reported as the most reasonable reserves estimates, but should be identified as sensitivity cases with the assumptions clearly stated. They illustrate the impact of different specified economic conditions on estimates of reserves.
2.7 2.7.1
REPORTING OF RESERVES ESTIMATES Risk-Weighting of Reserves Estimates
Remaining proved reserves, as defined in Section 2.3.1, are those reserves for which there is an 80 percent or greater probability that at least the estimated quantity will be recovered. In instances where additional reserves are estimated in the probable and possible categories, both the estimated quantity and the adjusted (riskweighted) portion should be reported, particularly when the estimates are being aggregated. Proper statistical procedures may be used to derive the expected or risk-weighted reserves from the data. In the deterministic procedure, the best estimate of each parameter is used in the calculation of reserves. The probabilistic procedure quantifies the uncertainty in the resource estimate by using the evaluator's opinion to describe the range of values that could possibly occur
8
for each variable.' If a deterministic procedure is being used and a probabilistic determination is not available, the following equality is recommended to approximate the expected reserves: expected = (proved ) + (p x probable) + (p x Possible) reserves reserves b reserves S reserves where Pb
probability of recovering the probable reserves (80-40%) P, = probability of recovering the possible reserves (40- I0%) =
For individual pools, the amount for the expected or risk-weighted reserves provides the evaluator's best judgement as to the quantity that will be recovered from the pool. The probability used to adjust the estimated quantity for a specific pool should be that considered by the evaluator to be appropriate for the particular circumstance, taking into account the available geological, geophysical and engineering data. It is likely, however, that the quantity actually recovered from a specific pool will be more or less than the risk-weighted estimate. If the number ofpools for which estimates ofreserves are being prepared is sufficiently large, then the sum of the expected reserves should be the evaluator's best judgement as to the total quantity that will be recovered from all the pools. According to the ranges specified in these definitions, the risk-weighting should result in an average risk-weighting of 60 percent for probable reserves (the mid-point ofthe 80 to 40 percent probability range) and 25 percent for possible reserves (the mid-point of the 40 to 10 percent probability range). When the value of the risk-weighted reserves is being determined, the unrisked reserves must be used in the economic analysis. Risk for both the reserves and values should only be applied after the economic forecasts have been completed using total costs to develop the unrisked reserves.
2.7.2
Aggregation of Reserves Estimates
Traditionally, when deterministic approaches are being used, the aggregation of a series of reserves estimates will have been made using the arithmetic method. However, with the increase in the use of statistical methods in reserves determination, the arithmetic method of aggregation may not always be appropriate. Although
• Theseprocedures are described in more detailin Section 4.3.
DEFINITIONS
use of a statistical method of aggregation may be better for reserves estimates, the method of aggregation may be dictated by regulators, auditors or management. Thus, when aggregating a series of reserves estimates, the evaluator should state whether the method of aggregation is arithmetic or statistical. If a statistical method is used, the evaluator should state how it is done. If the proved reserves, which represent an 80 percent confidence level, are summed arithmetically, the total reserves will represent a confidence level that is much higher than would be achieved if the proved reserves were totalled using a probabilistic approach of all the entities and an 80 percent confidence level. Conversely, . the proved plus probable reserves and the proved plus probable plus possible reserves will be overstated when summed arithmetically using a deterministic as compared to a probabilistic procedure. On the other hand, the sum of the expected reserves, as defined in the preceding sections, should be the same as the deterministic (using arithmetic methods) and the probabilistic procedures. This relationship is extremely important in summing reserves, and therefore it is recommended that risk-weighted reserves be used in the aggregation of reserves. In any event, the evaluator should state whether the method of aggregation is arithmetic or probabilistic.
2.7.3
Barrels of Oil Equivalent
From time to time, it may be desirable to report reserves ofoil, gas and related substances in common units. This
is generally done by converting reserves that are not oil to barrels of oil equivalent (BOE). The conversion can be made using either an energy equivalence or a relative value procedure, depending upon the purpose of the conversion. The energy equivalence is only relevant at the burner tip and, since the value in the marketplace is different for various types of reserves and the costs to move the various types from wellhead to the end-user vary considerably, the value of the reserves at the wellhead (or in the ground) is only somewhat indirectly related to energy content. Consequently, for making value-based comparisons, the conversion should be based on the relative values of the gas and related substances compared to the values of oil reserves at the field level. The conversions to BOE are usually made to barrels of "light" oil equivalent. Since medium and heavy oil have values much lower than light oil, it may be desirable that the medium and heavy oil reserves be converted to BOE of light oil as well as converting the gas and related product reserves, to better indicate their real value. Some companies may prefer to convert their reserves using gas as the common unit. The procedure would be similar, except that the converted reserves would be quoted as thousand cubic feet of gas equivalent. It is important, when reserves are reported in BOE or gas equivalent, that the method used and the respective conversion rates be disclosed. A more detailed description ofthe procedure is presented in Chapter 26, Uses ofReserves Evaluations.
9
Chapter 3
GUIDELINES FOR ESTIMATION OF OIL AND GAS RESERVES
3.1
INTRODUCTION
The quantification and classification of estimates of reserves are, by nature, rather subjective processes. Estimates of reserves are developed under conditions of uncertainty, and their reliability and classification are directly related to the quality of the data available, as well as to the competence and integrity of the individual responsible for preparing the estimates. The purpose of this chapter is to elaborate on the classification of estimates ofreserves derived using the two primary reserves determination procedures: deterministic and probabilistic. The categories of proved, probable, and possible have for some time provided a basis for differentiating estimates of reserves to reflect the probability of recovery considered appropriate by the estimator. Stated in another way, the assignment ofthe estimate ofreserves to the three categories has provided a qualitative measure of the probability that a particular estimate of reserves will, in fact, be realized. However, for some time there has been discussion as to whether a more rigorous approach should be adopted to describe the degree of probability associated with the specific reserves categories. Some observers view the use ofterms such as "high degree of certainty" to describe reserves classification categories as too subjective, and believe a definitive statistical probability of recovery would give users more confidence in utilizing- the estimates of reserves provided for each of the categories. For this reason, consideration has been given to a means to further quantify the degree of probability associated with each of the categories. The probability ranges adopted by the Standing Committee on Reserves Definitions for the definitions ofproved, probable, and possible reserves are intended to more explicitly quantify the probability of recovery associated with each of the reserves categories, both on an absolute and on a relative basis. The ranges provide an assessment that is more quantitative in nature than some prior definitions.
10
The use of probabilities to assist in the categorization ofreserves does not eliminate subjectivity from the process. It remains incumbent on the evaluator to ensure that the basis for the estimate of reserves and the category to which the estimate is assigned are clearly reported. The guidelines and examples given are intended to assist in this regard. The reserves classifications and associated probability ranges are applicable to estimates of reserves derived using either deterministic or probabilistic (stochastic) calculation procedures.
3.2
METHODS OF CALCULATING RESERVES
3.2.1
Deterministic Procedure
The deterministic procedure is the most commonly used method ofreserves estimation in Canada. Ifthe true values of all parameters used in any calculation were known, a true or deterministic value could be calculated. However, due to the uncertainties in the geological, engineering and economic data, for the purposes of reserves estimation using the deterministic procedure, the "best estimate" ofeach parameter is used in the calculation of reserves for each specific case. As a result, the probability distribution of the input parameters is generally not formally considered in the classification of reserves calculated using this method. Estimates ofreserves calculated using the deterministic procedure should be assigned to the proved, probable, and possible categories based on the probabilities inherent in the estimates. The assignment ofthe estimates of reserves to the respective classification categories should be consistent with the prescribed ranges ofprobability, taking into account factors such as the stage in the producing life ofthe reservoir, the amount and quality of geological and engineering data available, the availability of suitable analogous reservoirs and, perhaps most importantly, the evaluator'sjudgement as to the uncertainty inherent in the estimate.
GUIDELINES FOR ESTIMATIONOF OIL AND GAS RESERVES
The assignment of reserves estimates to the respective categories using the deterministic procedure normally uses one of two approaches. In the first, the evaluator develops a "best estimate" of reserves for each of the categories, using consistent parameters. Using this methodology, the evaluator effectively establishes a range of estimates of reserves, with the proved estimate based on parameters for which a high probability can be attributed, and additional estimates of probable and possible reserves based on parameters for which there is a lesser probability of occurrence. The effect of this is to progressively increase the estimated quantity as it is moved from the proved to probable to possible categories, with the overall range of estimates dependent upon the uncertainty inherent in the specific parameters upon which the estimates are based. In the second approach, a single estimate of reserves is derived for the pool and then allocated to the respective reserve categories based on an assessment of the portions of the estimate that would satisfy the probability ranges for each of the reserves categories. In making this determination, the evaluator must make a subjectivejudgement as to the uncertainty inherent in the single estimate and, therefore, the extent to which it can be allocated to the proved rather than the probable or possible category. As already noted, where probable or possible reserves have been estimated in addition to proved reserves, they should be adjusted (risk-weighted) and added to the proved reserves to result in the expected reserves. In summary, using the deterministic procedure, estimates of reserves are calculated and assigned to the proved, probable, and possible categories using primarily subjective criteria, the overall basis being that the assigned quantities satisfy the probabilities established for each of the categories. It is incumbent on the evaluators to provide the supporting rationale for the categorization of the reserves estimates.
3.2.2
Probabilistic Procedure
The probabilistic or stochastic procedure is less commonly used in Canada. It is more suitable for circumstances where the uncertainty is high, such as for reservoirs in the early stages of development, frontier areas, or areas where new technology is being applied. As the level of uncertainty increases, it is generally agreed that the probabilistic procedure becomes more relevant and the deterministic less reliable. Rapidly expanding computer applications also facilitate the use of the probabilistic procedure.
This method uses the statistical analysis of data. Relative frequency curves established for each variable describe the range of possible values for each, as well as the probabilities that these values will occur. After frequency distribution curves have been established for each variable to be used in a reserves classification, the Monte Carlo (described in Section 22.4.4) or a similar method is used to estimate a value for reserves. A single sample of each variable is taken randomly from each probability distribution and used to calculate a single value of the dependent variable. This procedure is repeated a large number of times to ultimately create a frequency distribution curve that describes the range of estimates of reserves and the probabilities of achieving particular estimates. Once the measures of central tendency (the mean or arithmetic average, the mode or "most likely" value, and the median or "middle" value) and the dispersion (range, standard deviation, and percentiles), have been determined using this technique, estimates of reserves may be assigned to each of the proved, probable, and possible categories. The assignment of the estimates of reserves to the respective categories should be consistent with the probabilities outlined in the reserves definitions, proved reserves being those with an 80 percent or greater probability, and probable and possible reserves having lower probabilities. The relative cumulative frequency distribution curves may be used as the basis for the assignment of estimated quantities to the reserves categories. Again, the evaluator must clearly describe the supporting rationale for the categorization ofestimates ofreserves. Like the estimates derived using the deterministic procedure, the probable and possible reserves should be adjusted (risk-weighted). Since the probabilities have been established through the probabilistic process, they should be used to adjust the respective estimates. It should be noted that the probability associated with the estimate of reserves for a pool should increase as the pool is developed and produced over a period of time. As the overall probability of recovery increases, the estimate of the proportion of reserves considered to be proved is likely to increase, with a diminishing proportion in the probable and possible categories. The objective of the evaluator should be to minimize the extent to which it is necessary to reduce estimates of proved reserves over the life of a pool for reasons other than production, although there may be circumstances (i.e., a significant price decline) where such reductions are necessary.
11
,
DETERMINATION OFOIL AND GAS RESERVES
3.3
GUIDELINES FOR SPECIFIC METHODS
The guidelines and examples that follow are designed to provide guidance for evaluators on the calculation of proved, probable and possible reserves, using the following methods for determining reserves: • Volumetric • Material balance • Decline curve analysis • Reservoir simulation This section also deals with the calculation of reserves of natural gas liquids (NGLs) and sulphur. It must be emphasized that the guidelines touch on some key factors related to reserves estimation, but are not all-inclusive. In the final analysis, the calculation and categorization of reserves depend upon the judgement ofthe evaluator as to the probability of recovery of the reserves of oil and gas. It is intended that the guidelines will lead to more uniform practices of reserves calculation in each category, and thus to reserves estimates that will be more comparable and consistent throughout the industry, the financial community, and the government agencies that use them.
3.3.1
Volumetric Method
The volumetric method is the most commonly used approach to estimating reserves in the early stages of production from an oil or gas field. As more data become available, the estimate may be refined, sometimes through the use of other reserves estimation methods. Often the volumetric estimates are useful for comparison with other methods. The volumetric method is used by employing the standard reserves equation with the appropriate choice of parameters. For various parameters in the equation, the guidelines provide suggestions for choosing the appropriate value, according to the category of reserves being calculated. Pool Area
The parameter that often has the greatest variability in the reserves equation is the area chosen to represent the areal extent of the pool. Thus, the choice of the value for the area plays a particularly important role in computing reserves in each category.
Single-Well Pools
For single-well pools, the area must be consistent with the reserves category, recognizing the geological and engineering information with respect to the single wellbore and the geological and other information available for single-well pools in similar formations. In the case of an isolated gas well with little or no geological control, it is a frequent practice to assign reserves to one section,* a frequently used regulatory spacing for gas wells. However, one section should only be .assigned as proved reserves if a review of similar wells in the same or a similar formation has satisfied the evaluator that such an area can be assigned 'with a probability of 80 percent or more. If the review of similar wells shows that a smaller area, such as one halfsection or even one eighth-section, can be expected to have a high degree of probability, this reduced area should be used for proved reserves. On the other hand, in situations such as a blanker sandstone, the review of similar wells may justify the assignment of more than one section if it can be demonstrated with high probability that the well will drain reserves associated with the larger area. In the event that an evaluator is reasonably confident that gas would be recovered from an area, say one section, but not with a high enough probability for the reserves to totally qualify as proved, then some lesser area for which there is a high probability, say one halfsection, should be assigned as the proved area. The remaining half-section ofthe normal spacing unit might then be assigned to the probable or possible category, depending on the degree ofprobability that such reserves would be recovered. For single oil wells, the area assigned would generally be less than for gas wells because the flow characteristics for oil result in smaller drainage areas. A typical practice is to assign proved reserves to areas ranging from one quarter-section for light crude oils to one sixteenth-section or less for heavy crude oils. Such assignments should be made only when a review of similar wells demonstrates that such reserves can be expected with a probability of 80 percent or more. The process used to assign areas to single oil wells should otherwise be similar to that for gas wells, with an assignment that reflects the probability that the area can be drained at the level required for each reserves classification.
*One section = 259 hectares, 640 acres,or 1 square mile.
12
z
GUIDELINES FOR ESTIMATION OF OIL AND GAS RESERVES
In certain cases such as sheet sandstones, even though only one well has penetrated gas or oil, information may be available, as a result of knowledge about nearby abandonments and the regional geology, that justifies the preparation of an isopach map. This situation is illustrated for an oil pool in Figure 3.3-1, which shows the zero pay limit. If the probability of a one quartersection pool is very high, based on a study of similar pools in the area, then the one quarter-section containing the well could be assigned as proved reserves. The remaining three quarter-section parcels offsetting the well, and within the zero limits ofthe isopach map, could also be assigned reserves as additional proved or probable or possible depending on the degree of probability that the oil will be recovered. These reserves, however, should be in the undeveloped category. The assigrunent of reserves for single gas wells with considerable geological control can be handled in a manner similar to that detailed for the oil well in Figure 3.3-1, except that the estimated drainage area for gas will usually be larger, depending on the available geological and other data. An area larger than the assigned area determined as described may be used
I
------------
\ -, -<>-
Figure 3.3-1
0
t-...
Multi-Well Pools
In multi-well pools, the area between wells should be considered to contain proved reserves if the areas assigned on a single-well basis overlap or are separated by a very small area, or if material balance calculations or production data and pressure response clearly demonstrate that the wells are in the same pool. There will, however, be many situations where such conclusive information is not available and the evaluators must use their judgement, based on geological and other data, regarding the areal extent and the assignment of additional reserves to adjacent areas. For wells that are in separate pools, the preceding methodology for assigning reserves for single-wellpools should be followed. If more than one well can be included in a pool, the type of procedure described in the following example might be used. Example
Figure 3.3-2 shows the zero pay limits for a multi-well conventional gas pool. It is important to emphasize that
",-<>-
-<>- L-0
depending on information on pressure, drillstem test results, and seismic data.
\
0
-~---~-~-~--
•
WELL LEGEND
o
Location
-<>-
Abandoned 011
* •
--
)
Gas
[7 1 mi 1 km
Single Well Oil Pool with Good Geological Control
13
DETERMINATION OF OIL AND GASRESERVES
RANGE 36
~
/
31
-.
V
1/
\
/
\
Figure 3.3-2
~
6
/
1/
31
~
f-
WELL LEGEND
0 1
36
-
• ~
Location
Abandoned Oil
Gas
L-4
D
One Section
Conventional Gas Pool, Zero Limit of Net Pay Map
this example illustrates a procedure that is useful for conventional oil and gas pools. The areas assigned relate to the particular natural gas reservoir and would differ in other geological settings. Knowledge respecting the geological formation is such that the evaluator is prepared to make a proved area assignment offour sections for a single well. The map is constructed using all pertinent data, such as the net pays encountered in the three gas wells, and information on dry holes that indicates the limits of the pool. Perhaps seismic information and pressure data, along with experience in the area, suggest that the two wells in the west are in communication with the one in the northeast The first step is to block in a 2 by 2 section area around the productive well and within the zero net pay isopach limit, as shown in Figure 3.3-3. These areas would be assigned proved reserves. Proved reserves would also be assigned to corridors of one mile or less in width between the proved areas around each well and any intervening corridors between the proved areas (Figure 3.3-4). After limits had been established for the proved reserves, a border one mile wide would be drawn around
14
s:0
/ 36
<,
J:
en z
I
1
1\
0-
~
~
-?-
36
the proved limits as an indication of the proved plus probable limits (Figure 3.3-5). As with the proved reserves, any corridor less than one mile between the proved plus probable limits would also be assigned to the probable category (Figure 3.3-5). The procedure would be continued for possible reserves ifany area were left within the zero pay limits. For oil, a similar approach for assigning areas in multiwell pools can be used, but the area assignments would usually be smaller.
Presence of Hydrocarbons In order to estimate any reserves, the presence of hydrocarbons must have been confirmed by production data or by a demonstrated ability to flow based on the results of drillstem tests. If production and test data are not available, reserves may be estimated based on information from cores, or provided that the reservoir is analogous (from the standpoint ofgeological and petrophysical characteristics) to producing or tested reservoirs in the same area. Reserves should be assigned only to reservoirs that have been penetrated by a wellbore; otherwise, quantities should be categorized as a resource.
GUIDELINES FOR ESTIMATION OFOILAND GASRESERVES
RANGE 36
7
I
/
~
/'
'r>.
~~V '/ / /
36
\
/ 0-
V~~
6
/6
en Z
l-
WELL LEGEND
0
-?-
• ~
1
V -----
:s:0
/
/
31
V \ ~/ ~ / vv
J:
/
V
V
Figure 3,3-3
//
v //- V/
.v / / 'V
\
r-.
V-
/
/ /
31
36
IZZ2I
~
D
Location
Abandoned Oil Gas
Proved
One Section
Conventional Gas Pool, Zero Limit of Net Pay Map with Individual Well Assignments
RANGE 36
31
r-,
V-
;7V /
V/-
~/ ~ ~/
~
~ \~ / ~ 'V '< VV
Figure 3,3-4
/
/
0-
J:
en Z
:s:
/
~
/
~/ //V r/
\
/
:/~ ~
/ / '/ //V V
/. / / / / // / V //V V
\
//
/
36
/
6
6
V
31
--
~
WELL LEGEND
0
-?1
36
• ~
IZZ2I
D
Location
Abandoned 011 Gas
Proved
One Section
Conventional Gas Pool, Zero Limit of Net Pay Map with Area of Proved Reserves
15
?
DETERMINAnON OF OIL ANDGASRESERVES
RANGE 36
a. :c
en z
s:0
t-
WELL LEGEND
0
-
•
36
*
IZZa
Location
Abandoned Oil
Gas
Proved
Probable
D Figure 3.3-5
Conventional Gas Pool, Zero Limit of Net Pay Map with Area of Proved Plus Probable Reserves
Reservoir Parameters
Values ofpay thickness, porosity, and fluid saturations, when combined with area, permit a calculation of the volume of oil or gas contained in a reservoir. These parameters are estimated from analyses of cores or petrophysical well logs. In many situations, core analysis is not available, and well logs indicate the presence of oil or gas, but do not allow reliable estimates ofporosity or fluid saturations. Where the geological formation is known to be productive in the region, a pay thickness based on the logs for the well and the values ofporosity and fluid saturations taken from nearby wells in similar formations (values that may be expected to have a high probability), may be used to calculate proved reserves. In such cases, where these parameters can only be estimated with a lower probability ofoccurrence, probable or possible reserves might be calculated. In the estimation of reserves, the values used for pay thickness, porosity and fluid saturations must always be consistent. This requires proper use ofcutoff values, well log calibration, and proper petrophysical calculation methods.
16
One Section
Averaging these parameters over multi-well pools is also important in estimating and categorizing reserves. If reliable estimates for many wellbores exist for any or all of the parameters, they should be applied over the intervening area between wells and the edge ofthe reservoir by contouring or other appropriate averaging methods. The calculation of the average, particularly if it is by contouring, should have regard for the geology and for any other factors that might influence the shape of the reservoir. Where insufficient individual well data respecting any of these parameters are available to allow for contouring, averaging should be done in a manner consistent with the probability necessary to support the particular category of reserves being estimated. Certain other reservoir parameters are needed to estimate reserves, particularly for purposes of converting the volumes of oil or gas contained in a reservoir to volumes that will be recovered and marketed. These include reservoir pressure and temperature and hydrocarbon fluid composition. The choice of such parameters does not usually dictate the categorizing ofreserves estimates as proved, probable, or possible.
h
GUIDELINES FOR ESTIMATION OF OIL AND GAS RESERVES
However, particularly for proved reserves, the parameters must be based on reliable data or be determined through valid comparison to similar reservoirs in a manner that reflects the appropriate level of probability.
are available. When economic producibility limits are coupled with the material balance, reserves are determined. In its simplest form, the material balance equation may be written as initial volume
Recovery Factor
=
volume remaining + volume removed
Estimates of recovery factor are based on analysis of production data from the pool in question, by analogy with producing pools in an analogous reservoir unit, or by engineering analysis, without analogy or production data. The estimator should keep in mind that recovery factors may be influenced by other factors, such as well spacing, the anticipated compression, the drive mechanism, and reservoir and fluid properties.
Since oil, gas and water are present in petroleum reservoirs, the material balance equation may be written for the total fluids or for anyone ofthe fluids present. For gas reservoirs, the frequently used plot ofreservoir pressure, adjusted for the gas compressibility (P/Z), vs. cumulative production is a material balance method.
For proved reserves, the recovery factor may be determined from a detailed reservoir study, or by comparison with detailed studies of analogous reservoirs where the recovery factor can be estimated with a high degree of probability.
• Reservoir pressure and temperature
For probable and possible reserves, the value used for the recovery factor may be similar to that used for the calculation ofproved reserves, the different categorization ofreserves being accounted for in other parameters. However, a larger recovery factor may be justified on the basis ofgeological data that indicates improved reservoir parameters such as porosity and permeability in certain portions of the field. Where a range of recovery factors is known from analogous reservoirs with similar characteristics, values corresponding to the middle to upper end of the range may be used for probable and possible reserves estimates. In some cases the recovery factor for proved reserves has been estimated on the basis of an 80 percent or greater probability, and yet the characteristics of the formation indicate that better recovery might occur. In other cases the recovery factor for proved reserves has been estimated lower due to an anticipated recovery problem (i.e., water influx in a gas reservoir), but there is only a chance that the problem will occur. In these situations the evaluator might use a higher recovery factor and assign the incremental reserves to the probable or possible categories, depending upon the degree of probability of their recovery.
3.3.2
Material Balance Method
The material balance method is employed to estimate the volume ofhydrocarbons in place in a reservoir when appropriate geologic, production and laboratory data
Four groups of data are required for a material balance: • Fluid production • Fluid analysis • Core analysis and petrophysical logs In addition to these data, it is highly desirable to know the type of reservoir mechanism that is operative in order to expedite estimation ofthe volume of the initial hydrocarbons in the reservoir. As with other methods, the better the quality of the data, the higher the degree of confidence in the results. The evaluator, in categorizing reserves, must consider the probability that the reserves in question will be recovered. The volume in place estimated by the material balance method might have a high enough probability to be considered as proved in the following situations: - Where significant data are available, particularly fluid production and reservoir pressure data, and the reservoir drive is known - Where production and reservoir data are limited, but the reservoir is analogous to reservoirs in the immediate vicinity and same geologic horizon - Where such data are of sufficient quantity and quality to have established the reservoir drive mechanism - Where production and reservoir data are limited, but the estimate is supported by a calculation of the hydrocarbons in place by the volumetric method For gas reservoirs, where there is a strong linear relationship between P/Z and cumulative production (Figure 3.3-6), the probability ofrecovery is likely high enough to assign the quantity indicated as proved reserves. However, no additional reserves should be assigned beyond the proved reserves, since no significantly different interpretation ofthe data would be reasonable.
17
_
DETERMINATION OFOILANDGASRESERVES
30000
20000
r-."
PIZ (kPa)
<,
~
10000
Average Pressure (kPa)
Z
72102 86107 87/08 86107 89/02 89/09 90/0B
21540 20078 18705 14623 13258 10742 7357
0.875 0.871 0.869 0.874 0.879 0.894 0.922
PIZ Cum. Prod. (kPa) (10 8m') 24621 23063 21532 16740 15086 12018 79n
i."i
0.0 96.0 209.0 582.0 724.0 920.0 1 210.0 i'·i.
I: :
ii~
I:i I:: IiiIX
f~
<, 1-----1----- I- - -
o
r-,
Oat.
o
250
500
-E=r~.'!'~C-r~!. 750
1 000
--r-. ~ ------
1 250
1 500 8
~
1750
2000
a
Cumulative Production (10m) PROVED 8 ,
OGIP=1800x 10 m ReserveS=1600xl0
Figure 3.3-6
Material Balance (Gas Reservoir)
There are a number of other situations where reserves estimates from material balance or some portion of the estimate might have an associated probability level that results in their being considered probable or possible reserves: - Where significant production data are available, but the reservoir drive mechanism is uncertain or the magnitude of the reservoir drive is uncertain - Where production and reservoir data are limited and there are no analogous reservoirs in the immediate vicinity - Where production and reservoir data are limited and the estimate is not supported by volumetric determinations For a gas reservoir, where the P/Z data do not give a definitive linear correlation, asin Figures 3.3-7 and 3.3-8, the resulting reserves that should be classified as proved are those that represent the quantity that can be estimated to be recoverable with an 80 percent probability. Proved plus probable reserves might reflect a larger volume than the data indicate may be recovered. In Figure 3.3-7, the scatter of points should encourage the evaluator to analyze the quality of the
18
6m3
data in terms of the type of pressure measurement (bottom-hole, drillstem test), and the type of recorder (mechanical or electronic). With respect to Figure 3.3-8, the evaluator should develop an understanding ofthe reservoir drive and depletion mechanism in order to accurately classify proved and probable or possible reserves. The apparent bending of the material balance plot may be interpreted as gas migration from edge or tight areas of the reservoir, or pressure support from an underlying aquifer. Use ofa reservoir simulation model might assist in this analysis.
3.3.3
Decline Curve Analysis
The analysis of a production decline curve provides estimates of three important items of information: • Remaining oil and gas reserves to be recovered • Future expected production rate • Remaining productive life of the well or reservoir In addition, an explanation of any anomalies that appear on the graph is useful. The analysis is only valid provided the well will not be altered (i.e., fractured or acidized) and the reservoir drainage is constant.
GUIDELINES FOR ESTIMATION OFOIL AND GASRESERVES
Average 40000
-,----,----,----.---rI
a
P/z
Date
Pressure
P/Z Cum. Prod. (kPa) (10'm')
Z
(kPa) INIT. 76/06
30991
0.987
31109
81/06 84/09 86/09 86/09 87/05 89/07
21380 20277 16602
0.988 0.922 0.919 0.913 0.915 0.915 0.913
19001
18519 18471
31391 31479 23182 22075 18179 20762 20237 18036
0.0 13.4 72.7 103.4
122.9 122.9
128.6 144,8
(kPa)
10000
+----+---+----j----"'f.,;;:-=:O""d----l-----l Economic Limit
o+---+---+----I----I-_ _~.L......:::>.I-..J--"'-I 100 o 50 150 200 250 300 350 Cumulative Production (10'm') PROVED
OGIP=300xl0
PROVED + PROBABLE
6m3
6m3 Reserves = 270 x 10
Figure 3.3-7
6m3
6m3 Reserves = 315 x 10
Material Balance (Scattered Datal
Date
Average Pressure (kPa)
85/11 75/08 76/10 76/11 77/08
21067 16858 14844 15306 13631
78/09
12604 13411
30000
20000
.... ....
PIZ (kPa)
~
10000
80/07 86/05 87/08 88/06
~ '<,
---- -----
- ~CEClOli9!~.!!-
o
100
200
300
8936 8556 8494
PIZ Cum. Prod. (kPa) (lo'm')
Z 0.924 0.899 0.902 0.901 0.906 0.910 0.920 0.929 0.932 0.932
22800 18761 16451
16989 15044
...
0.0 102.9 138.4 138.4 162.3
13852
190.5
14 SIT 9618 9184 9115
237.1 340.6 358.4 368.2
~~ -----
o 400
Cumulative Production (10'm')
500
:"'-..l ::::--600
700
,, OGIP=620xl0 m PROVED
6m3 Reserves=550x 10
Figure 3.3-8
OGIP=350xl0
PROVED + PROBABLE 6 ,
OGIP=675xl0 m 6m3 Reserves = 590 x 10
Material Balance (Reservoir Drive and Depletion Mechanism)
19
DETERMINATION OFOil AND GAS RESERVES
As with all other methods, the categorizing of reserves by decline curve analysis is dependent upon the judgement ofthe estimator. Important considerations include the amount and quality of data, the variability of the profile, and an understanding of past and future production policy and the depletion mechanism.
In a case where well-established trends are not evident, proved reserves should be restricted to the minimum quantity that the evaluator believes will be recovered with an 80 percent probability. Figure 3.3-10 shows such an example. Proved reserves are estimated by projecting the steepest portion ofthe production decline data. In this case, the incremental volume of oil that may be recovered if the lower rate of production decline prevails might be classified by the evaluator as probable or possible.
Because ofthe empirical extrapolation, a decline curve can usually have a wide range of interpretations. The range depends upon the production history of the property. For example, if there is limited prior production history, a wider range of interpretations is possible than for a well or property in the stripper stage of production. It is valuable to understand the production recovery . mechanism of the formation (or the same formations in the area) and the various characteristics of the well (net pay, permeability, and zone of completion). Also, each specific interpretation is a function of the experience, integrity and objectivity of the individual doing the evaluation. Determining reserves from historical graphs of production data that exhibit strong consistent decline characteristics should be straightforward. When a high degree of probability exists, as in Figure 3.3-9, proved reserves only would be assigned.
This situation could also apply when the type ofdecline pattern is not obvious. Figure 3.3-11 illustrates a case where either an exponential or a harmonic decline could be used to extrapolate the data. In this example, reserves determined from the exponential curve might be assigned to the proved category, since there is a high probability that at least this volume will be recovered. The harmonic curve reserves might be classified as probable or possible, depending on the probability of recovery, as judged by the evaluator. In this example there is a large difference between the estimates using the different interpretations, and this suggests considerable uncertainty. Thus portions ofthe quantity in excess of the proved reserves could be classified as probable and possible.
20
PROVEO ~
....."
~
-.. E
n.
~
co c
II: 0
;:
10
.LA
'Y~
U
::J
."
2
e,
(5
.
~v\
5
o
3m3
Reserves = 33 x 10
15
''\.."
~
f----------- _ E'<.O!!~,,!!"-Ll"!!!.. _ ----------- ~---o
30
20
10
3
Cumulative Oil Production (10 m3)
Figure 3.3-9
20
Decline Curve, Proved Reserves
40
GUIDELINES FOR ESTIMATION OFOILANDGAS RESERVES
20
PROVED
15 - I t - - - - , - - - - - + - I - - - - - - H M
--
PROVED + PROBABLE + POSSIBLE
Q
* :s II:
c
6m:3 Reserves = 155 x 10
Reserves = 176 x 10
6m3
10
o
::J
'D
e
Q.
i5
5-++----+-1----Economic Limit
o+--------+------\--------f-l-"'"---..L.::::>...-J o 50 100 150 200 3
Cumulative Oil Production (10· m
Figure 3.3-10
)
Decline Curve, Cumulative Gas Production
16 PROVED
3m3 Reserves = 17 x 10 ~
:E §.
12
PROVED + PROBABLE + POSSIBLE
M
3m3
Reserves = 25 x 10
Q) ~
'c"
II:
., 0
8
~
U
::J
'D
0 ~ Q.
i5
4
1\
-, .~
ECOnOm!C Limit
o
t
o
5
.. """-
~Harmonic
Exponential/f~
.T
10
~
15
20
25
30
3
Cumulative Oil Production (10 m3)
Figure 3.3-11
Decline Curve, Cumulative Oil Production
21
DETERMINATION OF OIL AND GASRESERVES
3.3.4
Reservoir Simulation Method
A reservoir simulator is a tool that is used to simulate the processes that take place in producing a reservoir. Simulation is often done to optimize recovery by analyzing various reservoir development plans, methods of production, and the complexity of the reservoir itself. Although reservoir simulation methods are complex, they include a combination of the physical principles and analytical techniques of one or more of the other methods of reserves estimation. The criteria for categorizing reserves would include the amount and quality ofproduction and pressure data, the validity of the model and its demonstrated reliability with comparable reserves, and the ability to history match. To illustrate, if sufficient amounts of good geological and performance data are available to allow for a reasonable history match, and if the estimator is using an appropriate simulation model that has been used successfully in reservoirs similar to the one being studied, projections ofrecovery under primary mechanisms and the specified economic conditions might be considered proved reserves. Ifthe situation being modelled is an improved recovery mechanism, these criteria and the guidelines given in Section 3.3.5 for categorizing improved recovery reserves generally apply. This means that for existing and operating improved recovery projects where an appropriate simulation model is being used, adequate data exist, and the response to the data is consistent with the simulation results; or where future projects can be expected with a high probability in reservoirs ofthe type where the model has been shown to give reliable results, the simulated recovery can be considered as proved reserves. At least a portion of the simulated recovery should be categorized as probable or possible or not considered as reserves, depending on the evaluator's views on the probability that the additional oil or gas will be recovered in the following situations: - Where the model has not been shown to give reliable results for the same type ofimproved recovery project in a similar reservoir Where insufficient data are available to properly use the model - Where the installation of an improved recovery project cannot be predicted with a high probability
3.3.5
Reserves from Improved Recovery Projects
The calculating and categorizing of reserves from improved recovery projects should be based on the
22
judgement of the evaluator and on information such as observed performance, the results of pilot projects, the performance of projects in analogous pools, and engineering studies. To illustrate, reserves attributable to improved recovery projects may be considered as proved reserves provided certain conditions are met. Such situations occur when the production response from a project corresponds to the results predicted by engineering analysis or where the improved recovery project has been in operation for a considerable period and the analysis of a decline curve can be used with confidence, or where a history-matched simulation is available. If the production response has fallen short of original predictions, the observed production data should be used for calculating and categorizing reserves. Proved reserves may be attributed to other areas of the pool where an improved recovery project has not yet been applied, provided that it is highly probable that a project will proceed, and that the geological and other reservoir characteristics are equivalent or superior to those for areas where an improved recovery project is in operation. If a successful scheme has been implemented in a similar pool that has analogous reservoir characteristics, proved reserves due to improved recovery may be assigned, provided the evaluator is convinced that the analogy is sound and that there is a high probability that a project will proceed and be successful. Probable or possible reserves can be assigned in other cases where the improved recovery method has been applied successfully to analogous reservoirs, but where there is a lower probability that a project will go ahead and be successful. Proved, probable or possible reserves may be attributable to a planned workover treatment, improvement to equipment, or other procedures, depending on the evaluator's judgement respecting the probability of success.
3.3.6
Related Products
Natural gas liquids (NGLs) are the liquid hydrocarbon components recovered from natural gas. If they are recoverable, they must be calculated and reported as reserves of either natural gas or natural gas liquids, but not both. The first test of recoverability ofNGLs is whether they will be produced as part of the stream of raw natural gas. If the fluid properties and reservoir pressures are such that the composition ofthe produced raw gas stream will significantly change with time due to retrograde or other phenomena, this must be reflected in the calculated reserves. Components of natural gas that wiil
GUIDELINES FOR ESTIMATION OF OIL AND GAS RESERVES
liquefy in the reservoir and not be recoverable must not be included in reserves of either NGLs or natural gas.
and therefore neither the natural gas nor the NGLs could be categorized as remaining proved reserves.
If cycling or other special means of producing the reservoir is planned in order to reduce the liquid losses that might otherwise occur, the NGLs that would be so recovered can be categorized only as proved reserves where their recovery can be estimated with a high probability. This would require sufficient reservoir and fluid data to make an accurate detailed projection ofproduction and, also, the special means of production would have to be actually in operation or expected with a high degree of probability.
Where the removal of NGLs from the raw gas is not required but is being planned, the removal of the liquids must be economically feasible or else the NGLs cannot be categorized as proved reserves. If the removal ofthe liquids is not economically feasible, they would be included as part of remaining proved reserves of natural gas.
Where the prevention of losses in the reservoir is less certain, such NGLs should be categorized as probable or possible or not considered a reserve, depending on the probability of their recovery. For most reservoirs, the composition of the produced gas will not significantly change with time. For these reservoirs, the only test of recoverability of NGLs relates to whether they will be recovered as liquids or remain in the natural gas. This is also a second test for those reservoirs previously mentioned where the NGL content of the produced gas will change with time. The first criterion in terms of their classification as reserves is that NGLs can only be categorized as proved ifthe raw gas from which they are to be removed will, after processing, result in marketable gas that is categorized as remaining proved reserves. Some portions of some hydrocarbon components, particularly ethane and propane, may not have to be removed from raw gas to make the gas marketable. Since the technology for removal of essentially 100 percent of all NGLs is well-proven, the only test of their recoverability from a proved natural gas reserve relates to whether the liquids would be recoverable at the specified economic conditions under which the estimates of proved reserves are being made. Where the removal of NGLs from the raw gas is necessary in order to make the gas marketable, the removal of the liquids must be economically feasible or the natural gas would not be economically recoverable as marketable gas at the specified economic conditions,
If raw gas containing NGLs that will be marketable natural gas after processing is categorized as either probable or possible reserves, the NGLs must be categorized in the same manner. With respect to sulphur reserves, essentially all of the hydrogen sulphide and other sulphur compounds must be removed from the raw natural gas and converted to elemental sulphur to meet environmental and other standards. The necessary technology exists, and the key question is whether the recovery of the sulphur is economically feasible at the specified economic conditions. If it is not, the natural gas would not be economically recoverable as marketable gas and thus could not be categorized as reserves. If the sulphur in question is economically recoverable but is contained in natural gas categorized as probable or possible reserves, the sulphur must be categorized in the same manner. In some reservoirs, usually where the gas has a very high HzS content, the pressure, temperature and fluid properties are such that some ofthe sulphur will liquefy or solidify in the reservoir and will not be producible without special production measures. Where such conditions are known to exist or can be expected, the sulphur that would liquefy and remain in the reservoir cannot be categorized as proved reserves unless special production measures for dealing with the problem have been demonstrated to work successfully. They would have to be feasible at the specified economic conditions, and either have been implemented or have a high probability of being implemented. Where these criteria are not met, at least a portion of the sulphur should be categorized as probable or possible or not considered as a reserve, depending on the overall probability of its recovery.
23
r
PART TWO DETERMINATION OF IN-PLACE RESOURCES
n !
_
- I
.. Chapter 4
OVERVIEW OF PART TWO
4.1
INTRODUCTION
Part Two deals with the estimation of hydrocarbons in place, the economically recoverable portions of which are classified as reserves. The estimation of initial in-place resources involves contributions from several disciplines, primarily geology, geophysics, petrophysics, and engineering, but contributions in varying degrees may also be required from specialists in chemistry, physics, economics, and other geological-engineering disciplines. It is important that the size, or at least the range in size, of a potential resource be determined using consistent approaches and considering the interrelationships ofthe parameters used to make the estimate. The size of the resource forms the basis for the determination of how much oil, gas, and related products may ultimately be produced for society's use, and for the formulation of operation plans and the necessary business decisions. Volumes ofthese discovered resources may be estimated by either a volumetric or a material balance method of calculation. These methods are described in Section 4.2. Section 4.3 describes the deterministic and probabilistic procedures for estimating in-place resources. Sections 4.4 through 4.7 briefly discuss sources and reliability of data, the interrelationship ofparameters, the ways in which resource estimates are used, and the background and experience of evaluators.
4.2
RESOURCE ESTIMATES
4.2.1
Volumetric Estimates
Reservoir Volume The first step in a volumetric calculation ofhydrocarbon resources is an estimation of the volume of subsurface rock that contains oil and gas. The volume is derived from the thickness of the reservoir rock containing the hydrocarbons and the areal extent of the accumulation. The important geological considerations in establishing a realistic estimate of reservoir volume include
the depositional environment of the reservoir beds, the history of any structural deformation of those beds, the trapping mechanism for hydrocarbon accumulation, and the positions ofthe various fluid interfaces. Mapping the extent and configuration of the hydrocarbon accumulation requires the evaluator to have an understanding ofthe geological concepts of sedimentation and the structural attitudes ofthe reservoir rock that control the limits and define the geometry of the deposit. Well samples and cores, well logs, seismic and well test data, and pressure information are all used to interpret the extent of the oil or gas pool. Visualization of the accumulation in three dimensions is necessary to portray a realistic mapped interpretation.
Rock and Fluid Properties The properties of the reservoir rock and the particular hydrocarbon are also important factors in the volumetric estimate of the resource. Although the volumes of hydrocarbons are calculated at subsurface depths, they are converted to standard surface conditions oftemperature and pressure for measurement and recording. The standard surface conditions in a particular location become the "base" temperature and pressure. The following properties are. important in volumetric procedures: 1. Porosity, $, which is the measure of the void space (fraction of rock volume) 2. Permeability, k, which is the measure of the fluid transmissivity in millidarcies (mD) 3. Fluid saturation, Sw'which is the portion ofthe pore space that is occupied by oil, So, gas, Sg, and interstitial water (fraction) 4. Capillary pressure, Pc' which is the force per unit area resulting from the interaction ofthe fluids with the medium in which they exist in kilopascals (kPa) or pounds per square inch (psia) 5. Electrical conductivity of fluid-saturated rocks
27
b
DETERMINATION OFOIL AND GAS RESERVES
6. Formation volume factor, Bo, which is used to convert subsurface volumes of oil to surface conditions (the conversion is a consideration ofa phase change resulting in the liberation of gas (solution gas) from the oil and the compressibility of reservoir oil)
Oil The calculation of oil in place is based on the following equation:
I N=VRx
Bo
7. Gas compressibility factor, Z, which adjusts for the compressibility characteristics in mixes of natural gas in the conversion of ideal gas volumes to actual volumes
where
Cutoff Values Reservoir rock and fluid properties are used to help determine the thickness of the reservoir rock that contributes oil or gas production based on testing or actual production. Relationships between porosity, horizontal permeability, and water saturation can be developed from core and capillary pressure data to determine cutoff values below which any known economic recovery method will be ineffective, based on present technology. The limiting factor in oil and gas production is the permeability, a measure of the flow characteristics of the reservoir fluids through the rock pores. The permeability to each of the three fluids-oil, gas and water-varies in relation to the content of each of the other fluids in the reservoir. The contribution to production is best measured by the relative permeability of the rock-a flow characteristic of a fluid in the presence ofanother fluid or fluids. For example, the relative permeability of the reservoir rock to oil or gas may be almost nil in the presence of a high saturation of interstitial water, which would render the hydrocarbons immobile. The magnitude of the in-place resource has this limitation from a thickness perspective, being limited to the reservoir rock from which it is possible to recover the hydrocarbons. Hydrocarbons in Place The volumetric calculation of hydrocarbons in place consists of the following steps: I.
Determine the volume of rock containing hydrocarbons from thickness and area considerations or from an isopach map of net pay.
2. Determine the average effective porosity. 3. Determine the volume percentage containing hydrocarbons (from fluid saturations). 4. Correct for the volume of hydrocarbons measured at the surface.
28
(I)
N = oil in place (ml) VR = rock volume (m') = 104 x A x h A = drainage area in hectaries (ha) (I ha = 104 m2) h = average net pay thickness (m)
In Imperial units, the equation is as follows: I N=VRx7758x
Bo
where
(2)
N= oil in place (bbl) (1 acre-foot = 7758 stb) VR = rock volume (acre feet) = A x h A = drainage area (acres) h = average net pay thickness (ft)
Natural Gas The in-place volume of natural gas is adjusted for temperature and pressure in order to measure volumes at standard surface conditions. The compressibility factor adjusts for the compressibility characteristics for different mixtures of natural gas components in changing from reservoir to surface conditions to account for the variance from the Ideal Gas Law. Natural gas resources may be classified as follows: • Solution gas • Associated gas (gas cap) • Nonassociated gas Solution gas is the gas liberated from oil produced from a reservoir. The rate of production of solution gas depends on the rate of oil production, the relative flow characteristics of the reservoir fluids, and the state of depletion of the reservoir.
r
OVERVIEW OFPART TWO
For calculation of initial solution gas in place, Gs ' the folIowing equation is used: G, = N XR,i where
(3)
G, = solution gas in place (m") N = oil in place (m') R,i = gas in solution at Pi (m3/m3) Pi = original reservoir pressure (kPa)
In Imperial units, the solution gas in place is as follows: G, where
=
NXR,i
(4)
G, = solution gas in place (scf) N = oil in place (stb) R,i = gas in solution at Pi (scf/stb) Pi = original reservoir pressure (psia)
Associated gas is the gas associated with an oil reservoir as a gas cap. Most, if not all, of the energy in the gas cap is required for maximum oil recovery, so associated gas reserves usually remain shut in until most of the oil reserves have been produced. Nonassociated gas is gas that is not associated with an oil reservoir. Production is limited only by market availability and contract terms. For the calculation ofnonassociated and gas cap in-place volumes, the folIowing equation is used:
G = VR x ljl x (l-Sw) x where
T
P
" x -'P"xT, Z;
(5)
G= raw gas in place (m') VR = rock volume (rn') = 104 XA x h drainage area (ha) A (I ha = 104 m 2) average net pay thickness (m) h porosity (fraction of pore volume) ljl Sw = water saturation T sc = standard base temperature (OK) (273 + 0c) Psc = standard base pressure (kPa) T f = formation temperature (OK) (273 + 0c) Pi = original reservoir pressure (kPa) Z, = gas compressibility factor at Pi and T f
In Imperial units, the equation is as follows: G = VR x 43,560 x ljl x (l-Sw) x where
T P " x-'P"xT, Z;
G = raw gas in place (scf)
(6)
VR = rock volume (acre feet) = A x h A drainage area (acres) (1 acre = 43,560 square feet) h = average net pay thickness (ft) ljl = porosity (fraction of pore volume) T so = standard base temperature (ORankine) (460 + OF) Pso = standard base pressure (psia) T f = formation temperature (ORankine) (460 + OF) Pi = original reservoir pressure (psia) Zi = gas compressibility factor at Pi and T f The base pressure used varies with the location of the resource, but is related to the pressure ofone atmosphere at some elevation above sea level (e.g., in Alberta, 14.65 psia and in British Columbia, 15.25 psia). The base temperature is normally 15.6°C (60°F). The determination of the compressibility of the gas involves the use of a gas analysis to provide a factor for a particular mix of natural gas. The equations set out in this section give in-place volumes of raw gas expressed at standard surface conditions. Before the gas is delivered to the point of sale, there are losses at the surface due to processing shrinkage, fuel consumption, and metering errors. These losses must be deducted from the raw gas volumes to arrive at the pipeline gas resources. In sweet, dry gas fields, the shrinkage is related only to fuel consumption and line losses. For wet or sour gases, the shrinkage may also be a result of recoveries of related products and an allowance for plant fuel. The shrinkage may be estimated from a representative gas analysis to obtain the content of the related products, and an estimate of the recoveries of each product. Actual shrinkage for a producing field may be obtained from the ratio ofthe saleable pipeline gas to the raw gas delivered to the plant.
Related Products Natural gas liquids may be calculated from the volume percentage of the product based on a representative gas analysis and the gas-in-place volume. The volumes in place ofnatural gas products expressed in standard volumes per volume of raw gas are shown in Table 4.2-1. Sulphur, which may be calculated from the weight percentage, is also shown in Table 4.2-1. The recovery factor assigned to the in-place volumes depends on the method and efficiency of recovery. Actual gas plant statistics are a source of recovery factors for related products from a producing gas field.
29
DETERMINATION OFOILANDGASRESERVES
Table 4.2-1 In-Place Volumes of Related Products
Liquid Volume per Volume of Raw Gas Product
Vol % Product multiplied by SI'
Propane n-Butane i-Butane n-Pentane i-Pentane n-Hexane n-Heptane n-Octane n-Nonane n-Decane
(m31l06m3 ) 36.88 42.22 43.80 48.53 49.02 55.10 61.80 68.59 75.42 82.26
Imperial' (bbVI06 cf) 6.54 7.48 7.77 8.60 8.69 9.77 10.96 12.16 13.38 14.59
SUlphur Weight per Volume of Raw Gas Product Sulphur
4.2.2
Vol % SUlphur multiplied by (tonnesIl06m3 ) 13.60
(It/I0 6cf) 0.377
Material Balance Estimates
Calculation of in-place volumes of hydrocarbons by material balance requires equating the incremental expansion ofthe reservoir fluids upon pressure drop to the reservoir voidage caused by the withdrawal of oil, gas and water, corrected for any fluid influx or injection. The process requires an accurate history of reservoir performance, including volumes ofoil, gas and water produced or injected, and pressure changes. Five to ten percent ofthe oil or gas must have been produced before a reasonably accurate calculation can be made.
4.3
PROCEDURES FOR ESTIMATING IN·PLACE RESOURCES
The calculation of an in-place resource volume of hydrocarbons does not yield an exact answer. The accuracy of each parameter used in the calculation depends on the validity ofits source and the accuracy of its measurement. When all the individual factors in an estimate are combined, the degree of variance can lead to substantial differences in the answers obtained. The
• Standard conditions of pressureand temperature are 101.325 kPa, 15.6°Cfor 81; 14.65 psia, 60°F for Imperial units.
30
uncertainty associated with any estimate of volumes of hydrocarbons in place is handled differently in the two procedures used for the calculation: the deterministic and the probabilistic. The deterministic procedure is the one most commonly used. The best estimate of each parameter is used in the calculation ofreserves. The accuracy ofthe estimates is only as good as the quality and source of measurement of each parameter used in the calculation and will reflect the experience of the professionals in selecting the best estimate for the parameters. After recovery factors have been applied to the in-place estimates, the reserves are classified as "proved," "probable," and "possible" to reflect the degree of uncertainty, in the view of the evaluator, associated with each category. Degree of uncertainty is discussed in detail in Part One. The probabilistic procedure quantifies the uncertainty in the resource estimate by using the evaluator's opinion to describe the range of values that could possibly occur for each variable, and producing relative frequency curves to describe the probability of the values occurring within that range. A combined relative frequency curve is then generated to describe the possible range for the in-place resources and the associated probability of occurrence of each of the volumes within that range. A variety of methods exist to generate the reserves volumes, the most common being the Monte Carlo computer simulation, which uses a computer to iteratively calculate enough in-place values from the variable parameter ranges to construct the in-place frequency distribution. With rapidly expanding computer applications, the probabilistic procedure is gaining popularity in portraying the uncertainties associated with a range ofestimates. However, there are alternative procedures to generate the in-place resource frequency distribution. The alternative presented in Chapter 6 is a "short-cut" that can be performed on a hand-held calculator. It must be stressed that, as in the deterministic, the reliability of the results using any probabilistic procedure is dependent upon the quality of the data and the experience of the evaluator in selecting the range of values for each variable. If properly derived. the probabilistic estimates of resources in place and recoverable reserves should compare closely with the proved. probable, and possible volumes obtained using the deterministic procedure. In order to understand the uncertainty associated with all reserves estimates, the evaluator must have a good appreciation of probability theory and statistical methods. This knowledge is critical when applying
OVERVIEW OFPARTTWO
classifications such as proved, probable, and possible to the values of resources or reserves. Uncertainty in reserve estimates is covered in more detail in Chapters 3,6, and 22.
4.4
SOURCES AND RELIABILITY OF DATA
Reliability of data is covered in various sections of Chapter 5 in the discussions of the individual parameters used in the calculation of in-place volumes, and in detail in Section 5.11, Quality and Reliability of Data and Results. The source of data and the accuracy of measurement are the two key elements in selecting parameters with some confidence. There can be several different sources of data from which a given parameter can be selected. Evaluators are usually faced with some conflicting values from which they must select either their best estimate or a realistic range of values for each parameter. The experience ofthe evaluator in assessing the validity ofthe data derived from each source is critical in explaining the difference and establishing the best value to be used in the calculations. Table 4.4-1 summarizes the sources for each of the variable parameters used directly in volumetric estimates. The source ofeach factor is shown, with a priority of source given for derivation ofthe specific parameter. The priority is valid only if the testing methods and measurements are considered to be adequate. Resource estimates are valid only with the available data and at the time they were prepared. Constant revision is necessary as other sources of data become available.
4.5
INTERRELATIONSHIP OF PARAMETERS
The various parameters used in the volume calculation are interrelated and, despite their sources, must be compatible to one another. For example--as mentioned in the discussion ofcutoff values-porosity, permeability, and water saturation are related through the geometry of the pore spaces in the reservoir rock. Pressure and temperature are both dependent upon the depth ofburial ofthe reservoir rock. The parameters selected must make sense when viewed together. The subject of recovery of hydrocarbons is covered in Part Three, which discusses the derivation of the recovery factor chosen to convert the in-place resources to reserves. Since the selection of recovery factor may be affected by other reservoir parameters that are discussed in Part Two, a few comments are in order here.
Recovery factor may be dealt with independently when adequate values for parameters such as drainage area, net pay thickness, and pore volume can be assessed. When the information available allows only an estimate of gross productive interval (gross pay), or when the area assigned may represent spacing or total pool area rather than effective drainage area, the recovery factor commonly incorporates the allowance for portions of the reservoir that may not contribute to the production in a given well. Allowance for this undrained volume would probably be better accounted for by adjusting the parameters of thickness and area. Competitive operation is another consideration that may affect the recovery assigned to an individual well. Hydrocarbons in the subsurface do not recognize boundaries of area ownership. Where reservoir continuity allows the movement of hydrocarbons across ownership boundaries, factors such as the date that production commenced and the rate of production have a greater influence on recoveries from individual wells than the in-place resource underlying the individual companyowned tract. In such circumstances, a share of pool reserves based on past production and current production rates provides an acceptable method ofestablishing recovery for individual wells. Extrapolation of well-established production decline curves is the most accurate means ofcalculating reserves and establishing recovery factors to be used with volumetric estimates ofin-place volumes. Decline curve estimates, which are dealt with in detail in Chapter 18, may also lead to re-evaluation of other volumetric parameters. Decline curve methods may be used only when there is sufficient production data to define the rate of decline, and when the capacity of a well to produce is actually declining. At times, apparent decline in production may be due to mechanical limitations. Extrapolation of past performance into the future assumes that the forces acting in the reservoir in the past will continue to act in the same fashion in the future.
4.6
USES OF RESOURCE ESTIMATES
Resource-in-place estimates are the starting point for volumetric estimates of reserves. Regular reserve estimates provide most exploration and production companies with a yardstick of their performance. When current inventory is compared to production rates, an indication of the life ofthe current resource is available at any time. Companies also report their reserve inventories to conservation authorities, securities commissions, and shareholders.
31
m
_
DETERMINATION OFOILANDGASRESERVES
Government agencies require reserve reporting to prepare resource inventories of the province or country for the purpose of determining requirements for pipeline construction and establishing a rationale for approving spacing changes, setting allowables, and approving secondary recovery schemes. Evaluations of reserves of oil and gas are used for acquisition and disposition of these assets, borrowing requirements for banking purposes, and illustrating investment returns to investors and joint venture partners. Individual property evaluations (reserves analyses) are
used for purposes such as land sale acquisitions, exploratory drilling operations, development prospects, participation in third-party ventures, and implementation of enhanced recovery schemes. Uses ofestimates ofin-place resources and reserves and evaluations based on these estimates are many and varied; the amount ofdetail required is dependent upon the accuracy required for the particular purpose. The uses of resource estimates are covered in more detail in Chapter 26.
Table 4.4-1 Sources of Data Units Parameter Area
Thickness net pay
Porosity
Symbol
SI
A
hectares
h
metres
decimal
Imperial Order
Source of Data
Requirements
Isopach map net pay
acres
2
Assigned area
3
Spacing units
feet
Sufficient well control, geophysical control, and identification of depositional pattern and type of trapping
}
Establishing relation to drainage and adequately applying average thickness
Core analyses
Representative recovery Applying proper cutoffs
2
Porosity log determination based on log core relationship
Establishing proper core-log relationship Correlation for hole conditions
3
Log combinations
4
Porosity log
S
Other wireline log
6
Geologist's log
fraction
Core analyses
}
Proper identification of lithology or rock matrix
}
Assessment of gross pay only may be possible
}
Assessing weighted average porosity of net pay Varied with lithology or matrix Lithology identification and use of empirical relationships
2
Log analysis based on log core relationship
3
Log combination
4
Single porosity log
}
S
Derived from another well in the same pool or another pool in the same zone
}
Acceptable comparison
(cont'd)
32
r
OVERVIEW OFPART TWO
Table 4.4-1
(cont'dl Units
Parameter Water saturation
Formation volume factor
Symbol
SI
Sw
decimal
Bo
Gas compressibility factor
Z
Formation pressure
Pr
Formation temperature
m 3/sm3
Imperial Order fraction
bbl/stb
Noncontamination of sample
2
Capillary pressure test
Representative samples for testing
3
Log analyses based on core correlation
Adequacy of determination of formation water resistivity, R", from water sample or logs
4
Log analysis using combination logs
Adequacy of determination of R" from water sample or logs
S
Resistivity vs. estimated porosity
Variation of porosity affecting resistivity
6
Cores and/or logs from samepool orname zone
Validity of comparison
7
From correlation with porosity or permeability
Establishment of correlation
1
Oil analysis
Acceptability of sample
2
Comparison to similar gravity crude
Similar reservoir conditions
3
Correlation curves
Validity of correlation
Gas analysis reservoir and pressure
Acceptability of data
Comparison to reservoir at similar depth with similar gas
Validity of comparison
Bottom-hole pressure bomb gauge
Adequate pressure buildup
2
From other wells in pool
Representative of subject well
3
From other pools at same depth
Acceptability of pressure-depth relationship
4
Estimated from depth vs. pressure correlations
Adequacy of correlation
Bottom-hole temperature measurement - bomb
Mechanical operation of equipment
2
Logs
Temperature of mud reflecting formation temperature
3
From other wells in pool
Adequacy of data
4
Other pools at same depth
Validity of particular depth correlation
S
Depth vs. temp correlation
Validity of particular depth correlation
2
Tf
°C
Requirements
Oil base core
dimensionless
kPa
Source of Data
psia
OF
33
DETERMINATION OFOILANDGAS RESERVES
4.7
BACKGROUND AND EXPERIENCE OF EVALUATORS
An evaluator, in estimating oil and gas resources, must play the role of a modem-day Sherlock Holmes. The investigative process-sifting through conflicting evidence, checking the validity of data, selecting the best parameters, putting together the conclusions in terms of an answer, and testing the reasonableness
34
of that answer-is a test in deductive reasoning. The process may be considered partly an art and partly a science. The depth of experience of the evaluators plays a large role in the acceptability oftheir answers. Drawing from many disciplines-geology, geophysics, engineering, petrophysics, and statistics--evaluators require the full background of knowledge in order to arrive at the best answer possible given the available data.
r Chapter 5
ESTIMATION OF VOLUMES OF HYDROCARBONS IN PLACE
5.1
RESERVOIR AREA AND VOLUME
5.1.1
Introduction
The two methods for estimation of original in-place volume of hydrocarbons are volumetric mapping and material balance. During the initial delineation and development of a field, volumetric mapping is the key to estimation, possibly aided in the very early stages by analogous field data. As depletion proceeds and adequate production history becomes available, material balance may represent a practical second method and may eventually become the most accurate procedure. Reasonable confirmation between the two methods can provide assurance that appropriate data and assumptions have been used for each estimate. Certain reservoir factors tend to reduce the applicability of material balance and reinforce the importance of volumetric mapping throughout the life of the field: • Moderate to strong water drive • Low average permeability • Complex internal architecture and poor lateral or vertical continuity Any of these factors may make it difficult to obtain a representative average pool pressure in response to production. The capability of mapping the "container size" as the basis for volumetric estimation is primarily determined by the interrelationship of the geological complexities and the amount, quality, and type of data. Well control, and the spacing of wells compared with the size of the accumulation are usually the most important considerations. Where applicable, the quality, amount, and positioning of seismic data may also be very important. Information on the following is also important to volumetric mapping: • Formation tops from logs and sample data • Cuttings samples
• Core for lithology, environmental analysis and measurement of parameters • Log response and evaluation • Pressure and pressure transients • Fluid composition • Fluid contacts • Test data and more extended production data
5.1.2
Acquisition of Data
The scope of reservoir study and data acquisition starting at field discovery and extending over the life of the pool must meet the technical objectives, but must also realistically reflect the cost and potential benefits. The information collected should meet both short- and long-term requirements. Ifimportant data is not collected when it is available even though it is not yet needed, there are likely to be serious regrets later when the information is no longer available or can only be obtained at prohibitive costs. The very basic data items such as logs and samples for formation tops are acquired rather routinely. Some of the other items are discussed in the following subsections. Seismic Data
Seismic can be a useful tool for mapping, depending on the geological setting and reservoir objectives. Traditional seismic has been used to provide the transit time from reflection horizons to define depth and form of subsurface structures. Seismic technology has advanced tremendously in the last several decades. Digital recording leading to common depth point seismic and the growing computer capabilities for data processing, have been keys to this advance. However, in many instances the depth response is still all that can be extracted from seismic. In a good number of geological settings where seismic quality permits, "stratigraphic seismic" may also be important. With this method, the amplitude response ofthe recorded
35
DETERMINATION OFOILAND GASRESERVES
signal may provide data relative to lithology, porosity and-when the seismic signalis particularlyclear-fluid content of the reservoir horizons. The "new kid" on the seismic block is 3-D seismic, which has quickly gained major importance in many geological settings, particularly as a development tool with great potential for assisting in reserves mapping. The basic response and data provided by 3-D seismic are no different than those obtained from conventional seismic. The difference is in the configuration of a 3-D survey, which is set up to provide a closely spaced grid ofdata points. This grid allows a more continuous threedimensional definition of structuraland other geological variations. If seismic data is applicable to a reservoir, it probably will have been gathered early in the exploration phase. Ongoing interactive review incorporating new wells must continue well into the development process. It also may be appropriate to shoot additional seismic to help resolve problem areas or incorporate new technical advances. The potential benefits from 3-D seismic, for instance, should be carefully considered, usually the sooner the better. Another aspect to consider in incorporating seismic data is that processing capability is continual1y improving. When field reviews are undertaken, reprocessing may offer data improvement without the added cost and possible timing delays of new shooting.
Analogous Fields
Another aspect to be kept in mind is that others may have drilled or be drilling in analogous field settings relative to the field of specific interest. Utilizing this data as it becomes available may help maximize usefulness of limited data sets. Extended Flow Tests
Horror stories are told of significant investment in facilities and equipment for a well that declined precipitously when put on production. If analogous field data indicates a significant risk, an extended flow test might provide insurance against such an occurrence. As the size of the project increases, the risk exposure increases proportionally. The running of extended flow tests must be weighed against certain considerations: • The level of apparent risk • The cost of the test • Whether it is practical, given the properties of the reservoir, to run the test long enough to resolve possible lower size limits • Environmental and conservation concerns Another option is to put a pool on production with minimum investment to allow extended production data to be gathered. If warranted, further development optimization can then be undertaken at minimal risk. 5.1.3
Original Pressure Data
Undisturbed original pressure data can only be obtained before significant production is taken. Unless the reservoir is very small, normal production testing will not be a problem in this regard. Good initial pressures in gas, oil and water columns allow construction of pressuredepth plots for fluid contact definition. A geographic spread of original water phase pressures will assist in determining whether hydrodynamics are important in the region. Pressure Transient Analysis
In recent years high-resolution pressure recorders have provided another possible source of information relating to reservoir limits. This process relies on the interpretation of very subtle pressure changes. Careful design of the procedure in consultation with experts is necessary, as well as care in the acquisition ofdata. One of the difficulties, as with any kind ofreservoir simulation, is that the results are not unique and must be corroborated with other information.
Data Analysis
Data accumulates rapidly during the delineation and early development phase of field development. The volume of data and pace of activity can often lead to a tendency to handle one well at a time and lose some perspective on the big picture. Periodic field-wide reviews of all geological data and related engineering material provide the best chance for optimal solutions. Also, careful management of data accumulation and study scheduling will ensure that holes in data sets are minimized and that the cost-benefit ratio of data acquisition is efficient. Depositional Environments
The recognition of depositional environments and their relationship to reservoir development is basic to petroleum geology. The study of recent deposits as "a key to the past" is a common theme of sedimentary geological training. Outcrops and producing fields provide a record of ancient depositional environments and resulting reservoir patterns. The extensive literature available on these subjects should be searched for analogous field data in any major field study.
36
s
ESTIMATION OFVOLUMES OFHYDROCARBONS INPLACE
Adequate core coverage is required to define environmental concepts in the subsurface. Proper core spacing and intervals depend on the complexity of the patterns of reservoir development. Data gathering must be appropriately resolved in the early stages of delineation and field development. Nearby analogous fields may add to tbe database. Once depositional environments are resolved from core, it may be possible to expand the study into noncored wells by calibration to log response. However, there is always more risk of error when using logs rather than core for environmental interpretation. Primary Porosity and Diagenesis
Primary porosity is retained in sedimentary rocks through deposition, initial burial, and lithification. This type of porosity and the patterns of its occurrence are easily related to depositional environment. Most sandstones and some carbonates are dominated by primary porosity. Subsequent to the formation of primary porosity, sedimentary rock is often subjected to increasing or varying temperature, pressure, depth of burial, and ground water regimes. As a result, minerals may be dissolved or precipitated. Also, the reservoir rock may be fractured. The processes creating tbese changes in the rock fabric and properties are called diagenesis. The diagenetic overprint and the resulting porosity and permeability changes mayor may not be closely related to original depositional features and patterns. Diagenetic porosity development may, in fact, be controlled by something entirely different such as fault and fracture sets or erosional surfaces. Diagenesis and its controls and results must be considered in reservoir mapping, particularly in carbonate rocks. Type of Trap
Petroleum deposits may accumulate in three basic types of traps: Structural Traps, which are formed by rock layers that have been folded or faulted Stratigraphic Traps, which are formed by depositional, diagenetic or erosional processes Hydrodynamic Traps, which are created by moving formation water, buoyancy, and density interaction with a hydrocarbon accumulation
These traps may occur alone or in combinations of differing dominance. Mapping patterns and style depend very much on the types oftrap. Petroleum geology texts, which usually contain extensive detailed material on
traps, may be used for reference. Analogous field data is also very important when considering trapping. The exploration concepts that led to a discovery would have included an interpretation of hydrocarbon source and trapping. This interpretation should be reviewed and refined or revised, if necessary, at an early stage. Most basins or play areas tend to have a limited suite of trap types of economic importance. Trapping should be understood within the limits of available data before detailed reserves mapping proceeds. Reservoir Continuity
Larger scale structural and stratigraphic features are of first-order importance in determining the limits ofa reservoir and the volume of gas or oil in place. Limits may be defined by faults, folds, facies changes, diagenetic boundaries, or erosional surfaces.
It is often unclear in the early stages of exploration and development whether an accumulation of oil or gas is in a single pool ora series of pools in close proximity. The keys to resolution ofthis question may be provided by pressure; pressure-depth plots; gas, oil and water compositional data; and indications from fluid contacts. The degree of internal continuity and homogeneity within a pool is an important geological feature relative to recovery efficiency. Detailed cross sections or fence diagrams are usually necessary to resolve the details of internal reservoir architecture. Fluid Interfaces
Fluid interfaces important in reserves determination include the following: • Gas-oil • Oil-water • Gas-water Pressure-depth plots provide the best technical resolution oftbese interfaces when good quality pressure and fluid density (gradient) data is available above and below the contact. This method defines a contact even in undrilled or untested intervals (Figure 5.1-1). In medium- and coarse-grained reservoirs of high porosity and permeability, the transition from hydrocarbons to water will be sharp and easily defined by well logs. Flow testing will also be conclusive to definition of contacts in this type of reservoir if wells and test intervals are properly located. Capillary effects in the small-diameter pore systems in fine-grained rocks result in long hydrocarbon-water transition zones and considerable difficulty in
37
DETERMINATION OF OIL AND GASRESERVES
1700
~
5.1.4 t- Gas Gradie~t 4.4 kPaim
1800
-;;;-
\
'" .§.
s: 1900
15. Ql
c
'~" o
'E 2000
\
1\ ~
\\
Ql
::>
>::
2028 TVD(m ss) \(ree Water Level
Resolution ofthe "container size and shape" by a map of the hydrocarbon-filled reservoir is the single most important step in volumetric reserves estimation. Since the reservoir is a three-dimensional form, vertical illustrations such as cross sections, fence diagrams, or isometric drawings may also be required to understand pool geometry. Examples of forms requiring vertical diagrams include complex faulting, major unconformities and salt dome intrusives. Once the vertical geometry is better displayed and understood, more accurate maps may be drawn.
Maps for Volumetric Estimation
2100 Water Gradient / 10'18 kPaim 2200 34
36
I~
38
40
Pressure (mPa)
Figure 5.1-1 Pressure-Depth Plot for Free Water Level Determination
resolving the water level. For example, in the Turner Valley Formation gas reservoirs in the Alberta foothills, the change from fully water-saturated zones to irreducible water saturations may occur over an elevation exceeding 100 metres. In this extreme case, accurately defining water levels is difficult using only log or test data. Gas-oil contacts may also be difficult to resolve. Pressure-depth plots offer a technical solution when quality data are available. Flow testing, including wireline repeat formation tester (RFT) data, may be helpful. The neutron and density log combination can be definitive where the contact is located within a drilled continuous porous section. On rare occasions, reservoir character and seismic quality may be sufficient to define fluid contacts by "flat events" on seismic sections.
Hydrodynamic trapping will result in tilted oil-water contacts with a tilt proportional to oil-water fluid density differences and flow velocity. Tilted contacts may not be evident where a very local area is under study, but they become evident on a larger scale. Accurateresolution of this type of contact may be extremely significant to reserves definition. Gas accumulations may also occur in hydrodynamic settings, but the density difference of water and gas is such that measurable tilts on gas-water contacts are unlikely.
38
Mapping
The interplay of structure, fluid contacts, and porous reservoir variations requires at least the combination of a structure map and an area or volume map. In many cases, construction of a series of maps prepared in a logical sequence may be the best technical approach. This could include some or all of the following: Structure Maps, which may be: • Top formation or top porosity, showing location of faults and fluid contacts • Base formation or porosity with limits as above • Fault plane structures Ifboth top and base porosity structure maps are drawn, then a gross pay isopach map can be derived by crosscontouring. Isopach Maps, which are maps of thickness variations of gross or net pay showing reservoir limits controlled by structural form, fluid contacts, depositional features, diagenesis, erosional features, or combinations of these controls. The isopachs of gross and net pay thickness variations are simple geometric depictions of the reservoir form that can be assessed for "geological reasonableness" with some confidence. Porosity-Thickness (
h)* Maps, which may be drawn directly or constructed by drawing maps on the individual parameters and cross-contouring. Porositythickness mapping is particularly important where porosity in the reservoir is variable and average porosity would not approximate the reservoir void space in all areas.
• Porosity is represented by thesymbol "1\>" inthismonograph and in the petroleum industry generally. The thickness of the reservoir is represented by the symbol "h."
ESTIMATION OF VOLUMES OF HYDROCARBONS INPLACE
Hydrocarbon Pore Volume (HPV) Maps, which may be drawn directly or by cross-contouring ljlh with I-S w values." HPV mapping is particularly important when water saturations are variable within the reservoir.
can remain uncertain well into the field development phase. Closely spaced drilling may provide the only method for resolution of reservoir limits in this circumstance.
Where a series of maps is drawn showing interrelated values, cross-contouring is required to ensure that the maps are compatible. If cross-contouring is being done by hand, maps on two separate variables are overlaid and, at each point where contours ofthe two maps cross, a related variable is calculated by the appropriate arithmetic manipulation ofthe individual values. Figure 5.1-2 shows the derivation of a porosity-thickness (ljlh) map on porosity and net pay thickness. The manual process is tedious, but current computer mapping software can handle it readily.
The Choice of Map Types
The use of cross-contouring to combine parameters in a technically rigorous process is warranted when individual parameters have consistent patterns that can be drawn with reasonable accuracy and with greater assurance than the combined value. For example, a ljlh map can be constructed by preparing a map of porosity variations and an isopach map ofnet pay, and then combining them by cross-contouring. The less rigorous alternative is to calculate and plot ljlh values at each well location and construct the map directly from the combined variable. If individual data such as ljl does have a defined trend, it may tend to be lost in this methodology. Reservoir Limits and Wedge Zones
Structure maps based on seismic depth data and available well control are often the first maps constructed on an oil or gas pool. Limits defined by structure and known fluid contacts may then be located. In dipping reservoirs, the area of fluid interfaces, for example, the oil-water interface, produces a wedge area where the geometry must be carefully handled. This wedge area is geographically defined when the structure is mapped on both the top and the base of porosity. Dipping faults may also create wedge areas, and solution of this geometry may require drawing a structure map on the fault plane. When faults are steep, the wedge area may become very small and may be reasonably represented by a median line. In stratigraphic traps, reservoir limits may not be defined by structure maps, evident gradational thinning, or other simple techniques. Seismic amplitude response might be helpful in some cases, but stratigraphic limits ·Water saturation is represented by the symbol "Sw" throughout the monograph.
The final map to choose as a basis for volumetric calculation is a matter oftechnical judgement: a simple productive area map, an isopach map depicting rock volume, a pore thickness (ljlh) map, or a hydrocarbonpore volume (HPV) map. The choice should be based on careful appraisal of the degree of complexity that can be fairly represented with the da!a available. Simple maps such as productive area or gross pay isopach maps represent physical forms that can be readily assessed for realism. Maps that combine parameters are not as easy to relate in detail to physical forms even though they often tend to be dominated by a single variable such as gross pay thickness. Interpretive geological mapping offers the potential of providing the best representation of the reservoir if adequate data is available and the practitioner is thorough. One general rule worth considering even with interpretive mapping is that the simplest interpretation that fits the data and the geological concepts is often the best. Even with a thorough and technically sound interpretation, if there is freedom to vary the reservoir size significantly, interpretation can introduce the risk of significant error. Careful assessment is required to define when this leads from the booking of "proven" to "probable" reserves. In summary, mapping concepts may be reduced to a few simple concepts to consider: I. Assessing specific data available, analogous fields, and geological concepts in order to understand and visualize the feature to be mapped 2. Separately mapping each significant data item that shows a definable pattern of variation 3. Combining maps by cross-contouring where appropriate (Figure 5.1-3 illustrates a series of maps) Mechanically Contoured Maps
Where a large amount of data is available at reasonable spacing, an alternative method of reserves mapping is to use evenly spaced (mechanical) contours. This amounts to linear interpolation between actual well data points. The method may require some interpretation to assign reservoir boundaries, but once this is done the freedom to vary the result becomes limited. For this reason it is often used in unit or joint venture projects
39
DETERMINATION OFOIL AND GAS RESERVES
I / I / / / / / / /
o
/
,,<:)
, <:).
,
/
/
/ /
/ /
/ /
/
/
/
/
/
/
/
/
/
/
/
/
/
_ O.Oe /
/
/
/
/
/
/
I
/
/
I
/
I
/
I
I
'b
I
I I I
\
/ /
<:)'0 <:).
I
I I I
(a) Net Pay Isopach, h (m)
;' /
/ /
(b) Average Porosity, $ (fraction)
(0)
(0)
/ / /
hX$=$h
/ /
/
(1.0)
\/? 7~
(1.0)
0
(0)
/
(0) (2.0) (a)
/ I
I I I I
...
X (1.6)
(.6)
I
I
,/ <:)'0
C).
(2.4) /
/
(1.2)
I
X (.6)
'"i::>
\ (2.4)
/
(c) Overlay h Contour and $ Contours (calculate $h at intersections)
Figure 5.1-2
40
Cross Contouring
(0)
'0
'"i::> 'b
o
(d) Contour Map of $h (m)
ESTIMATION OFVOLUMES OFHYDROCARBONS INPLACE
Gas-water contact (·2603) intersects top reservoir
Gas-water contact (·2603) intersects
base reservoir
I
1 mile
I
I
(a) Structure on Base Porosity (rn ss)
1 mile
I
(b) Structure on Top Porosity (m ss)
0.8
3.0
Note: Gross thickness map was developed by computer cross-contouring structure maps.
I
1 mile
I
(c) Gross Porous Thickness (m)
Figure 5.1-3
Series of Related Maps (zero edge from seismic, computer-contoured) (ZYCOR Software)
41
n
DETERMINATION OF OIL AND GAS RESERVES
where variations in interpretations can lead to disagreement and impasse. The mechanical method of contouring minimizes extension of high contour values into undrilled areas and, in contrast to an interpretive map, may provide conservative reserves volumes. The strength of mechanical contouring is that if done properly it honors the available hard data with minimal interpretation. Its weakness is that unless the patterns are very simple it does a very poor job of representing the geological patterns and reservoir variations. It should be recognized as only a simple approximation for joint venture and reserves assignment. It is not a geological map. An example of an interpretive and mechanically contoured map of the same data is shown in Figure 5.1-4. Computer Mapping
Advanced software is available for computer mapping ofreservoir parameters with a number of contouring options. The computer is very good at handling simple surfaces such as structure maps, but may have problems with complex surfaces and fault discontinuities. The
capability to adequately represent complex forms depends very much on the quantity and spacing of the data being mapped. Since computer mapping uses mathematically defined best-fit surfaces, the result is noninterpretive and tends to be somewhat mechanical. Combining computer mapping on individual values, editing for geological concepts, and cross-contouring the map series can produce a map that is geologically sound. A major benefit of computer mapping is the ability to use cross-contouring techniques and to calculate volumes. Even where hand-drawn interpretive maps are required to capture the geological concepts, it may be appropriate to digitize maps into computer format to use these computational capabilities. Another benefit (curse?) of computer mapping is that it is possible to test a range of different assumptions and analytical approaches. This can be very useful ifprobabilistic reserves estimates are being prepared. Preparing a range of map interpretations can be an onerous task without computer technology.
/I
Yankee
~
1 mile
(a) Offshore bar cut by meandering shale filled channel. Environmental concepts may be assisted by log, core, seismicdata, or nearby analogous fields.
1 mile
(b) Samedata as used in (a) but contoured ignoring environmental concepts. Apparenttrap integrity and volumes are quite different from (a).
Source: AfterWeinmelster, 1989.
Figure 5.1-4
42
Examples of Mechanical and Interpretive Mapping
ESTIMATION OF VOLUMES OF HYDROCARBONS IN PLACE
5.1.5
Refinement of Volumetric Estimates
With time and addition of data in any of the areas discussed, it is reasonable to expect that the uncertainty of volumetric estimates can be narrowed. The best answers are obtained when the maturity ofthe field provides an extensive database, all reasonable sources are incorporated in the solutions and-where discrepancies between sources arise-preconceptions are challenged and either confirmed or revised. On occasion, new technology such as 3-D seismic, wellbore image logs, or other similar advances may supply better answers. Using all of the data sources may require crossing technical discipline boundaries; thus working in multidiscipline teams is a growing trend in many companies.
References Weinmeister, M. 1989. "Calculating Recoverable Gas in Place from Volumetric Data." Shale Shaker, May-Jun. 1989.
43
R
DETERMiNATJONOF OIL AND GAS RESERVES
5.2
THICKNESS
5.2.1
Introduction Next to the areal extent of the reservoir under study, the thickness value referred to in engineering terms as "net pay" is the most variable component of the oil-in-place equation. It is frequently the most poorly defined and misunderstood term in discussions of reserves. The confusion stems mainly from the differences in focus of the two contributing disciplines: geology and reservoir engineering. The geologist is concerned first with mapping the discrete reservoir elements in question irrespective of any real or commercial segregation dictated by gas-oil or oil-water interfaces.
At this stage geoscientists will map "gross reservoir" and "net reservoir." Later, after the bulk reservoir elements have been adequately defined and mapped, economic considerations will come to the forefront as the reservoir engineer asks the geologist to produce a map showing only the outlines of the hydrocarbon accumulation. The terms "gross pay" and "net pay" are used to describe reservoir thickness. Gross pay, referring to the total hydrocarbon-bearing zone, frequently includes intervening nonproductive intervals that may be present in the reservoir (Figure 5.2-1). Net pay refers to the sum of the productive sections of the reservoir and is determined by the application of cutoffs, which are the specified lower limits of core or log data (porosity,
0.0
OPHI
0.6000
0.0
Porosity .-..."""-CNL FOe Porosity
,
-
-e..-_ _
--
--,•
,, \
\
c;:"
-
..... ---- , ,
\
:::
-.----ill \
CAL - caliper
DS - bitsize
GR - gamma ray (APi)
NPHI - neutron porosity
CNL - compensated neutron log
DPHI - densityporosity
ILd - deepinduction resistivity
CSU - cyber service unit
FDG - compensated formation density
urn- medium induction resistivity
SFL - sphericallyfocused laterotoq
Source: Schlumberger of Canada, 1985.
Figure 5.2-1
44
Reservoir Interval Terminology
SP - spontaneous potential
ESTIMATIONOF VOLUMES OFHYDROCARBONS IN PLACE
permeability, and fluid saturations) below which a formation will be unable to achieve or sustain economic production. Cutoffs are determined by using existing production information from the subject or similar formations, and by constructing correlations between production, porosity, permeability, and water saturation and the recoverable reserves requirements. While porosity and water saturation calculations (which are discussed in subsequent sections) are subject to certain inherent errors, none are large enough to change the results by several orders of magnitude. The same is not true for net pay. Net pay is also important in determining the total amount of hydrocarbons in a reservoir so that the total amount of energy in that reservoir can be calculated. Net pay in this context can be much higher than the value used in the oil-in-place equation because here it can include intervals located in transition zones and even below producing oil-water contacts. Another major criterion in determining net pay is the potential oil available for future secondary or tertiary recovery programs. In such programs displaceable net pay may not equate to net pay in a pressure depletion process, particularly in the case ofa very heterogeneous reservoir. Net pay may also be used during the unitization process either as a stand-alone figure in net pay maps or as a guide for development drilling programs. Clearly, the purposes for which net pay calculations will be used will dictate how they should be determined.
5.2.2 Defining Net Pay Logs
Wireline logs of all types have been incorporated into the process of defining net pay. Porosity tools, by their very nature, offer the most universally consistent net pay criteria. Where single porosity tools are utilized to characterize reservoir porosity, the analyst will typically determine the tool reading corresponding to the appropriate lower limit of porosity and draw a vertical line down the log. All reservoir exceeding this lower limit may be integrated to arrive at a value for net pay. Where multiple porosity tools have been run and a more sophisticated solution approach has been employed, cutoff values, typically in the 2 to 4 percent range for most carbonates and 7 to 10 percent for many sandstones, will be applied to the computed data. In this way logs are employed as the primary filter for net pay
because they represent the first available evidence of the productive potential of a well. Beyond the obvious quantitative porosity estimates afforded by neutron, density, and sonic tools, there are the spontaneouspotential (SP), caliper, gamma ray (GR), and microresistivity devices such as the microlog. These provide further qualitative evidence that a zone is capable of fluid production. In heterogeneous reservoirs with thin beds of widely varying quality, some logs may not properly define net pay due to their tendency to average or smooth porosity over larger intervals. This problem is most acute in previously explored areas with a high number of older logs. Core
Full-diameter or wireline-retrieved small-diameter cores offer a further level of definition beyond that accorded by logs alone. Permeability measurements may be matched to porosity to confirm or enhance the selection of the lower level of producibility. It is useful to note that the absolute value of permeability for a given reservoir and reservoir fluid dictates what the equivalent porosity cutoff will be, and not the reverse. Porosity-Permeability Cutoffs
The empirical selection ofporosity cutoffs to determine net hydrocarbon pay is best accomplished for normal oil and gas reservoirs by using core permeabilityporosity cross-plots. Using minimum air permeability values of 1.0 mD (for medium to high gravity oils), 0.5 mD (for wet gas), and 0.1 mD (for dry gas) will yield approximate effective porosity cutoff levels for commercial hydrocarbon production into wellbores. These cutoffs are empirical (i.e., based on testing and actual production) and are a function of many parameters such as fluid viscosity (mobility), rock grain size and pore size (pore geometry), rock cementation and infill, wettability, and capillary pressure properties. Porosity cutoffs usually increase with decreasing pore and grain size as illustrated in Figure 5.2-2. This plot was generated from a large database of actual core data acquired from dozens ofclastic and carbonate reservoirs scattered across the western Canadian sedimentary basin. Exceptions to the cutoffs listed are gas accumulations in the microdarcy range «0.1 mD) and heavy oil in unconsolidated sands. Although sophisticated, large fracture treatments have been employed on wells in the microdarcy range; however, such low-rate gas
45
DETERMINATION OF OIL AND GASRESERVES
when applied to net pay computations, but it is often essential in the evaluation process to estimate even semiquantitatively the effective permeability of the reservoir.
1000 :;;:
~
100
•
0-
~
~ 10
{j
:;;
'" ~
•e
0
~ LL
'" ~
§ ~
-e c e
0
g
Q.
i;::
.s
::;
~
E ~
()
1.0
r» c
~
0
'in
5.2.3
.=
Logs
~'"
()
0.1
o
10
20
30
Core Porosity ('Yo)
CGL::: conglomerate Source: PanCanadtan Petroleum Lid.
Figure 5.2-2 Air Permeability vs. Porosity production is considered to be uneconomic at the present time. The porosity cutoff for commercial primary production of heavy oil from wellbores is estimated to be approximately 27 percent. Air permeability cutoffs should not be used for heavy oil sands because the measurement of air permeability in disturbed and extracted heavy oil sand is quite meaningless. At this porosity level, the sand is becoming poorly cemented and mobile, permitting the heavy viscous oil to move sufficiently for economic production. These oils have the capacity to carry loose sand grains, as well as small amounts of connate water or gas bubbles. This flow mechanism is far different from that ofconventional oil and gas reservoirs.
Flow Tests The ultimate test of the ability of a reservoir to give up fluids is the actual flow test. During the drilling process and just prior to the decision to run casing in a well, an operator has two options available: I. Open-hole/closed-chamber drillstem test (DST) 2. Wireline formation test (WLT) Judicious use of these tests can enhance the reservoir analyst's ability to discriminate between pay and nonpay zones. Approximate values of in situ permeability can be calculated from WLT data, the object being to sample a cross section of the elements of a reservoir unit and project the permeability data to cover the entire reservoir. WLT techniques are at best "quick-look"
46
The open-hole drill stem test option affords the best overall assessment of net pay criteria because, under ideal circumstances, large volumes ofthe reservoir fluid can be recovered and studied in addition to the extensive drawdown and buildup pressure data that is obtained.
Data Acquisition Programs
The earliest methods for using logs to select net pay intervals involved the use ofSP or OR logs. Using curve inflection criteria for determining the top and the base ofeach reservoir unit remains a valid method ifthe stratigraphic unit is a simple clean sandstone-shale sequence with very porous and permeable sandstones present. However, the blanket assumption that all porous and permeable reservoir units are capable of production is dangerous. Bitumen can be present in different forms: a tar mat or solid pyrobitumen. Disseminated shale, pyrite particles, bitumen, or other blocking or cementing materials can seriously impair the capacity of a reservoir to produce hydrocarbons and thereby disqualify it as net pay. When conditions such as these are known to exist or where the reservoir approaches the lower limits of the producing porosity-permeability regime, more sophisticated logging methods must be considered. Here, all the porosity measuring devices may be employed depending on availability, cost constraints and hole conditions. In clastic sequences, the neutron-densitycaliper combination in conjunction with the microlog and a standard induction resistivity device will resolve most net pay situations satisfactorily. In mixed lithology carbonate reservoirs, where gas may be present, additional care must be exercised, particularly in the choice of the proper resistivity device. Where matrix porosity is low and water saturation is at or near irreducible conditions, resistivities can easily exceed 2000 ohm-metres. The choice ofa laterolog over an induction device may be advisable if resistivity is to be used as a net pay discriminator. An additional environmental consideration involves thin bed resolution. Thin beds are defined not only as vertical variations in lithology, but also may include any closely spaced changing petrophysical parameter that makes evaluation difficult. Rapid fluctuations in
ESTIMATION OF VOLUMES OF HYDROCARBONS INPLACE
porosity type, rock texture or pore type may combine to preclude proper evaluation with standard logging methods. Where thin hydrocarbon-bearing laminae are thought to be present, the addition of a mud-gas log to the open-hole logging program is advisable.
wetlability, relative permeability, and sensitivity to completion fluids and methods. In order to determine the appropriate analyses required, the core retrieval and analysis program must be designed so that all coring objectives may be achieved.
Core
A flow chart depicting the process of designing and implementing a core analysis program in net pay determination is shown in Figure 5.2-3. Ofcritical importance is identification of the reservoir properties that must be measured in the laboratory to aid in the determination of net pay. Once the coring objectives have been defined, the operator must design the retrieval and analysis programs in conjunction with the relevant service
Core data are used to supplement and calibrate log data when net pay is being determined. In addition to porosity and permeability, other properties may be measured in the laboratory to determine whether the interval of interest possesses the properties required for inclusion in net pay. These supporting properties include water saturation, electrical properties, capillary pressure,
I Establishment of Coring Objectives I I
rDesign of Core Retrieval Program I I Core Retrieval and Preservation j
I Core Gamma I j
Core Description and Sampiing For Basic Core Analysis
I Basic Core Analysis • Porosity • Permeability • Fluid saturation
j
I
I
I Sampling For Reservoir I Quality Analysis
Sampling ForSupplementary Core Analysis
I
I
Petrologicai Studies and Reservoir QuaiityAssessment
Sample Screening • x-ray methods
I
I
I Supplementary Tests • Electrical properties • Clay swelling • FInes mobilization • Wetlability • Capillary pressure • Relative permeability
I j
I Data Synthesis I I Net Pay Calculations
Figure 5.2-3
I
Flow Chart for a Core Analysis Program
47
rtn-.
_
DETERMINATION OFOIL ANDGASRESERVES
companies. Factors such as core barrel type, drilling fluid and core preservation methods may be important. Once the core has been retrieved, it is shipped to the laboratory for appropriate analyses. Well Testing
A wide variety oftesting services and equipment is available to accomplish the objectives of the reservoir engineer in a safe and efficient manner. If the limitations ofvarious systems are understood, factors such as excessive downhole pressure and temperature, rough borehole conditions, and the presence of highly toxic hydrogen sulphide can be dealt with in advance to arrive at an optimum testing strategy. Service company experience has shown that the presence of those three factors in the Foothills region of western Canada seriously limits the application of open-hole testing. Such limits apply to a lesser degree to the remainder of the basin except where the presence of H2S is suspected. An effective program must start with a clear idea of the priorities given to the following objectives: 1. Reserve definition for either primary or secondary horizons 2. Stimulation treatment design criteria for follow-up completion attempts 3. Gathering of reference data to allow drilling and completion engineers to plan future wells for maximum efficiency by reducing reservoir damage created by the drilling or completion process
5.2.4
Data Interpretation
Net pay has been defined as reservoir rock that meets various quantitative cutoffs such as porosity, effective permeability, and water saturation. The parameters used to distinguish net pay are usually well-defined for the formation and pool or area from a history ofproduction characteristics for the area. For a specific well to be kept for production, it normally must have a net pay thickness sufficient to contain enough hydrocarbon reserves to pay for the well completion plus an acceptable profit. Wells with less net pay than this should be abandoned ifthey are not required for other purposes such as water injector or disposal wells. Porosity
Porosity is the most popular reservoir quality indicator, and this is unfortunate because the same enviromnental and depositional factors that influence porosity also influence permeability. Although increases in permeability are frequently associated with increasing
48
porosity, post-depositional processes in sands such as compaction and cementation can shift the porositypermeability trend line. For example, increasing porosity associated with constant permeability might indicate the presence of more numerous and smaller pores. The concept of mean hydraulic radius is gaining acceptance as a better method to distinguish reservoir or hydraulic units (Amaefule et al., 1988). Mean hydraulic radius distinguishes pore morphological changes that porosity and permeability alone cannot characterize. Water Saturation
Water saturation is the next most frequently employed parameter used by reservoir engineers to describe the quality ofthe reservoir unit being investigated. Clearly, lower water saturations are indicative of better hydrocarbon production potential. Water saturation, or any fluid saturation for that matter, may be affected by a multitude of rock properties (composition, grain size or shape, packing, sorting and cementation); therefore, use ofa single saturation cutoff could have serious implications in rapidly changing rock types. Fluid Contacts and Transition Zones
The identification of the various fluid contacts, the location of the transition zone, and the determination of other petrophysical, geological, and production characteristics are essential for accurate assessment of what constitutes net pay in the wellbore. This data may then be used to estimate reserves, hydrocarbon column heights, productivity, water cut, and production economics. Fluid contacts may be identified using core analysis (capillary pressure), logs, or pressure data. In a reservoir that is thick enough, a plot of formation pressure vs. elevation can yield both formation fluid type and interface location. Several pressure readings in gas, oil and water zones are required. Plotting and connecting points of common slope identifies the fluid types. Extrapolation ofthe lines to points ofintersection yields hydrocarbon-fluid contacts as illustrated in Figure 5.2-4. The analysis of these plots to determine vertical pressure continuity in a single well or horizontal continuity from well to well is not straightforward because permeability barriers can also be present. Where determinable, the most useful values are the freewater level (the water level if no rock material were present), the 100 percent water level (the level to which water rises due to the presence of the rock material and
T
ESTIMATION OFVOLUMES OFHYDROCARBONS INPLACE
. .c
~
15. Q)
- - - - - - - - - - -
Cl
- - - - ::: ::: - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - ------------ - - -
-
-
Gas~~-
~
<;
Oil Water
-
- - - - - - - - - - - - - - - - - - - - - - - - - - - -
- - - - - - - - - - - - - - - - - - -
-
- »<>
-
-:;/
/
/
/ /
/
/ /
/
O~
/
/ /
Source: Computalog Gearhart Ltd., 1990.
Figure 5.2-4
\
\
\
\
-
.r-
/
/
Pressure
Hydrocarbon Fluid Contact Identification from Pressure Gradients
the resultant rock-water capillary forces), the bottom of the transition zone (the same as the lOa percent water level), and the top of the transition zone. Across the transition zone, water saturations will vary from 100 percent at the bottom (lOa percent water level) to irreducible water saturation at the top ofthe transition zone. Due to the varying relative permeabilities across the transition zone as saturations change, the hydrocarbon and water cuts will change from bottom to top, A "no flow" situation is also possible.
5.2.5
•
- -----.- A-/
/
~
Factors Affecting Data Quality
Adverse Borehole Environments The reliability ofthe various net pay parameters, water saturation, porosity, and net pay, when calculated from open hole logs or measured in full diameter cores, is directly related to the knowledge and understanding of the borehole environment from which this data was drawn. A number of factors can influence this reliability level. The principal factor is the physical condition ofthe borehole at the time oflogging. Bottom-hole temperature and pressure can affect the functioning of
every logging tool. Except in extremely hostile (hot and corrosive) environments (as encountered in deep sour gas reservoirs in the Alberta foothills), these two factors are normally manageable and ofminor importance. Other factors that can, and often do, contribute to critical errors are rugosity (roundness or smoothness) ofthe borehole and the depth of invasion of the drilling or coring fluids employed. Most logging service companies employ sophisticated algorithms to correct their porosity tools for hole irregularities, and use electrical devices to minimize the effects of drilling fluid contamination. However, the reservoir analyst must use caution when these corrections have been employed near "the edge ofthe envelope." In cases ofextreme borehole rugosity, for example, density logs become totally unreliable for porosity. Unless other tools that are less affected by rugosity (i.e., the neutron and sonic logs) are available, the use of nearby well control data might be moreadvisablethan porosity data that seems "a bit high." Similarly, a quick scan of the log header on the primary resistivity device for evidence of either anomalously
49 e
= DETERMINATION OF OIL ANDGASRESERVES
high mud weights or fluid loss characteristics is always a worthwhile precaution. Either ofthese conditions may lead to excessive overbalancing and consequent flushing of the reservoir which, in tum, can create thick mudcake buildup and lead to erroneous calculations of water saturation. Determination of net pay thickness is usually not susceptible to direct measurement errors except where directional or slant drilling techniques have been employed. Penetration ofany reservoir at anything less than a right angle to bedding will give erroneously high thickness indications. Routine examination of the geological framework for the area, coupled with due diligence in the area of borehole trajectory, should remove this as a concern in most instances.
Core Representation When core data is being used to assess the net pay interval, it is important to realize that the core may, in fact, not represent the true reservoir interval. The reason for this is that often the entire zone is not cored or core may be lost, and therefore, there may be pay that must also be considered above or below the retrieved interval. Proper sampling is essential if the resulting basic core analysis data is to be representative of the reservoir. Friable unconsolidated sandstones, fractured reservoirs and reservoirs with alternating competent and incompetent layers often are not fully recovered during coring operations. Small (em scale) to large (m scale) intervals may be ground up or washed out, leaving only the competent zones and some rubble. Unfortunately, it is the competent zones that are often tight and, therefore, the core may represent only the poor part of the reservoir. The sampling should be based upon the lithological distribution, porosity and permeability variations within the lithological units and the distribution of hydrocarbons. The samples should be representative of the interval from which they are chosen, with three to four samples being selected per metre. Where possible, sampling intervals and sizes should be uniform in order to minimize statistical errors. In certain intervals, plug samples may be taken rather than full diameter samples, but the latter type of sampling should be used in heterogeneous reservoirs such as those that are fractured, conglomeratic, or vuggy. Core gamma logs are used in the core analysis laboratory to aid in correlation of core depths with log depths and to determine the precise location of missing
core intervals. Occasionally they are also useful in helping to reconstruct the correct depth sequences of misoriented core. Normally a core gamma logger is operated as a "total instrument," measuring all radiation in a certain, wide range of wavelengths. However, spectral components due primarily to potassium, uranium and thorium emissions may also be measured. Methods for using the spectral components to determine clay types, cation exchange capacities, clay volumes, and even to evaluate source rock have been or are being developed. To properly assess the problem of representation, it is first necessary to measure the core and determine the amount ofrecovery vs. the length ofthe interval drilled. Ifthere is missing core, the lost core interval is customarily placed at the bottom ofthe interval. Often this does not represent the true picture. The actual missing interval can be determined by a detailed comparison of the core gamma log to the downhole gamma log.
Formation Heterogeneity Most logging devices respond to particular properties of a formation that are related to the depositional and post-depositional history of the rocks. The search for a better understanding of porosity and permeability distributions in reservoir rocks has inevitably led to the conclusion that geological environments may be recognized from log shapes in correlatable zones. The first clues to the presence of nearby reservoir boundaries or heterogeneities may be derived rapidly and cheaply even when very little physical sample material (cuttings or cores) is available from wells. However, as the multitude of examples in Figure 5.2-5 illustrates, care must be exercised because log shapes are much more characteristic than diagnostic. Log shapes also tend to be more predictable and reliable in clastics than in carbonates. Various logs are useful to calibrate geologic data. Spontaneous potential logs have long been used to infer not only the presence, but the depositional environment of sand bodies and thereby provide an indirect estimate ofareal extent. The gamma ray log will in most cases reflect lithology better than the spontaneous potential log, particularly where high hydrocarbon saturation exists. Acoustic logs can give clues to the presence of unconformities and faulting and may be an early waming that more than one reservoir unit is present. Resistivity logs are often helpful in qualitatively assessing vertical grain size variations. The recent introduction of formation imaging technology, which presents either an acoustic or an electrical image of the rock
50
•
ESTIMATION OF VOLUMES OF HYDROCARBONS IN PLACE
GENETIC SAND UNITS Cut and Fill
Offlap Fill·ln
Alluvial-Deltaic Point Bar
Alluvial Point Bar
Distributary Channel Fill
I
i
Delta-Marine Fringe
Barrier Bar
~
[ CC L[
Slightly Serrate
Smooth Bell
Slightly
Smooth Cylinder
Serrate
Bell
i
Transgression on Unconformity
\
Serrate
Smooth Cylinder
Cut and Onlap
Funnel
~II~
Serrate Funnel
Thin and Resistive
Bell
Thick
20-150ft.
10-150ft.
10 - 300ft.
10-100ft.
20-75ft.
5-20ft.
Form
Linear; m~ be very wi e
Linear
Linear
Blanket
Linear
Blanket
Trend
Parallelto depositional slope
depositional slope
Parallel to
Parallel to
Parallelto
shoreline
deroSitiona, slope, bu variable
AMPLIFIED SAND UNITS Offlap Fill-In
Cut and Fill
Fill-In i
Point Bar Buildup Alluvial Plain or Valley
( Smooth Bell Slightly Serrate Bell Thick Form Trend
5
M
Delta-Marine
Submarine Canyon BUildup
Barrier Bar Buildup
Fringe
Buildup of Graded Beds
Buildup
~
c:
Multiple
Multiple Smooth Funnel
Smooth Cylinder Slightly Serrate Cylinder
50 - 300 ft.
50-tOOOft.
50 - 500 ft.
30 - 300 ft.
Blanket
Linear, butmay be very wide
Fan
Linearto blanket
Parallel to
Normalto shoreline; normal
Parallelto axis of basin
Serrate
Funnel
1000 ft.
Linear to blanket Parallel to depositlonal slope
shoreline
orparallel toaxis ofbasin
Smooth Cylinder Slightly Serrate Cylinder
HYBRID SAND UNITS Systematic I
Progradation of Alluvial Over Delta-Marine Fringe
Serrate Trahsgression Over Della
Serrate Funnel WithThin Resistive Streak Blanket Smooth BellOn Serrate Funnel Blanket Source: After Shell Development Company, 1970.
Figure 5.2-5 Sand Unit Shape Diagram
51
3
DETERMINATION OFOIL AND GAS RESERVES
surrounding the borehole, shows great promise in assisting both the geologist and the reservoir analyst. Image data is particularly helpful in defining the areal extent of the pay zone before pressure transient data becomes available. In summary, the patient analyst has many tools available in the search for clues to the character of reservoir heterogeneity. Every avenue must be explored at this early stage to reduce the uncertainty regarding the most critical parameter in the volumetric equation: drainage area. Tool Resolution Many types of logging tools are utilized in the determination of reservoir parameters and net pay. The vertical resolution of each tool is dependent upon the requirements ofthe particular measurement. The deeper measuring tools, designed to overcome or minimize the effect of the flushed zone, are limited in their vertical resolution. Conversely, tools that are designed for shallow measurements often have superior vertical resolution. Knowledge of the limitations and differences
52
between the various tools and how these differences relate to geological variations will result in the analyst being better able to understand and evaluate the reservoir.
References Amaefule, J.O., Kersey, D.G., Marschall, D.M., Powell, J.D., Valencia, L.E., and Keelan, D.K. 1988. "Reservoir Description: A Practical Synergistic Engineering and Geological Approach Based on Analysis of Core Data." Paper presented at SPE, Houston, TX, Oct. 1988, SPE 18167. Computalog Gearhart Ltd. 1990. "The Selective Formation Tester." Calgary, AB. Schlumberger of Canada. 1985. Open Hole Log Interpretation. Course notes, Calgary, AB. Shell Development Company. 1970. Reservoir Geology ofSand Bodies. Houston, TX.
ESTIMATION OF VOLUMES OFHYDROCARBONS INPLACE
5.3
PERMEABILITY
5.3.1
Introduction
Permeability does not appear in the volumetric equation, but it is difficult to have any meaningful discussion about the concept of volumetrics without addressing this key attribute of all commercial hydrocarbon reservoirs. Permeability is a measure of how easily a single fluid (gas or liquid) will flow through the connected pore spaces when a pressure gradient is applied. The permeability, k, of a reservoir rock is related to the volumetric flow rate, Q, through the rock by means of'D'Arcy's Law: k
LlP
11
LlL
Q=-A-
(I)
where Q = volumetric flow rate (mLls) k = air permeability (mO) 11 = fluid viscosity (cp) A = cross-sectional area (cm-) t>P = pressure differential (atmospheres/em) t>L = unit length (em) This permeability is more properly termed specific (or absolute) permeability: the permeability of a reservoir to a fluid when the fluid fills 100 percent of the pore space. Specific permeability is not usually directly applicable to petroleum reservoirs. Essentially all reservoirs, whether they produce oil or gas, contain at least two components: hydrocarbon and water. Calculations relating to reservoir conditions require effective permeability: the permeability to the fluid of interest at the conditions of interest. Effective permeability may replace specific permeability in Equation (I) when the conditions are specified under which the permeability applies. The main "condition" in this regard is the fluid saturation. For this reason, there is yet another permeability measure termed relative permeability: the effective permeability at the fluid saturation of interest divided by the specific permeability. Relative permeability is mainly a function of fluid saturation, but also depends to varying degrees on other parameters such as saturation history, temperature, pore pressure, overburden pressure, and interfacial tension. Permeability is interpreted from well test data or logs, or is directly measured on core samples in the laboratory.
5.3.2
Permeability from Core
All laboratory methods for determining permeability rely on a measurement or an interpretation of a flow rate
through, and a pressure drop across, a sample of known length and cross-sectional area, for a fluid of known viscosity. This data is then analyzed by means of 0'Arcy's Law. In theory, the nature of the fluid should not be important; however, in practice, the nature ofthe fluid is very important if the rock and fluid interact. The measurement methods for permeability (American Petroleum Institute, 1952), which are currently under review, may be divided into classes based on the sample type (plug or full diameter core), the fluid used (gas or liquid), and the technique (steady or unsteady state conditions). The sample type controls the amount and quality of information that can be obtained. For a plug, only a unidirectional permeability can be measured, while for a full diameter sample, the vertical permeability plus the permeability in any horizontal direction can be determined. Although gas permeabilities are the simplest ones to obtain, they suffer from two major laboratory problems that are only occasionally encountered in the field: slippage flow (Klinkenberg effect) and inertial (Forcheimer) effects. These problems, although theoretically possible, are rarely observed when liquid permeabilities are being measured. Steady and unsteady state techniques may be used for both types of samples and both types of fluids. The gas permeability ofwhole core samples is typically determined and reported in three directions: one vertical and two horizontal. The two horizontal directions are at 90° to each other, but otherwise are not usually oriented in any particular direction. However, ifthe core was oriented when it was originally cut, the horizontal permeabilities can be related to actual directions in the reservoir. Liquid permeability may be measured using the principles ofgas permeability, but the fluid used is brine or oil instead of gas. Except for possible fluid-rock interactions, unsteady state liquid permeability measurements on plugs do not encounter any major problems that would affect reservoir applications. Test procedures are available to evaluate fluid-rock interactions. These tests involve measuring the permeability of a rock as a function of time (investigation of clay swelling) or as a function of flow rate (investigation of"fines" migration). The degree to which the clays, (most commonly smectite) in a sample have adsorbed water can significantly change the size of pore throats, and hence the value of permeability. Even when clays do not swell, they may contribute to fines migration. Mineral debris may become detached from the pore walls and entrained in the moving fluids above a
53
-DETERMINATION OFOILANDGAS RESERVES
certain critical velocity. These particles are then carried along with the flow until they come to pore throats through which they cannot pass. The particles lodge in the pore throats, accumulate, block the throats, and thereby decrease the permeability. Fines migration and clay swelling behaviours are encountered during liquid permeability testing. In gas permeability tests, neither phenomenon is normally observed. However, if clays have been dehydrated during the cleaning of hydrocarbons from the pore system, significant changes in gas permeability may result as the test progresses. The advantages of the steady-state plug liquid permeameter (the apparatus used for permeability measurement) are that the data interpretation is straightforward and liquid permeabilities are more applicable to reservoir calculations than gas permeabilities. However, the apparatus is complicated and relatively expensive and, consequently, the procedure is more difficult than in the case of the gas permeameter. Measurements of liquid permeabilities on whole-core samples are less common because of even higher costs.
5.3.3
Relative Permeability Measurement
Although the concept of relative permeability is very simple, the measurement and interpretation of relative permeability vs. saturation curves are not. There is evidence that relative permeability is a function of many more parameters than fluid saturation. Temperature, flow velocity, saturation history, wettability changes and the mechanical and chemical behaviour of the matrix material may play roles in changing the functional dependence ofrelative permeability on saturation. The best defined of these secondary dependencies is the variation of relative permeability with saturation history; relative permeability curves show hysteresis between drainage processes (wetting phase decreasing) and imbibition processes (wetting phase increasing).
There are currently no industry standard methods for determining relative permeability, and much research is ongoing, but there are two basic methods of obtaining relative permeability data: steady state and unsteady state. For the steady state method and a two-fluid system, the two phases are injected at a certain volumetric ratio until both the pressure drop across the core and the composition of the effluent stabilize. The saturations of the two fluids in the core are then determined. If this experiment is conducted at various volumetric flow ratios, a relative permeability vs. saturation curve may be derived. This method of testing is generally too timeconsuming and expensive. to be practical for many commercial reservoir engineering purposes. The unsteady state method is based on interpreting an immiscible displacement process. For a two-phase system, a core either in the native state (preserved) or restored to the saturation conditions that exist in the reservoir is flooded with one of the phases. Typically the flood phase is water or gas since in the reservoir one or the other ofthese phases usually displaces oil. The flood process to obtain relative permeability data is interpreted by means of a theoretical model or else by computer simulation. It is sometimes claimed that the steady state and unsteady state methods yield the same values of relative permeabilities. Although undoubtedly true under some circumstances, this statement is not generally true. For most cases, relative permeability is known to be a function of saturation history. Because the history of the core is completely different in the two cases, it is reasonable to expect a difference in the resultant relative permeabilities. The unsteady state test would seem to be the more physically realistic in the context of the usual reservoir processes, because all such processes involve one phase displacing another.
References American Petroleum Institute. 1952. "Recommended Practice for Determining Permeability of Porous Media." API RP 27 (3rd ed.), Dallas, TX.
54
•
ESTIMATION OFVOLUMES OFHYDROCARBONS INPLACE
5.4
POROSITY
5.4.1 Introduction Porosity is the fraction ofthe reservoir bulk volume that is filled with fluid or nonmineral matter-in other words, the "storage capacity" of the rock. While various methods for determining porosity by core and log analysis are described in Section 5.2.2, an understanding ofthe many ways pores may be distributed in reservoir rocks is necessary to fully appreciate the concept of porosity. Figure 5 A-I illustrates what is called "cubic packing" of spheres and is one example of the packing of spherical sand grains.
into whatever places they will fit and as the constituent spheres become irregular or nonrounded. The porosity ofrocks, therefore, decreases as the variation in particle size and shape increases. The porosity of competent rocks is also reduced as the amount of cementing material in the matrix increases, since the cementing material tends to bridge the contacting surfaces of mineral particles and line the pore surfaces. In addition to "primary" porosity created by the intergranular spaces in most clastic rocks and some uniformly deposited carbonates such as oolites, "secondary" porosity can result from vugs and fractures that are generally created after deposition. Vugs are those pore spaces that are larger than would be expected from the normal fitting together of the grains that compose the rock framework. They may originate in many ways, and the type of vug implies some features of its geometry and interconnection. Vugs may vary from tubes or planes that traverse the matrix to vesicles isolated from each other. Fractures and fracture porosity result from earth movements that create joints and faults through which fluids may move. Although fractures may only contribute up to I or 2 percent porosity to a reservoir, they will have a significant effect on reservoir permeability. Hydrocarbons have been produced commercially from rocks with porosities as high as 50 percent. Fractured carbonates, such as those in the Foothills belt of western Canada, are prolific, although matrix porosity may be as low as 1.5percent. Some nonproductive rocks also have high porosities. Clays and shales and certain chalky carbonates may have fractional fluid volumes or microporosity greater than 40 percent; yet these rocks are seldom productive. Porosity, therefore, cannot be considered the sole criterion for the determination of reservoir productivity.
5.4.2 Porosity, .p =
Figure 5.4-1
l
L' - (Lid)' (ltd' /6 0.4764
Sources and Acquisition of Data
Core Analysis (1)
L'
Porosity of Cubic-Packed Spheres
Even though porosity is independent of the size of the spheres, the porosity of a uniform sphere system can vary from over 25 percent to nearly 48 percent depending upon the packing geometry. Ifpart ofthe pore space of the model is filled with mineral particles of smaller size than the spheres, porosity is decreased. The porosity continues to decline as ever smaller particles are put
Core analysis has been called the cornerstone upon which formation evaluation rests, as it provides the only directly quantifiable measurement of fundamental reservoir parameters. Measurements are made on full diameter and plug samples obtained from conventional coring devices, and on plug samples obtained by rotary or conventional sidewall coring tools. The appropriate procedures are described in the Recommended Practice for Core Analysis Procedure (American Petroleum Institute, 1960). An overview of the most commonly used methods follows. 55
m
_
DETERMINATION OF OIL AND GASRESERVES
Porosity measurements are made after a sample has been selected and cut to form a right cylinder, and the hydrocarbons have been removed. The method of cleaning and subsequent drying can have an effect on the measurements. Samples are normally cleaned in a vapour phase unit or in a Dean Stark apparatus using toluene as a solvent. For tight, competent samples, a pressure core cleaner may be used. The samples are then dried in an oven to remove the residual toluene. If the samples contain significant amounts ofclays, the samples should be humidity (45 percent) dried or dried in a low temperature oven to minimize dehydration. Excessive dehydration results in porosity values that are too high.
I.
Clean liquids from the rock samples.
2. Measure the mass of each cleaned sample (dry mass). 3. Determine the volume of each sample (bulk volume). 4. Measure the volume of the open space in each sample (pore volume) or the volume of the solid in each sample (grain volume). The remaining properties may be calculated from the measured values of dry mass and any two of the three volumes (bulk, pore or grain). Methods for determining porosity are oftwo basic types: those that yield porosity directly, and those that yield values for grain volume, pore volume or bulk volume independently. Several analytical methods may be employed in the laboratory, as shown in Table 5.4-1. The following are the most commonly recommended ofthese methods:
A group of properties, including pore volume, porosity, bulk volume, bulk density, grain volume, and grain density, are generally determined in the laboratory by means of a single test procedure. Typically, the steps in this procedure are as follows:
Table 5.4-'
Comparison of Techniques of Determining Porosity
Measured Property
Method
Porosity
Summation of Fluids
Poor ±O.69%
Fair ±1.0%
No
Direct
Good ±O.OI cc
Good ±O.OI cc
Gas Expansion
Good ±O.02 cc
Steeping
Grain Volume
Pore Volume
Bulk Volume
Calculated
Accuracy
Sample Size
Sensitivity to Surface Vugs
Sensitivity to Calibration
No
Moderate
No
High
-
Yes
Any
-
Low
Good ±O.02 cc
-
Yes
Any
-
High
Good ±O.OI4 cc
Good ±O.05 cc
No
Yes
Any
Yes
Low
Gas Expansion
Good ±O.OI7 cc
Good ±O.05 cc
No
Yes
Any
No
High
Steeping
Good ±O.014 cc
Good ±O.05 cc
-
Yes
Any
Yes
Low
Mercury Archimedes Bulk Volume
Good ±O.OI4 cc
Good ±O.05 cc
-
Yes
Any
Yes
Low
Caliper
Good ±O.015 cc
Fair ±O.OI5 cc
-
No
Any
No
Low
Precision
Source: Geotechnical Resources Ltd., 1991.
56
Need for Need Measurement of for Noneffective Cleaning Pore Space
ESTIMATION OFVOLUMES OFHYDROCARBONS INPLACE
Gas Expansion Method. This is used for determining grain volume; it is also known as helium porosimetry and the Boyle's Law method.
Except in the presence of gas, the difference between apparent density, Pa, read by the tool and true bulk density, Ph' is trivial.
Mercury Archimedes Method. This method, used to determine bulk volumes, is based on the fact that a nonwetting fluid will not spontaneously invade a sample.
Acoustic logging tools employ one or more transmitters that emit a sound pulse and receivers that record the pulse as it passes them. The acoustic log represents a recording ofthe time required for a compressional wave to traverse one metre of formation. This interval transit time is the reciprocal of the velocity of the wave. Interval transit time, , as illustrated by Equation (4):
Caliper Method. This method is used to determine bulk volume by measuring the length and diameter of a right cylinder sample. Summation-of-Fluids Method. This method is used for quick determination of the porosity of uncleaned sa~ples.
(4)
Log Analysis
Porosity is also obtained from a variety of downhole measuring devices where tool response is a function of the formation porosity, the fluid in the pore space, and the matrix properties. When the fluid and matrix end points are known or can be determined accurately, tool response can often be reliably related to porosity. All three logging devices (acoustic, density, neutron) respond to the characteristics of the reservoir immediately adjacent to the borehole. The depth ofinvestigation is shallow (only a few inches on average) and usually completely within the flushed zone created by invasion of drilling mud filtrate from the wellbore. At present, the density log is the primary porosity log for most reservoir engineering applications. In operation, a radioactive source applied to the borehole wall emits medium energy gamma rays into the rock. As these gamma rays collide with the electrons in the formation, they lose energy, but continue to travel and are counted as an indication of formation density. Density tool response depends on the electron density which, in tum, depends on the density of the rock matrix, the formation porosity and the density, of the fluids filling the pores. For a clean formation of known matrix density, formation bulk density, Ph' is given by Equation (2): (2)
where Ph Pr
= formation bulk density (g/cm") = fluid density (g/cm") = porosity (fraction)
Pma=
Porosity,
<1>,
is therefore given by Equation (3):
= Pm, - Pb Pm, - P,
= interval transit time (us/m) =
=
transit time in the matrix (us/rn) transit time in the fluid (us/m)
Neutron logs respond primarily to the amount of hydrogen in the formation. In clean formations with pores filled by water or oil, the neutron log indicates the amount of liquid-filled porosity present. Rock has essentially negligible hydrogen content and therefore does not contribute to the porosity response. In the operation of the neutron log, high-energy fast neutrons are emitted continuously from a radioactive source in the sonde or tool. These neutrons collide with formation nuclei in a billiard ball fashion and at each collision lose some energy. Within a few microseconds, the neutrons have been slowed down from initial energies of several million electron volts (eV) to thermal velocities around 2.5 eV and proceed to diffuse randomly until captured by the nuclei of atoms such as chlorine, hydrogen or silicon. The capturing nucleus then becomes intensely excited, emitting a high energy gamma ray of capture. Depending on the type oftool, either the capture gamma rays or the neutrons themselves are counted by a detector in the sonde. The counting rate at the detector is inversely proportional to the hydrogen concentration. Therefore, low count rates infer high porosity and vice versa, and this relationship will generally hold true except where gas is present in the region of investigation of the tool. Industry Databases
matrix density (g/cm")
where
(3)
Except in rank wildcat environments, the reservoir analyst should be aware that an important source of reliable data exists in those wells that have already been logged or cored in the vicinity ofthe study well or area. Many governments, as part of the management of
57
$
DETERMINAnON OF OIL AND GASRESERVES
nonrenewable resources, require that data recovered during the drilling and completion of a well be submitted to the managing agency. In Alberta, for example, all activity is reported to the Energy Resources Conservation Board (ERCB), which maintains a core and cuttings storage and examination facility as well as copies of all data derived from the wells (logs, core analyses, special core analyses, well tests, and production histories). The ERCB also maintains a comprehensive database composed of all key reserves criteria for the oil and gas pools in the province.
5.4.3
Analysis of Data
Statistical Techniques for Core Data
Porosity values for each sampled interval, along with related permeability and fluid saturation data are tabulated in a core analysis report (Figure 5.4-2). Typically, the reservoir analyst will group core data measurements into beds or layers that closely approximate the stratification evident on the open-hole logs. The interpretation ofthis data is aided by cross-plots of horizontal permeability vs. porosity (Figure 5.4-3). By comparing core porosities to individual log response, the reservoir analyst can more accurately calibrate the open hole logs over the uncored portion of the interval of interest. Great care must be exercised in the use of core porosity data because many factors can affect the representativeness of this data. In reviewing core analysis reports, the reservoir analyst should ensure that a summary sheet describing all core retrieval and analysis procedures is included (Figure 5.4-4). this information provides the best basis for assessing the quality of core data. Porosity from Logs
Anyone or, more frequently, a combination of all three conventional porosity devices are typically run when a well has reached total depth or when a protective intermediate casing string is to be set prior to drilling deeper. The science and art of interpreting these logs for porosity and fluid saturation is embodied in the term petrophysics. Petrophysics seeks to express the physical and chemical properties of rocks as they pertain to the evaluation of hydrocarbon-bearing layers. Each log has its own unique application. Figure 5.4-5 illustrates the method used for computing porosity from a density log for a clean formation of known matrix density, Pm.' containing a fluid of average density, Pr. The lithology dependence ofthis tool is evident in the fact that a log reading of
2.54 g/cm'' produces porosity values ranging from 6.6 percent for sandstone to 17.5 percent for a dolomite matrix. Because sound travels more slowly in a fluid-filled pore than in solid rock, for each rock type a unique relationship exists that relates the measured transit time to porosity. The industry has adopted the Wylie TimeAverage Equation as the standard for computing porosity from acoustic logs in clean consolidated formations with uniformly distributed small pores. Figure 5.4-6 demonstrates this porosity vs, transit time relationship. For example, a value of216.5 us/m (66 ils/ft) produces three different values for porosity depending on the nature of the matrix mineral. Neutron log porosity readings are computed and recorded directly on the log. These logs record porosity in linear units for a particular lithology. An internal program automatically provides corrections for the varying effects of mud weight, salinity, temperature and hole size variations. Once the appropriate lithology has been determined, porosity can be read directly from the service company chart as illustrated in Figure 5.4-7. Cross-plotting techniques have evolved because use of a single tool to determine porosity is valid only where the lithology is known to consist of a single mineral that is clean and water-filled. In nature, very complex mineral assemblages are the norm. Here, even the nature ofthe pore structure itself can affect tool response. Under these circumstances, data from two or more porosity devices is needed to resolve the response to differing matrix minerals to the presence of gas or light oils, and to the pore geometry. By far the most universally accepted and utilized ofthese is the neutron-density cross-plot. Today it is almost standard practice to run the neutron and density logs in tandem or combination and present porosity from both logs on a compatible porosity scale. This overlay presentation provides the experienced petrophysical analyst with an additional qualitative interpretation of the nature of the porosity and the host lithology and can aid in the detection of gas-bearing zones in the wellbore. In Figure 5.4-8, a reading of 21 percent limestone porosity from the neutron log is cross-plotted against a 15 percent limestone porosity from the density log, defining a point, P, lying between the limestone and dolomite curves. If the lithology is known to be a mixture ofthese two minerals, it is appropriate to proportion the distance on a line connecting equal porosity values on both curves and assume that it represents the
58
=
m
"TI
<0'
en --<
CORE ANALYSIS REPORT
<:
~
~
'" :.,. '" ,
'" --i
-c
"C
(j'
Ol
o
0
~
'=:J" » -< Ol
Sample Number
Depth (m)
Thickness (m)
BELLY RIVER FORMAnON CORE # I 1023.00 m - 1041.00 m 1023.00-1025.41 2.41 0.19 1025.41-1025.60 I 1025.60-1025.80 0.20 2 1025.80-1026.42 0.62 3 1026.42-1027.32 0.90 4
Sample Depth (m)
Sample Length (m)
Permeability kmax (mD)
k".
(mD)
Porosity
k"".
(mD)
(%)
Saturation Oil H,O
(%)
(%)
Grain Density (kg/m' )
Remarks*
5 z
"T1
<
0
r-
c
RECOVERY/CUT: 17.85 m/ 18.00 m 1025.43 1025.67 1026.14 1026.60
0.13 0.12 0.13 0.14
82.94 9.53 5.12 0.12
78.49 8.57 4.82 0.11
8.59 1.89 3.16 <0.01
20.8 13.3 18.0 10.6
12.6 10.4 14.2 TR
11.9 35.8 35.4 68.4
2683 2668 2675 2677
sh FD FD FD FD
s:m en
0
"T1
:I:
-<
0 :0
0 n
»
:0
OJ
IJ)
::3J
0
"C
en Z
Z
'"0
.... ~
-e r-
10 11 12 13 14
20 21 22 23
1029.38-1029.63 1029.63-1030.07 1030.07-1030.47 1030.47-1030.75 1030.75-1031.18 1031.18-1031.39 1031.39-1031. 7 103 1.73~ 1032.00 1032.00-1032.28 Hl32.28-1032.60 1032.60-1032.81 1032.81-1033.64 1033.64-1034.66 1034.66-1035.61 1035.61-1035.80
0.25 0.44 0.40 0.28 0.43 0
1029.52 1029.71 1030.12 1030.55 1030.80
0.13 0.14 0.13 0.12 0.15
57.52 88.48 24.47 25.68 84.63
56.57 84.96 23.38 25.23 74.86
45.97 64 .29 17.80 7.63 3.04
19.7 19.9 23.4 19.4 16.3
11.6 8.4 9.3 11.9 8.2
31.4 24.8 25.2 35.0 39.6
2643 2640 2646 2647 2650
FD FD FD FD FD
0.21 0.83 1.02 0.95 0.19
1032.67 1033.21 1034.32
0.12 0.14 0.15
12.50 0.38 0.55
12.16 0.32 0.53
3.78 0.Q3 0.22
19.6 18.1 18.4
13.9 TR TR
33.0 20.3 30.9
2661 2682 2682
3.38
1.30
20.2
TR
46.6
2668
FD FD FD calc ss FD
1035. 66
0.09
3.44
calc ss MISSING
-c
--<
0
IJ)
v.
»
Source: PanCanadian Petroleum Ltd., PCP Ferrybank 6-23-43-28W4. Date: Nov. 17, 1987, File: 87-GC-422. * FD = full diameter, P = plugged sample, sh ~ shale, calc ss ~ calcareous sandstone. ** Plug permeability-sample not suitable for full diameter measurement.
» ()
m
DETERMINATION OF OIL ANDGASRESERVES
Equation: log (kh) = -2.7496 + 0.2128
Formation: Belly River Depth: 1025.41 m to 1037.12 m 1000
~
/ / / /
I~
100
c .§. >-
:.,a =
10
'E" a. '" iii
/
t
/ /
/ /
I-
/ /
~ ~
~
c 0
N
.;:
'" +
...
...+
+
...
/ / / / /
I-
0
...
T
/
/
+;1"
I-
~
J:
.+. .
i= ~ I-
...+,./ +~
/ /
.1
...
/ /
~
I-
.01
o
/ / /
,
,
18
24
/
6
12
30
Porosity (%j
Source: PanCanadian Petroleum Ltd., PCP Ferrybank 6-23-43-28W4.
Figure 5.4-3
Porosity vs. Horizontal Permeability
Core Intervals 1023.00-1041.00 m
Recovery/Cut 17.85/18.00 m
Coringequipment Coringdiameter Core fluid
Diamond 101 mm Water-base mud
CLEANING Solvent Extraction equipment Extraction time Dryingequipment Drying time Dryingtemperature
Toluene Vapour phase 22 days Convection oven 24 hours 150'C
ANALYSIS Pore volume measured by Grain volume measured by Bulk volume measured by Fluid saturation measured by
Boyle's Law heliumporosimeter Boyle's Law heliumporosimeter Mercury/caliper Retort
Source: PanCanadian Petroleum Ltd., PCP Ferrybank 6-23-43-28W4. Notes: Plugs are I inchdiameterunlessotherwise noted.
Figure 5.4-4 Core Analysis Report: Analytical Summary Sheet
60
Formation Belly River
No. of Boxes 16
40
$ = Pm.' P. Pm.' P,
30
Dolomite
~
~
-e-
.i'w
20
17.5%
e
e
Limestone
10.0%
3
Pma = 2.71 g/cm
~-----------
10
Sandstone
3
Pm. =2.65 g/cm P, =1.0g/cm
O+-L---'C-.L--'--,-2.9
2.7
.-
3
-,-
2.5
2.3 3 Bulk Density, Pb (g/cm )
2.1
--, 1.9
Figure 5.4-5 Porosity from Formation Density Log
50
$= 40 Dolomite
rf 30
Limestone
.".
.6.tma = 143 }.ts/m
---,,r/
./ltma = 156 jls/m
;6
.~
~
20 15.4%
~---------------------~--------------------
10
13.0%
-
Sandstone
Alma = 182 us/rn =161511s/m
I>t,
7.7%
O.J----L~-..L-_,_~-----~------~
100
200
300 IntervalTransit Time, I>t (lls/m)
400
Figure 5.4-6 Porosity from Sonic Log
61
DETERMINATION OF OIL AND GAS RESERVES 40
30
Sandstone
12.6% 10.0%
----------------------10
Dolomite
O+-----'~:::..---_+---_r---___,---o 10 20 30 40
Neutron Index (Apparent Limestone Porosity) Figure 5.4-7
Neutron Porosity Equivalence Curves
volumetric proportion of the two minerals. Therefore, the interval represented at P would be composed of 40 percent dolomite and 60 percent limestone and have a porosity of 18 percent. While knowledge of the matrix constituents is always important, an error in choosing or assuming the matrix pair does not have a great impact on the porosity determined except in very low porosity carbonates. This feature ofthe neutron-density cross-plot, combined with its inherent gas identification properties, makes it the most popular technique.
Correlation of Log and Core Porosity Many reservoir analysts prefer to use core analyses in reservoir studies, particularly where equity determination is a key issue. While computer-processed suites of log data may represent the only continuous source of computed reservoir parameters, it has long been recognized that log-derived values are not absolute numbers. In core-log matching exercises, the objective is to standardize the output results in such a way that differences in results from well to well represent relative changes in reservoir quality. Therefore, it is common practice to use the core data as the reference point and fit log analysis data to it. A paper by Hamilton and Stewart (1983) outlines a step-by-step procedure for conducting this type of analysis.
62
40 30
Limestone 30
Sandstone ~
25
20 ~
C
20
25
" <::
o
00 10 Q)
E
2-
....
o 10
0
5 ' " Dolomite -10
f~AnhYdrite
-20 + - - - - , - - - , - - - - - ; r - - - - - , - - - ; o 10 20 30 40
CNL (Limestone) (%)
CNL = compensated neutron log
Figure 5.4-8
Porosity and Lithology Determination from Neutron-Density Log
ESTIMATIONOF VOLUMES OFHYDROCARBONS IN PLACE
5.4.4
Factors Affecting Data Quality
Preservation of In Situ Conditions The quality ofthe results obtained from core analysis is directly related to the quality ofthe core when it reaches the laboratory. Therefore, in cutting and retrieving the core, precautions must be taken to preserve, as much as possible, the conditions that exist downhole in the reservoir. The cutting and retrieval ofcore to surface results in the removal of overburden pressure, the introduction of dril1ingfines, and some modification ofthe clays, al1 of which can affect porosity measurements.
contribute to reductions of porosity with increasing depth. There is strong evidence of a continuous reduction in porosity with increasing pressure differential applied between the interior and exterior of a sample. The analyst should be aware that in situ porosity will be lower than that measured under atmospheric conditions in the laboratory. Pore volume compressibility tests may be conducted to determine the appropriate reduction factor for the reservoir under study, and this type of measurement is now virtual1y routine.
Reservoir Heterogeneity Shale Content The most important problem that has eluded solution since it was recognized by early logging over 50 years ago is that of shaly sands. The presence of shale or clay minerals in the interstices of sedimentary rocks affects log analysis by moving the resistivity of the porous and permeable zones toward the normal shale resistivity on the log. Shales also impact porosity measuring devices. With densities between 2.4 and 2.7 g/cm-, shales can show up on density logs as having nil to moderate porosity. On acoustic and neutron logs, shales may appear to have moderate to high porosity. In extreme cases the effects on resistivity and porosity logs can cancel out in the computation of water saturation. However, ifthey do not cancel, the analyst may misinterpret or overlook prospective pay zones. The amount of shale must therefore be determined to permit its contribution to be subtracted from the measured parameters. The impact of clays on the results of core analysis is equal1y difficult to resolve. The main obstacle encountered is in distinguishing pore water from nonliquid clay mineral water. In addition to retaining the clay lattice water, the core analyst must be careful to preserve the last few molecular layers of adsorbed water on the clay minerals. Figure 5.4-9 illustrates the complexity that the presence of clay minerals can introduce to the process of porosity determination from either cores or logs.
Rock Compressibility In the assessment of data quality and reliability, it must be remembered that most laboratory porosity determinations are based on information obtained at surface conditions. Rocks are elastic media and can be compressed and decompressed when subjected to the stress and release of overburden pressure. Mineral elasticity, grain movement and, final1y, grain failure al1
The results of sampling with wireline logging tools or core samples can be misrepresentative of the reservoir. The actual volume of reservoir sampled even with well logs is insignificant in comparison to the unsampled reservoir volume and is never statistical1y random. Certain geologic environments such as marine sands can be predictable over distances in the order ofkilometres, while carbonate reservoirs may vary significantly over distances in the order of centimetres. The effects ofreservoir heterogeneity on the quality ofthe data being used to characterize the reservoir can be minimized only by careful geological investigation. With respect to reservoir heterogeneity, three main criteria should be considered: sample homogeneity, the presence of fractures, and sample size. As a basic rule of thumb, the larger the sample, the better it will represent the range of microscopic variations in the rock. Most reservoir rocks, even those that visually appear to be homogeneous, exhibit variations in permeability over relatively smal1 distances. In highly fractured reservoirs, there are real1y two permeabilities of interest: matrix and fracture permeability. To determine the matrix component in such reservoirs, plug samples are used because al1 fractures must be excluded from the samples. In this case, the general rule "the bigger the sample, the better the sample" does not apply. Fracture permeability should be measured on whole core samples. To get representative values, however, the samples should be restressed to overburden conditions. The procedures utilized for fractured reservoirs are also applicable to vuggy carbonate reservoirs.
Measurement Precision and Tool Resolution Anyone who has ever attempted to use wel1 logs and core analysis data to accurately characterize a reservoir knows that even with the wide range of tools available one rarely gets the same answer from each tool.
63
r+
DETERMINATION OFOILANDGAS RESERVES
---VC1ay Petrophysical Qualities
'I'
Clay ,
Effective
I
,
Bound
Dry : Clay Clay :, Water
Water
Total
Free Fluid
•
•
Free Water
,
~~cor.
(After Overburden Correction)
f---
•
NML
Density
Log Measurements
1----------V Clay = volumeof clay NML = nuclear magnetic log
Neutron
Sonic
Source: Schlumberger, 1988. Figure 5.4-9
Impact of Clay on Log and Core Measurements
The sources of errors in logs and core analyses are both random and systematic and are introduced by the implicit limitations imposed on the measuring device by design considerations. Statistical variation in radioactivity measurements is an example ofa random error; improper or degrading calibration in a logging tool or pressure recorder is an example of a systematic or constant error. By far the most serious source of error is introduced by the unavoidable complexity ofthe reservoir rock. What is referred to here is any closely spaced variation in petrophysical parameters. When petroleum engineers are confronted with thinly bedded strata, they must be even more aware of the vertical resolution limitations of the measuring device.
References American Petroleum Institute. 1960. "Recommended Practice for Core Analysis Procedure." API RP 40, Dallas, TX. Geotechnical Resources Ltd. 1991. "Porosity." In The Science and Technology of Core Analysis (2nd ed.). Course notes, Calgary, AB. Hamilton, J.M., and Stewart, J.M. 1983. "Thin Bed Resolution and Other Problems in Matching Log and Core Data." SPWLA 24th Annual Logging Symposium, Calgary, AB. Schlumberger. 1988. "Measuring Porosity, Saturation and Permeability from Cores: An Appreciation of the Difficulties." The Technical Review," Vol. 36, No.4, Oct. 1988. "Material from The Technical Review is printed with the permission of The Oilfield Review.
64 7
ESTIMATION OF VOLUMES OF HYDROCARBONS INPLACE
5.5
HYDROCARBON SATURATION
5.5.1
Introduction
The saturation of a given fluid is defined as the fraction of the pore volume occupied by that fluid. This definition, while simple, provides no insight as to how or where the fluids are held within the porous network of the rock; it merely states that some fraction of the pore network contains the given fluid.
5.5.2
Saturation Determination From Core
The saturations of hydrocarbons (both liquid and gaseous) and water in petroleum reservoirs are two of the most important properties of interest to the reservoir analyst. However, because these fluids are generally mobile, they are not always recovered during conventional coring operations. Therefore, by the time the core is analyzed in the laboratory, the fluid saturations do not necessarily represent those that exist in the reservoir. For this reason, fluid saturations measured by core analysis are generally treated as qualitative numbers rather than precise values. With proper precautions, such as drilling with lease crude and using pressurized or sponge coring techniques, saturation measurements may be made more accurately. However, these techniques add considerable expense to the core retrieval. It should be noted that the inaccuracy ofthe measurements is not due to the laboratory techniques, but to the difficulty in obtaining proper samples. For accurate estimates ofsaturations in a reservoir, both core and geophysical well log data must be used; furthermore, the log data must be interpreted accurately. This means that calibration constants for electrical properties should be measured on core samples. When proper care is taken, reliable saturation values can be obtained from logs. More accurate saturation data may be obtained by using sponge core or oil-base core techniques. With the sponge core technique, core is recovered by means of an aluminum inner core barrel that has a sponge lining. Fluids escaping from the core are absorbed by the sponge. Samples are cut from the core and analyzed for fluid content using the Dean Stark technique. In this process the sample is weighed and placed in the Dean Stark apparatus, and the extraction solvent is boiled and condensed repeatedly. The water-solvent vapour mixture rises and condenses, with the water collecting in a graduated collection tube. Solvent cleans oil out of the sample. The volume of water is measured directly and
the mass of oil originally in the sample is calculated by difference. The sponge corresponding to each sample is similarly analyzed in order to obtain the total fluid content of the core. Oil-Base Coring for Connate Water Saturation With oil-base core, the core is drilled with lease crude or an appropriately designed fluid as a lubricant. The crude will only displace oil and, therefore, it is possible to accurately determine connate water saturations. The recovered core is kept immersed in this fluid until it is ready for analysis in the laboratory. The recovery of an oil-base core and the successful measurement of an average connate water saturation, Swo' requires balancing the need for accurate water saturation data with the realities of conducting a potentially hazardous coring operation with minimal risk and reasonable expense. Careful consideration must be given to the selection of the proper coring fluid to preserve the native wettability in the core. (Wettability is defined in Section 5.5.5.) To determine a reliable connate water saturation, the optimum placement of the core location is, as far as is feasible, above the local oil-water contact. Detailed knowledge ofreservoir pressure permits maximum overbalance reduction to minimize the stripping of connate water during the coring process. When all elements of the operation are carefully controlled, laboratory analysis (Dean Stark) on fulldiameter core samples for connate water saturation compares favourably with other methods such as single well tracer testing and open-hole log evaluation. The economic attraction of such an operation is easily appreciated, considering that reductions in recognized water saturation may approach or even exceed 50 percent and may result in increases ofas much as 20 percent in the perceived original oil in place. Changes of this magnitude can impact not only estimated reserves, but also field development plans and production through increased maximum rate limitations. Saturation Measurement Three general families of techniques are available for the measurement ofsaturations in rocks: chemical, which includes retort and distillation methods; electrical, which includes both laboratory and geophysical log methods; and nonintrusive, which includes X-ray and nuclear magnetic resonance. The chemical techniques are
65
DETERMINATION OF Oil AND GASRESERVES
currently the universal choice for routine core analysis operations, and electrical, for wellbore measurements. The nonintrusive techniques are gaining acceptance as on- line saturation methods for displacement and enhanced oil recovery studies, but are not generally used to determine routine oil and water saturations and will not be discussed further. Chemical Methods The procedure for determining fluid saturations by the retort method is based on taking two companion samples. One is weighed, thoroughly cleaned, and then its porosity determined (porosity sample); the other is crushed, placed in a retort oven, and heated for analysis of its oil and water contents. In the distillation method, the sample is placed in a Dean Stark apparatus with toluene. As the toluene is heated and condensed, fluids are removed from the rock, and the water is captured and measured. Oil values are determined by calculation. Generally, the sum ofthe water and oil saturations does not total one, but is a fraction of the porosity because a gas saturation has developed with the depressuring of the core sample.
ljl = porosity (fraction) m = cementation exponent S; = water saturation (fraction) n = saturation exponent As a consequence ofArchie's work, the exponents m=2 and n=2, and the coefficient a= I are generally used in formation evaluation; "a" is a constant, also used in Equation (3). However actual values of "a," "rn,' and "n" can be determined in the laboratory for any specific
reservoir, At this time, a recommended procedure does not exist for formation factor measurement. Although most laboratories use custom-built apparatus, all have the same basic principles of operation. The sample is capped with mandrels and placed inside a pressure containment cell fitted with electrically insulated end caps. The chamber is pressurized and the sample is then saturated with brine. Ifthe tests are to be performed at reservoir temperature, the pressure containment cell is placed in an oven. The resistivity of the sample is measured and the formation factor, F, is calculated using the following equation:
Electrical Methods
(2)
Because brine is electrically conducting, it seems reasonable to expect the electrical conductivity, or its inverse, the electrical resistivity, to vary with brine saturation. This expectation is the basis of the electrical method of saturation determination. During the 1930s, a large number ofworkers performed tests to determine the relationship between the resistivity of rock samples and the brine content. In general, it was found that correlations existed, but it was not until the comprehensive work ofArchie (1942) was published that these correlations were placed in their modem context. Archie's work was based on GulfCoast sandstones in the porosity range of 10 to 40 percent, saturated with brines of salinity between 10 000 mg/L and lOa 000 mg/L of NaCI. The work covered both fully saturated and partially saturated samples, and presented the classical empirical equation still employed today by petrophysicists and formation evaluation experts:
aR" R, = ljlms: where
66
R, = true formation resistivity (ohm-m) a = constant R,. = formation water resistivity (ohm-m)
(I)
where R, = resistivity of water- saturated formation (ohm-m) The ultimate objective offormation factor measurement is to determine the values of "a" and "m" that characterize a reservoir. For this reason, a suite of samples should be chosen having a range ofporosities that spans the range found in the reservoir. A prerequisite to obtaining representative values of "a" and "m" is a very careful sample selection procedure. Formation factors and porosities (preferably measured under stressed conditions) are determined for this suite of samples. The values for all samples tested are then plotted on log- log paper as illustrated in Figure 5.5·1 and fitted with an equation of the form: log F = log a - m log ljl
(3)
The determination of the "n" exponent in the Archie equation (Equation I), is considerably more complicated than formation factor measurement because it necessitates measurement of not only a resistivity, but also a saturation at each data point. Samples are commonly desaturated by one of two methods: centrifuging, or using a porous diaphragm.
ESTIMATION OF VOLUMES OF HYDROCARBONS IN PLACE
50
,
~, , -.
'.
<,
", <,
'. .
30
Hard 1 <, y F = , $'
"
20
~~ <,
<,
<,
'.
~
>!i. g....
,
". <,
-:
-e~ 10 Ul
eo
,
.... :-:. "-
SOf!(HUm/ F = 0.62 $,.1.
a..
". . '" ~ ".~
c..,(S"-'"
- 1 87 + 0.019 m-.
$
5 ". <,
'.
,
/ . ...., -, ". <,
F = 0.81 $'
'.
~ <,
<,
<,
2
-,
<, <, <, <, <,
<,
"
1 10 Source: Schlumberger, 1972.
Figure 5.5-'
10' Formation Factor, F
10
Porosity vs. Formation Factor
Once a set of saturation-resistivity data has been obtained, the saturation exponent is found by plotting this data in log - log format as illustrated in Figure 5.5-2 and fitting the data with an equation of the form: log I = -n log Sw where I
3
(4)
= formation resistivity index
Capillary Pressure Studies It is usually accepted that hydrocarbons displace water in a reservoir rock during the normal process of accumulation. Because sedimentary rock is usually deposited in a water environment, the pore network must have been originally full of water. To gain a better understanding ofpresent fluid distributions, it is necessary to understand how hydrocarbons displace water to form the hydrocarbon accumulation in the first place.
where the tubes represent pore throats interconnecting individual pores. For a hydrocarbon accumulation to occur, the pore spaces must be continuously interconnected and the capillary pressure of a water-filled pore must be exceeded by the pressure of the encroaching hydrocarbons. This threshold pressure, also referred to as the displacement pressure, determines whether or not hydrocarbons can accumulate in a pore on the microscopic scale or in a particular geologic structure on the macro scale. In the case of a cap rock or reservoir seal, it determines the maximum height a hydrocarbon column can reach before the seal is breached. The density differences between the hydrocarbon and water phases results in a force called buoyancy effect, which is the principal motive force causing oil or gas to migrate upwards through water-saturated rocks in the subsurface.
The pore geometry of sedimentary rocks is frequently described in terms of the "bundle-of-tubes" concept,
67
DETERMINATION OFOIL ANDGASRESERVES
Basal Belly River Ferrybank Alberta
Formation: Field: Province:
Company: PanCanadian Petroleum Ltd. Well: PCP Ferrybank 2-23-43-28 Location: LSD 2-23-43-28W4M \
I
I
!
I
\
R, 1.00 -=-Ro
S~·68
1\
\
\
\
10
-.
\ \
\-
\ ~
~
~ ,\. ~
10" Source: PanCanadian Petroleum ltd.
~
~ 1.0
Brine Saturation (fraction)
Figure 5.5-2 Formation Resistivity Index Opposing this upward force, however, is the capillary pressure of the reservoir which depends on three factors:
1. Radius of the pore throats of the rock 2. Interfacial tension of the two fluids 3. Wettability of the rock Capillary pressure data is generally obtained from small core samples which represent a tiny fraction of the reservoir. In the laboratory, an air-mercury fluid system is often used to represent the reservoir system. Air-brine and oil-brine systems are also used. It is essential for
68
the analyst to combine data from many samples to more appropriately model the reservoir under study. Several methods are available to average capillary pressure curves. A frequently used method is one developed by Heseldin (1973) in which he uses a displaced rectangular hyperbolic function to relate porosity to bulk volume hydrocarbon for varying levels ofpressure and, in tum, relates capillary pressure to water saturation for various levels of porosity. This method has been used successfully in Alberta in the Waterton, Jumping Pound and Virginia Hills fields.
T ESTIMATION OF VOLUMES OFHYDROCARBONS INPLACE
Another method used is one developed by Leverett (1941). This method employs a correlating function commonly called the "J function," which was originally proposed as a means to convert all capillary pressure data to a universal curve. However, experience has shown that significant differences in the correlation of the J function with water saturation occur from formation to formation. The prime use ofcapillary pressure curves is to confirm water saturations in difficult evaluation environments. Other uses include determination of rock characteristics such as average pore throat size, pore throat size distribution and permeability; calculation of depth of free water level or oil-water contact; and determination of the extent of the transition zone. The manipulation ofcapillary pressure curves is fraught with many uncertainties, and only an experienced reservoir engineer or petrophysicist should attempt such an exercise. Accurate knowledge ofthe specific gravities ofthe reservoir fluids, interfacial tension between fluids and rock, and rock wettability is required for translating capillary pressure data into equivalent oil-water or gas-water data. Table 5.5-1 lists commonly used values for wettability, 0, of a water-wet system and interfacial tension, o, in
dynes/em. Table 5.5-1 Wettability and Interfacial Tension System
0
Air-water-solid 0 Air-mercury-solid 1400 Oil-water-solid 00 0
Cos 0
cr
a Cos 0
I -0.766 I
72 480 35
72 -370 35
When all data has been assembled, the process for interpreting water saturation in an oil-water system from air-mercury capillary pressure curves is a four-step process: I.
Determine the capillary pressure - height relationship in the reservoir.
Pc w/ ,
= 0.433h (SGw -
SG,)
(5)
where PC w10 = capillary pressure of the wateroil system (kPa) h = height (m) SG = specific gravity, relative to water 2.
Convert the reservoir water-oil pressure system into the laboratory air-mercury pressure system using the appropriate rock-fluid values and fluid specific gravities.
_
PCa/ Hg - PC w/0
(cr cos 0)'/H. (o cos 0)'/w
where PC olllg
(6)
= air-mercury pressure (kPa)
PCw10 = water-oil pressure (kPa) 3. Calculate PC olllg for any height above the free water level for the selected rock type. (cr cos 0)'/H. Pc,,". = 0.433h (SG w - SG,) (o cos 0),/w
(7)
4. From the air-mercury capillary pressure curves (Figure 5.5-3), read the percentage bulk volume occupied by Hg at that level for the selected rock type and convert it to Sw' or read Sw (wetting phase saturation) directly. For reservoir systems with fluid characteristics similar to the laboratory systems, conversion factors are not required. However, if the characteristics differ, adjustments similar to these steps must be taken.
5.5.3
Saturation Determination From Logs
All water saturation calculations in theoretically shalefree formations assume a homogeneous intergranular pore system. These determinations are made from resistivity logs and are based on some form of Archie's water saturation equation. As with the computation of porosity from the various geophysical logging combinations, the determination of fluid saturation from various resistivity and porosity logs has generated many unique approaches. Nearly all these techniques are derived from the classical Archie equation, and the results are wholly dependent on the accuracy of the basic input parameters: R", F and R,. The analyst usually selects the deep resistivity reading from either the induction or the laterolog device and after correcting it for environmental, borehole, bed thickness and invasion effects, adopts it as true resistivity, R,. Porosity derived from the sonic, the neutron-density, or some combination of log and core coverage will be matched with the appropriate lithologically dependent porosity-formation factor relationship. Finally, R" will be determined either from log calculations, test recovery, or a sample of produced water from a nearby water-bearing zone in the same geological formation. In shale-contaminated reservoirs and in low porosity complex carbonate rocks, Sw can only be accurately calculated by employing the most
69
it
DETERMINATION OFOIL ANDGASRESERVES
Company: PanCanadian Petroleum Limited PCP Ferrybank 2-23-43-28 Well: Location: LSD 2-23-43-28W4M
Formation: Field: Province:
Basal Belly River Ferrybank Alberta
105
14
Air-Mercury Capillary Pressure Curve
12
Air-Mercury Capillary Pressure Curve
I-
10' 10
o,
""x
'"0
a.
8
~
:::::..
~
~
~'"
:::J
'"'"OJ
:::J
6
a,
10'
\
~
a. 4 10 2
o
\
r-;
o
.2 .4 .6 .8 1 Wetting Phase Saturation (fraction of pore volume) Source: PanCanadian Petroleum Ltd. Figure 5.5-3
\
1 1~
1
1~
1~
Bulk Volume Occupied Hg (volume fraction)
Air Brine Capillary Pressure Test
advanced computational routines that in themselves rely heavily on data support from special core analysis studies. The casual analyst is well-advised to seek expert advice in these areas because improper selection of input parameters could lead to solutions that grossly misrepresent true reservoir conditions. Figure 5.5-4 represents a flow diagram of a typical petrophysical evaluation based on saturations determined from electrical resistivity relationships. The resultant water saturation is the fraction ofthe pore volume of the reservoir that is water-filled. That portion not filled with water is assumed to be filled with hydrocarbons.
70
I~ r-...
5.5.4
Flow Test Procedures for Gas and Oil Saturation
Well test analysis has always held great interest and attraction for drilling and reservoir engineers because it offers the potential to assess not only the true saturation condition ofthe formation, but also formation transmissibility. As advances were made in mathematical modelling theory, early field data that was frequently ambiguous became more amenable to resolution. With the advent of very sophisticated electronic pressure gauges, high speed computers and advances in the field ofmathematics, a new frontier has opened. Addition of the pressure-time derivative to log-log type curves now permits the identification of multiple reservoir boundaries and heterogeneities such as fractures and layered formations.
, ESTIMATION OF VOLUMES OFHYDROCARBONS IN PLACE
Rock Type
Formation Fluid Tests ~-------------------------+
---------------------+
k Permeability
Cores. Sidewall Samples. Drill Cuttings
1-0--
---------~
+---
----
Drilling Time
I I I I
------------+1I I I I I I I I I I I I I I I
Natural Radiation
Induced Radiation
I--
----
-
Spontaneous Potential
I--
-----
Quantitative Under Special Circumstances
I
Acoustical Velocity
Quantitative
Porosity
Rw Formation Water Resistivity
R. Water-bearing Formation Resistivity
Produced Water
Electrical Resistivity
------------------------+ Rt True Formation Resistivity
r--
Sw Water Saturation
Source: After Shell Development Company, 1969,
Figure 5.5-4
Log Interpretation Flow Chart
71
It
DETERMINATION OFOIL AND GAS RESERVES
In designing any test, reservoir engineers integrate as much open-hole logging and geological information as possible. Some of the flow regimes that can be recognized during a pressure test include infinite acting, pseudo-steady state, and steady state. It is important that the test be designed to recognize and capture data from all flow regimes. Critical formation properties like permeability and skin factor can be determined only from the infinite acting flow period. Reservoir size and shape can be deduced from the pseudo-steady state phase, and the steady state phase can give clues to that most-soughtafter parameter: drainage volume. Pressure transient tests can be conducted either in the open hole or in perfor. ated casing. The open-hole drillstem test (DST) employs a valve, packer, and pressure gauge. A more sophisticated production logging tool string run in a cased hole can measure temperature, pressure, fluid density, and flow rate in addition to gamma ray activity and borehole diameter. In both cases, the goal is the same: to assess the fluid content and transmissibility of the reservoir as well as the extent of the producing formation away from the wellbore.
5.5.5
Factors Affecting Data Quality
Presence of Shale or Clay Shale- or clay-free environments are rare occurrences in nature. Shale is, in fact, one of the most common constituents of sedimentary rocks. Aside from the negative effect on porosity and permeability, as previously discussed, the unique electrical properties ofthese complex mineral assemblages greatly influence the determination of fluid saturation. Most analysts resort to one oftwo techniques to resolve water saturation in a shaly sandstone environment. The Waxman-Smits relationship (Smits and Waxman, 1968) attempted to relate the resistivity contribution ofthe shale to the cation exchange capacity (CEC) of the shale: .
,
s; BQ,Sw -=--+-I
R,
F*R w
F*
where F* = formation resistivity factor for shaly sand B = equivalent conductance of clay exchange cations (sodium as a function of C; at 25°C (mho ern- meq") Q, = concentration of clay-exchangeable cations per unit pore volume (meq
ml,")
72
(8)
In 1968, continuous measurements of rock CEC in situ were not possible and, for practical purposes, a Dual Water Model was proposed as a solution. In this approach, clay is modelled as consisting of two parts: bound water and clay minerals, with the clay minerals assumed to be electrically inert. The Dual Water Model as applied to shaly formations is illustrated in Figure 5.5-5. Fluids
Solids
Matrix
Matrix
Silt
Dry Clay
Bound Water
Free Water
Effective Porosity
Shaie
Source: Schiumberger, 1987.
Hydrocarbons
Total Porosity
Figure 5.5-5 Dual Water Model The analyst determines R. and R.b and inputs them to any ofa number ofgeneral computer interpretation programs for clastic sequences, such as the schematic of a typical process illustrated in Figure 5.5-6. To evaluate a shaly formation, four parameters must be determined: water conductivity, C; (or R.), conductivity of bound water, C wb (or R.b)' total porosity, li>, and bound Water saturation, Swb' In practice, a cross-plot of neutron and density logs generates acceptable values of li>,. Any of a variety of shale-sensitive measurements, usually the gamma ray, can be the source of Swb'
Presence of Bitumen Bitumen, in either the fluid or solid (pyrobitumen) phase, is observed in significant quantities in many reservoirs in western Canada, particularly in the Devonian carbonates that account for nearly 70 percent of all oil and 20 percent ofall gas produced. When present, pyrobitumen is a major source of uncertainty because of its effects on porosity, permeability, wettability and chemical adsorption, properties that can have a major impact on hydrocarbon recovery processes. On the other hand, bitumen in the liquid phase can be a reserve in itself, as for example, the 50 x 10' m 3 of resources assigned to the Devonian Grosmont Formation ofnorthern Alberta and Saskatchewan. When a reservoir engineer encounters a reservoir with either bitumen or pyrobitumen, careful study and analysis are necessary to adequately gauge the impact that its presence could have on production and production
ESTIMATION OF VOLUMES OF HYDROCARBONS INPLACE
j
I
I
Correlate logs
j
I
Mark permeable beds (SP, ML)
I
I
SP ML
Break beds into zones Induction
Lateroloo
R'h P'h Llt'h Igr $N Ro•
<;
/ Conductive" 2' Resistive " 5'
I
Zones" 2'
I
I
In shale zones, determine average values for R'h P,h Llt'h eneutron.,
zone
= shaliness index = effective porosity = formation water resistivity v; = volume, shale Rt = true resistivity
Z $. Rw
I Determine shale volume usingshale scalar Chart 1 and Chart 2
I
I
Start zone analysis j I I
Read conductivity (or resistivity) and Igr for zone
I
I
Density P'h = 2,65 Yes
No
I
= spontaneous potential = microlog = resistivity shale = bulk density, shale = sonic travel time, shale = gamma ray index = neutron porosity, shale = wet resistivityof undisturbed
j
One of: density, neutron, acoustic
Cross-plot two logs
Correctfor shaliness Llt Chart 3 Chart 4 P neutron Chart 5
1$ from Chart 41 j
Solve shaly sand equation to get R 0.' Z (need $., Rw' Vsh' R'h)
i Solve for Sw (need Roo, a, Z)
I Last zone Yes
No
\ The End Data Required • Resistivity-induction, dual induction, laterolog • Gamma ray • Porosity log(s) - density neutron, acoustic • Water resistivity
Charts Required 1. Shale scalar 2. Relationship for gamma ray vs. percent clay (V'h) 3,4,5. Acoustic, density, neutron response
Figure 5.5-6 Shaly Sand Interpretation Process
73
't
DETERMINATION OF OIL AND GASRESERVES
strategies. Reservoir rocks with organic-based pyrobitumen frequently exhibit strong tendencies to oil wetness, resulting not only in abnormally low calculated connate water saturation, but also in high effective water permeability. And more important to those concerned with reserves and estimating ultimate recovery, these reservoirs frequently suffer "premature" water break-through on waterflood recovery schemes. Reservoir Heterogeneity As noted in the discussion on the use of capillary pressure curves, each plug represents the characteristics of only the rock type present in that tiny sample. It is imperative, therefore, that the reservoir engineer have some appreciation ofthe variability that can be encountered within the total reservoir under study. Each discrete layer is itself susceptible to subtle changes, both vertically and horizontally, that may escape the eye of even the most careful investigator or lie beyond the depth of investigation of any borehole logging devices. While logs and cores provide data that is useful in calculating water or hydrocarbon saturation, logs represent a moving observation point. This running average, when compared to the stationary observation data point derived from core data, can result in a lack of conformity between samples of differing geometrical character. It is important, therefore, that common sense be employed when comparing saturation data derived from differing measurements and differing rock volumes. Good correlation between widely diverse measurements might indicate the presence of a homogeneous reservoir and permit the analyst to employ fairly large-scale approximations ofthe reservoir. Conversely, poor correlation could signal the presence of extreme heterogeneity in the larger reservoir sense.
Wettability Wettability is defined as the tendency of one fluid to spread on or adhere to a solid surface in the presence of other immiscible fluids. Wettability is a major factor controlling the location, flow and distribution of fluids in the reservoir. The wettability of originally water-wet reservoir rock can be altered by the adsorption ofpolar compounds or the deposition of organic material. The wettability of the reservoir can affect the estimation of in-place hydrocarbon volumes as well as estimates of hydrocarbon recovery.
The estimation of hydrocarbons in place is affected because the understanding of fluid saturations, resistivity measurements, capillary pressures, relative permeability and residual saturations is changed when the system varies from being strongly water-wet to strongly oil-wet. Recovery estimates can also be significantly affected because the initial and residual saturations, relative permeability, primary, secondary and tertiary recovery processes are different for the oil-wet and water-wet cases. References Archie, G.E. 1942. "The Electrical Resistivity Log as an Aid in Determining Some Reservoir Characteristics." Trans., AIME, No. 146, pp.54-62. Heseldin, G.M. 1973. "A Method of Averaging Capillary Pressure Curves." Canadian Well Logging Society, Vol. 6, No. I, Dec. 1973, pp. 33-46. Leverett, M.C. 1941. "Capillary Behaviour in Porous Solids." Trans., AIME, Vol. 2, T.P. 1223, pp.152169. Schlumberger. 1972. Log Interpretation Charts. Houston, TX. - - - . 1987. Log Interpretation Principles! Applications. Houston, TX. Shell Development Company. 1969. Petrophysical Engineering. Course notes, Houston, TX. Smits, L.J.M., and Waxman, M.H. 1968. "Electrical Conductivities in Oil-Bearing Shaly Sands." Trans., AIME, Vol. 243, pp. 107-122.
74
ESTIMATIONOF VOLUMES OF HYDROCARBONS IN PLACE
5.6
TESTING AND SAMPLING
5.6.1
Introduction
The flow capability of a well is generally found by measurement of actual production. Two general types of flow tests, the drillstem test and the production test, are often used to measure production rates and obtain flow pressures. In addition to collecting this data, flow tests provide good opportunities to gather samples of produced fluids for further analysis. This section will discuss flow tests, as well as the reasons and procedures for collecting fluid samples.
5.6.2
Drillstem Tests
The drillstem test (DST) is often the first opportunity to observe the flow characteristics and record the pressure of a reservoir. A DST meets three objectives when conducted properly:
1. To obtain a stabilized initial reservoir pressure 2. To obtain an indication of stabilized flow rates 3. To obtain samples of reservoir fluids The majority of wells today are drilled using the rotary drilling technique, which consists of rotating a bit that is fastened to a drill string made up of pieces ofthreaded pipe called the drill stem, and drill collars. The drill collars are heavy pieces ofdrill stem and allow a downward force to be applied to the bit. The bit is rotated by the drill string which, in turn, is rotated at the surface by the drilling rig. Using this rotary drilling technique, the drill hole is deepened until the prospective zone is reached. A DST is conducted by replacing the drill bit with a drillstem test tool, attaching it to the bottom ofthe drill string and lowering it into the hole. The tool consists of one or two sets of isolating packers, a valve for allowing reservoir fluid to flow, and locations where pressure recorders may be placed. A packer is an expandable rubber element that is squeezed up against the hole. When the packers are expanded, or set, the zone of interest is isolated from the fluids trapped between the hole and the pipe (also known as the annulus). Figures 5.6-1 and 5.6-2 illustrate a typical DST tool in unset and set position. It is important to ensure that the packers are set in a zone that will allow a tight seal. Once the packers have been set at the proper depth, the valve inside the tool is opened, allowing reservoir fluids to flow up the drill string to the surface. Produced liquids (oil, condensate, water) are sent to tanks, and gases are generally sent to a flare pit or flare stack. After a set period oftime, the downhole valve is closed, and the reservoir pressure is allowed to increase to
stabilized conditions. The valve in the tool may be opened and closed as often as required once the packers have been set. A typical DST would include a 5-minute preflow, a 30-minute shut-in, a main flow of 60 minutes, and a final shut-in of 90 minutes. The flow rate during a DST is usually measured when reservoir fluid appears at the surface. Gas and liquid rates are easily measured by the service company providing the DST equipment. Flow rates may be estimated in cases where reservoir fluid does not reach the surface by observing the amount ofliquid recovered in the drill string after the test is complete because any fluid that travelled past the valve in the DST tool during the test would be trapped in the drill stem after the valve was closed for the final buildup. Many experienced rig supervisors are able to accurately determine the amount of fluid recovered in the drill stem while retrieving the DST tool. Average flow rates are estimated by dividing the flow times into the volume of liquid recovered. Pressures are recorded by gauges inserted in the DST tool. Drillstem test tools allow the placement of gauges in a variety of locations so pressures can be measured above the DST valve, outside the tool, and below the tool. The most important measurements are those recorded inside the tool itself. Analysis ofthese pressures indicates the hydrostatic head of the mud column and drawdown and buildup' pressures. Pressure gauges are discussed in more detail in Section 5.8. Closed-chamber DSTs are run in much the same manner as regular DSTs, but the fluids are not allowed to flow to the surface. A pressure gauge at the surface records the increase in pressure as fluids enter the drill string. A detailed analysis of the pressures obtained at surface, the pressure measurements recorded downhole, and the liquid recoveries will yield production rates.
5.6.3
Production Tests
Production tests are performed on completed wells; the tests provide the engineer with insight into the production potential of the reservoir. Production tests may be conducted immediately after the well has been completed or after the well has produced for several years. This is an important consideration as reservoir characteristics do change through the life of the reservoir. Parameters such as pressure and flow potential all change as fluid is withdrawn from the pool. The equipment necessary for a production test can vary from well to well. The basic requirements are pressure recorders to continuously measure flowing and buildup pressures, and surface equipment that is able
75
b
DETERMINATION OF OIL AND GASRESERVES
Figure 5.6-1 Drillstem Test Tool (Unset Position)
Figure 5.6-2 Drillstem Test Tool (Set Position)
76 r
ESTIMATION OFVOLUMES OFHYDROCARBONS INPLACE
to accurately measure the flow rates of the well. Generally, pressure recorders are placed downhole close to the producing formation. Pressure recorders are available in various pressure ranges. It is unwise to expose the recorder to more than 75 percent of its maximum range. Flow rates are best measured on single phases, so test separators are used. A two-phase separator separates gases from liquids, while a three-phase separator separates gas, oil, and water. Surface equipment must be sized correctly to ensure that it will not be a bottleneck for the producing stream. In the design, implementation, and analysis of a production test, several factors must be considered: the purpose ofthe test and the data that is required, the reservoir and fluid characteristics, the type of test, the test equipment necessary, and any operating difficulties. The purpose of a production test often depends upon a number of considerations, the first of which is the life of the well. The data needed from a well that has just been completed may be different from the information needed for a well that has been on production for several years. The determination of information such as delivery rate, reservoir damage, drainage area and boundaries, stabilized flow conditions, and the need for formation stimulation treatment must all be factored into the purpose ofthe test. Many tests are conducted to examine the success offormation treatments, or to recover representative reservoir fluid samples. Knowledge of the characteristics of the reservoir and the fluid is important for the design ofa production test. Many tests yield inadequate data because of avoidable problems. Reservoir damage may occur due to high flow rates, or the well may freeze offdue to hydrates. Knowledge of reservoir and fluid characteristics will lead to the collection of data that is as accurate as possible. There are many different types ofproduction tests, each of which will yield important data. Each type may be run alone or in combination with other tests. The following are common types oftests: • Interference • Absolute open flow (AOF) of gas wells • Constant and variable rate • Pressure buildup A well is generally "cleaned up" after a zone has been completed or worked over. This allows completion fluids to be withdrawn from the reservoir to prevent further formation damage. A segregation test helps to determine if one zone is in pressure communication with another. AOF tests can be one of three types:
conventional, modified isochronal, or single point. All three yield information about the AOF potential of a gas zone. Drawdown tests are conducted to determine reservoir characteristics such as damage, permeability and flow potential. Pressure buildup tests yield much the same information as drawdown tests with the addition of stabilized reservoir pressure. Another aspect ofproduction test design deals with the duration of flow rates and any corresponding buildup times. Generally, flow rates should be of sufficient time to allow the flow rate to reach stable conditions. As this time period is usually dependent upon the parameters the test is designed to determine, field experience and rules of thumb are generally used. Typically, buildup times must be at least twice as long as the preceding flow rates. Exceptions to these rules occur in some AOF tests, where the flow time and buildup time are set and are independent of whether the reservoir has reached stabilized conditions. The question of the actual flow rate is usually dependent upon previous production data. It is recommended that the well be flowed at the anticipated delivery pressure, or if the delivery pressure is not known, at 50 to 70 percent of the well's AOF.
5.6.4
Sampling
Collection ofrepresentative samples of reservoir fluids is necessary for many reservoir engineering applications. Gas, oil and condensate samples are needed for compositional analysis and PVT (pressure-volumetemperature) analysis. Water samples yield information relating to the water salinity and solids content. Special care must be taken when sampling, so that samples collected are representative of the fluids found in the reservoir. The inability to gather good samples may compromise many calculations and studies performed at a later date. Two methods are commonly used in obtaining reservoir fluid samples: subsurface sampling and surface recombination sampling. In subsurface sampling, a sample chamber is run to the bottom of a flowing well on wire line. The sample is collected at bottom-hole pressure and then isolated through the closing of the inlet valves. The well is kept flowing during the process to avoid fluid segregation (and thus an unrepresentative sample). The seemingly simple process of obtaining representative samples is easily hampered by the presence ofmore than one phase in the wellbore. If the reservoir is initially undersaturated above the bubble-point pressure, an accurate sample is easily obtained. However, if the reservoir is initially at the bubble point, it is difficult to assess
77
q DETERMINATION OFOIL AND GASRESERVES
whether the oil and gas are being collected in the correct volumetric proportions. Wen conditioning can alleviate this problem. The principal drawback of subsurface sampling is that only small volumes of wen fluids are sampled. Furthermore, it is necessary to take several downhole samples so that saturation pressure can be compared at the same temperature. In surface recombination sampling, separate volumes of oil and gas are taken at separator conditions and recombined to give a composite fluid sample. Sampling points should be chosen in order to provide homogeneous, preferably single-phase, sample mixtures. Surface sampling allows the collection of fluid samples at the operating conditions of the surface production facilities. Samples are usually collected by fining a cylindrical container with valves at both ends. Due to the location of the sampling point, a much larger sample may be obtained. Because the samples are taken over several hours of flow, this method gives a fairly accurate producing gas-oil ratio (GOR). As in subsurface sampling, the well must be conditioned to ensure stability during sampling. If done correctly, both sampling techniques should yield identical samples. Prior to any sampling of fluids, whether at surface or bottom-hole, it is important to consider whether the fluid to be collected represents the reservoir fluid. When fluid is withdrawn from the reservoir, the pressure changes caused by the withdrawal sometimes cause liquid and vapour to separate. If a collected sample contains a disproportionate part of either of the two phases, the subsequent fluid analyses will give erroneous results. To prevent this problem, it is recommended that the wen be conditioned to remove from the sample point any fluid that may compromise the sample. Conditioning is generally accomplished by flowing the wen at low drawdown rates so that any altered fluid is displaced by true reservoir fluid. Well conditioning reduces the amount of free gas present at the wellbore by essentially pushing it back into solution. The first stage of a conditioning program involves producing the well at a low stabilized rate at constant temperature and gasoil ratio. This reduces the free gas saturation below the critical gas saturation for gas flow in the formation. This first stage may take as little as a few hours or as long as several days. Any remaining gas is forced back into solution through pressure buildup (the wen is shut in). The shut-in period is dependent upon the transmissibility of the formation and can last up to 72 hours. If the wen was initially undersaturated, it is flowed at a low rate during the sampling. If the well was initially at
saturation pressure, the samples are taken while the wen is shut in. Well conditioning procedures are given in API Recommended Practice No. 44 (American Petroleum Institute, 1966). In the course of sampling, care must be taken to ensure that the sample containers are properly purged to prevent air contamination.
Gas Samples By regulation in Canada and as good operating practice, gas samples are obtained whenever a drillstem or production test results in flows of gas. Usually, the samples are obtained at the surface from the outlet of a separator at relatively low pressures (200 to 700 kPa range). Steel sample containers are used in the case of sweet gas. Where hydrogen sulphide (HzS) is present in the gas, it is important to use special containers because steel containers will absorb small concentrations ofHzS and thus prevent its detection during laboratory analysis of the sample. If these are undetected, the results could be small concentrations ofHzS, improper design of gas processing facilities, and high costs to effect changes. Determinations for HzS are often made at the wellsite to ensure that any small amounts of HzS are detected. In fact, when there is any doubt, Tutweiler or Gas-Tech measurements should always be made at the wellsite. Gas samples are sometimes obtained in conjunction with PVT sampling at bottom-hole conditions and transferred to special high-pressure containers for transportation and analysis at appropriate laboratories. Conventional analyses usually identify the mole percentages of various hydrocarbon components as well as carbon dioxide, hydrogen sulphide, helium, and nitrogen. The specific gravity and heating content of the gas are also determined. Gas analyses are used in reserves determinations to calculate the compressibility factor of the gas mixture and to estimate the volumes of sales gas, recoveries of natural gas liquids, and processing shrinkages. In Canada, gas analyses can generally be obtained quite readily through public sources and, in particular, through the conservation and regulatory authorities of each province. Water Samples
Formation water samples can be obtained from the recoveries of drillstem, wireline, and production tests, and during routine production operations. Care must be taken to use only analyses of samples that are
78
...
ESTIMATION OF VOLUMES OF HYDROCARBONS INPLACE
uncontaminated by drilling mud filtrate and the various chemicals used during production and treating. In many cases, determining whether the sample is representative of the formation is based on rather subjective judgement. Analyses of oil field water samples usually identify the major constituents and total solids in milligrams per litre or parts per million. Total solids can range from a few hundred to over 200 000 parts per million in Canadian oil field formation waters. Specific gravity and resistivity are also measured. Water analyses are generally used to identify the source of the water or to obtain the resistivity of the water in order to calculate interstitial water saturations from porosity information and electrical well logs. Analytical results are often presented graphically to enable visual comparisons or "fingerprinting" of waters to be made. The Stiffdiagram (Stiff, 1951) is widely used for this purpose. Useful compilations offormation water resistivities are available for the majority of productive reservoirs in the western Canada sedimentary basin and other parts of Canada. One such compilation is published by the Canadian Well Logging Society (1987). The published formation water resistivities represent the best information available at the time of publication, but care must be taken to use the data most appropriate to the specific application.
Oil Samples Conventional Surface Samples
Crude oil samples are obtained and analyzed for a variety of characteristics that are ofimportance in reservoir work, production operations, wellsite treating, pipelining, and refining. This brief discussion is restricted to crude oil samples as they apply to reservoir engineering. A distinction will be made between conventional crude oil samples obtained at the surface and crude oil samples obtained at reservoir conditions in order to measure PVT characteristics in the native state. Conventional surface crude oil samples are generally obtained from crude oil storage tanks, at the wellhead, and from drillstem test recoveries. The American Petroleum Institute has published guidelines that should be followed in obtaining reliable oil samples (American Petroleum Institute, 1966).
PVT Samples
For a better understanding of the physical properties of a reservoir fluid, a PVT study should be performed early in the life of the reservoir to obtain truly representative samples ofthe reservoir fluid. Generally, it is better that PVT studies be performed on subsurface samples. Tests
After a representative sample has been obtained, the following five tests are normally performed to assess the fluid behaviour and properties: Pressure-Volume Test. A pressure-volume (PV) test involves the constant composition expansion of the fluid sample at reservoir temperature. The sample is initially undersaturated (reservoir pressure is greater than bubblepoint pressure). As the pressure is reduced towards the bubble-point pressure, the oil compressibility is identified. The actual bubble-point pressure is also measured. Below the bubble point, the two-phase volume is measured as a function of pressure. Differential Liberation or Vapourization Test. In a differential liberation test, the sample is subjected to an incremental pressure reduction from the bubble point to zero. As the solution gas evolves, it is removed from the system. As a result, the composition of the fluid sample is always changing. This test identifies the relative density of gas, the gas deviation factor, the gas formation volume factor, the relative oil volume factor, and the gas-oil ratio (the gas remaining in solution at a given depletion pressure as compared to the volume of residual oil at stock tank conditions). During this process, the oil density at each pressure increment is determined by mass balance. A quality control check compares the calculated oil density at the depletion pressure (through mass balance) to the measured oil density at this point. Viscosity. Viscosity is measured at reservoir temperature at a series ofpressures above and below the bubble point. Flash Liberation or Separator Tests. In a flash liberation test, the sample is again subjected to a pressure reduction from bubble point to zero. The oil and liberated gas, however, are kept in equilibrium throughout the expansion. This test identifies the formation volume factor and the solution gas-oil ratio at separator conditions. One or more flash liberation tests should be done to determine the behaviour of the reservoir fluid as it passes up the tubing, through the separator(s), and into the stock tank.
79
DETERMINATION OFOILAND GASRESERVES
Compositional Analysis. Most reservoir fluid parameters can be estimated fromcompositional analysis. In general, the more fluid parameters sought, the more detailedthe analysis must be. A typical compositional analysis includes a separation of components through C10 as a minimum. More sophisticated equationsof statemayrequire analysis through C30orhigher. It is important to note that due to the nature of the expansion, the flashand differential liberation processes yield different vapour-liquid splits. The degree of difference depends mainly on the composition of the initialsystem. In general, in low volatilityoils in which the solution gas consists mainly of methane and ethane, the resulting oil volumes for either form of expansion are essentially the same.For highervolatility oils,which contain a relatively high proportion of intermediates, the resulting oil volumes can be significantly different.
80
In an undersaturated oil reservoir,depletionbegins as a flash processand eventuallybecomes a combinationof flash and differential liberation processes. Because of this, care mustbe taken to ensure that the correct data is being used in engineering calculation. Flash data must be adjusted using differential liberation volumes to reflect the variouspressureregimesinthe reservoirduring depletion.
References American Petroleum Institute. 1966. "Sampling Petroleum Reservoir Fluids." API RP 44, Washington, DC. Canadian Well Logging Society. C,J. Struyk (ed.). 1987. Formation Water Resistivities ofCanada.
Sep. 1987, Calgary, AB. Stiff, H.A., Jr. 1951. "The Interpretation of Chemical Water Analysis by Means of Patterns." JPT., Vol. 192,pp. 15-17.
ESTIMATION OF VOLUMES OFHYDROCARBONS INPLACE
5.7
RESERVOIR TEMPERATURE
5.7.1
Introduction
Reservoir temperature is of prime importance in the determination of in-place volumes and recovery factors for gas and oil. In estimating gas reserves, a knowledge of temperature is necessary to calculate the gas compressibility factor and gas formation volume factor. T.o estimate oil reserves, knowledge of the temperature IS critical if laboratory PVT data is to be measured under reservoir conditions. Temperature also affects other parameters such as oil viscosity and miscibility, and thereby impacts reservoir engineering estimates of OIl recovery. Often values ofreservoir temperature are estimated from data in the literature or from readings obtained during logging or testing operations. Such data may be acceptable under initial conditions, but should always be confirmed or adjusted using more reliable data as it becomes available. The most reliable source of temperature data is a bottom-hole temperature (BHT) measurement taken with a continuous recording subsurface temperature gauge under stabilized bottom-hole conditions. Other methods, such as using maximum reading thermometers during testing or logging operations, are considered less reliable. Although temperature is usually a function of depth, a number of other factors affect temperature as well. Isotherms at depth may not always follow surface topography. This section describes various techniques used for measuring or estimating BHT and points out the shortcomings in some ofthe values obtained.
5.7.2
after the casing is cemented could be affected by the heat released in the setting reaction of the cement. The most representative BHT data is probably obtained following the completion of the well after it has been shut in long enough for temperature equilibrium to be established between the wellbore and the formation. An ideal time to measure BHT is during a static bottomhole pressure survey, or pressure buildup following a flow test, or during the process ofbottom-hole sampling following the completion ofthe well. In some instances, it is a good practice to run a temperature-depth profile on each producing well using a continuous recording thermometer. In the early stages of development and production of a field or reservoir, measured temperature data may be too sparse to provide a reliable estimate ofinitial conditions. In this case, regional correlations may be helpful. The following list provides correlations for estimating formation temperature for several regions in North America [T, = formation temperature in °C (OP); D = depth in m (ft)]: Alberta (average)
=
T,
=
1.7 + 0.0366D
= 35.0 + 0.0201D) 0.0 + 0.0341D
Alberta Bashaw (carbonale complex)
(T f
Alberta Rimbey-Meadowbrook (carbonate complex)
T, = 9.4 + 0.0304D (T, = 49.0 + 0.0167D)
AlbertaWindfall-SwanHills (carbonatecomplex)
T, (T,
Louisiana Gulf Coast (hydropressure zone)
= 32.0 + 0.0187D)
=
0.0 + 0.0352D = 32.0 + 0.0193D) Tf = 23.3 + 0.0228D (T, = 74.0 + 0.0125D)
North Texas
T, = 15.6+ 0.0306D (T, = 60.0 +0.01675D)
OklahomaAnadarko Basin
T f = 18.9+ 0.0202D (T, = 66.0+ O.OlllD)
Oklahomadeep Anadarko Basin (below 21,000 ft)
Tf (T,
Data Sources
Temperature measurements are made in conjunction with a number of operations conducted on a well. Many of these measurements will have varying degrees of accuracy. Measurements taken while the well is being drilled will likely be influenced by the cooling effect of the circulated drilling mud and will be only approximate. During open hole logging, errors may occur in BHT measurements unless sufficient time is allowed for the wellbore to reach temperature equilibrium with the formation. Measurements taken during flow tests could be detrimentally affected by the cooling effect created by gas expansion when fluids enter the wellbore or flow through any mechanical restriction in the wellbore such as a bottom-hole choke, mandrel or flow nipple. Temperatures recorded on logs run immediately
Tf (T,
= =
18.9+ 0.0255D 66.0+ 0.014D)
In general, formation temperature in the hydropressure zone may also be estimated from thermal gradient maps published by the United States Geological Survey ~nd the American Association of Petroleum Geologists (Cronquist, 1990), using the equation:
T,= Tsa + goD
(I)
where T, = formation temperature, °C COP) Tsa = average surface temperature, °C (OP) geothermal gradient, °C/m (Op/ft)
=
depth, m (ft)
81
7
--, DETERMINATION OFOIL AND GAS RESERVES
Reservoir temperatures obtained using these correlations should be considered preliminary, and they are not a substitute for actual measurements. In Alberta, temperatures measured at a mean depth for each oil and gas pool are shown in the annual reserves report published by the Energy Resources Conservation Board (1991). Reservoir temperature is considered to be constant over the life ofa reservoir, and most reservoir processes, with the exception of in situ combustion and steam or water injection, are considered to be isothermal. Waterflooding can cause significant cooling. In some of the West Pembina Nisku reefs in Albertawhere pools were converted to hydrocarbon miscible flooding after many years on waterflood, reverse temperature gradients were still noted years after the pools had been converted.
5.7.3
Data Analysis
It is pertinent to give some thought to the means of arriving at a value for reservoir temperature. The term bottom-hole temperature or sand-face temperature is applied to the temperature opposite the producing horizon. The logical place to record a single representative measurement would be at the centre of the producing interval or at the pool datum depth. Frequently, measuring tools cannot be run to the desired depth, and therefore the temperature must be extrapolated. For this reason, an accurate determination of the temperature gradient should be established at the run depth. This can be done most conveniently while running a pressure bomb to measure the static bottom-hole pressure. It is likely that temperature would be extrapolated to the same datum as pressure.
If it were desired to estimate BHT from open-hole well logs, some adjustment to the recorded temperatures might be necessary (Deming and Chapman, 1988). Following the cessation of drilling, the usual practice is to condition the borehole by circulating drilling mud throughout the hole for a period of time known as the circulation time. Because the temperature of the drilling mud is usually lower than the undisturbed formation temperature at the bottom of the well, temperature in the wallrock drops during mud circulation. When circulation of drilling fluid is stopped, the well is "shut in," and the temperature in the borehole rises. It is during this period of time, usually 4 to 30 hours after the end of circulation, that the well is logged and the BHT measured at a shut-in time, which is the time elapsed since circulation ceased. Thus the BHT measured is higher than the temperature of the circulating mud, but lower than the true equilibrium or formation
,
temperature. To estimate true formation temperature from raw BHT data, a correction must be applied. The corrections may be made using a Horner plot method. This method owes its name to the fact that it is identical to the equation developed by Homer to predict reservoir pressure recovery. In this method, BHT = T_ + A In [(t + tcire)/t]
(2)
where T_ = undisturbed formation temperature A the negative slope of the Horner straight line (an unknown constant) t shut-in time teire -- circulation time Thus if two or more BHTs-measured at the same depth in the same well, but at different times-are known, the equilibrium temperature may be estimated. The Horner method was used by Deming and Chapman (1988) to analyze BHT data gathered from microfilm copies of log headers on file at the Utah Oil and Gas Commission. Figure 5.7-1 shows 18 Horner plots for BHT data from oil and gas fields in the Utah-Wyoming thrust belt. Although the quality of these data is comparatively high, some scatter about the Horner line is inevitable.These plots show that successivelogging runs generally yield a series of temperatures that are consistent with the Horner model of conductive heat transfer into the borehole during shut-in.
5.7.4
Data Analysis on a Regional Basis
Recently published technical data provides good insight on variations in BHT in western Canada (Lam and Jones, 1984; Lam et aI., 1985). One of the key areas ofinterest has been southern Alberta where a high density of wells provides an opportunity to measure and explain temperature variations from one region to another. Figure 5.7-2 shows the main topographic features of southern Alberta with respect to the eastern limit ofthe disturbed belt. As might be expected, variations in the temperature gradient from the calculated mean value (referred to as "spread") are more frequent in the vicinity of the disturbed belt. Figure 5.7-3 shows these spreads. The spread values vary from a low of 2°C in the plains area of southern Alberta to 10° - 13°C in areas near the edge of the disturbed belt. These "spread anomalies" occur between Hinton and Edson, to the southeast of Hinton and Edson, and south of Calgary. One notable feature on this map is the coincidence of the high-spread area south of Calgary and a fault zone as indicated on an ERCB Paleozoic surface map. In addition, relatively high spread values occur in the Medicine Hat area to the east (as indicated by light
82
s
ESTIMATION OF VOLUMES OF HYDROCARBONS IN PLACE
Shut-in Time, t (h) 50 30
140
20
15
10
5
Shut-in Time, t (h) 20 15 10
50 30
140
5
130
130 Wells
120
AR34-02 5657m
G
~ 110
~ 110 Q)
CC 846 AI 5011 m
~
::>
li! Q)
100
ARE 28 - 06 4595m
0.
E
~
90
ARE 20 - 16 4145 m
Q)
"0 ::c 80
,
E a
'6
•
120
G
AR34-02 3544m
•
70 AR 3-2 2737m
CO
60
ARE 30 -14 3305m AR 10-1 2367m
50 ARE30-14 1462 m
40 .1
.2
.3
C
.4
Wells
~
::>
•
li! 100
.5
458 F2 4700m
Q)
0.
E
ARE36-14 4540m
90
~
45801 4011 m
Q)
"0 80 ::c
E a '6CO
ARE 28 - 01 4221 m ARE 28 - 01 2996m
70 AR 4-1 2332m
60
•
50
CC846 BI 2186 m AR 34 - 2 1933m IRD #1 1688m
40 .1
.6
+tel") In t
(a)
•
.2
.3
C
.4
.5
.6
e;" ) + tIn t
(b)
Source: After Deming andChapman, 1988.
Figure 5.7-1 Representative Horner Plots from Wells in the Utah-Wyoming Thrust Belt
'M~-~'- Lake
Elevation Above Sea Level (tt)
12000 6000
1 .,....... ...•••....••.... ••••••... ••
3000 4000 2000 1000
o
Source: AfterLam at al., 1985.
Figure 5.7-2 Relief Map for Southern Alberta
I Faull From ERCSPaleozoic \ surface map
'C
1Il~~ •.
c::J
6
oI
Source: After Lamat at., 1985.
Figure 5.7-3 Contour Plot of Spread for BHT Values in Southern Alberta
83
DETERMINATION OF OIL AND GAS RESERVES
shading in Figure 5.7-3), but these are surrounded by ~L'nerally low spread values. It has been observed that isothermal surfaces do not alwuys tallow topographic surfaces for a number of reaS,HIS: a rapidlyvaryingtopographic surface and smooth is,'lhemlal surfaces at depth, or distortion to the isothermal surfaces due to a number of influences such as taults, water movement, or subsurface temperature or L,,'ndueti,ity anomalies. Water movement along faults and fractures increases spread values when water is heated at depth andtravelstowardthe surfacealong fault planes. Thiscan cause large horizontal temperature difti:renees at the same depth level and, consequently, a large spread in the temperature values. \\' atcr movement through permeable strata is another tactor that can strongly influence the temperature ,,'giIlle. Hydrodynamics appears to play an important 1\'1<' in the distribution of subsurface heat and so influenccs the temperature distribution. Gravity-imposed ,l,lwnward water movement occurs in the upper strata as",ciatN with surface recharge while, in other areas, l',-nneable beds allow upward water movement. These
Temperature ('C) 100
50
upward and downward water movementsin the sharply dipping permeable formations cause the isotherms to dip sharply, giving largespreadvalues as indicated. Such water movement is thought to occur in the western Alberta basin. In the east, where the water movement is mainly lateral, the isotherms lie parallel to the surface and the spread values are smaller. Another reason for an increase in spread values may be the effect ofthermal conductivitycontrasts between adjacent dipping beds. For example, the clastic and shaly formationsabovethe Palaeozoicsurfaceare oflow thermal conductivity,whereas the calcareous and evaporitic formations below the Palaeozoic surface are more conductive. In steeplydippingbeds,suchchangesin thermal conductivity may distort the isotherms so that temperature vs. depth plots over a 3 x 3 = 9 TWP/RGE area will exhibit large spread values. This is illustrated in Figure 5.7-4 in which two 9 TWP/RGE areas of west-central Alberta are compared (Lam and Jones, 1984).Although these areas are in the same general region, they exhibit a totally different temperature gradient and spread.
Temperature (OC)
150
50
o-c-'-~-'----'--'-L->--'~-'----L--<"-'---'--+-o
100
Grad. = 24.2 QC/km
(13.3'FI103ft)
1-
, ,
,
..
,
,, " ,
-
\.
2-
...
, ~,
\~
2
\
\\ ",.
150
Grad. = 31.7°C/km (17.4'F/l 03ft)
.. ... .. . '
,),
tit
"
, '~ , ,, ',,,, ,, ,, "
.
to :::.
% m
. ....'-
c
... ... ... ... .... .... ... . ..• ..
'~
,-
,
4
5
,
scar-care Error ofEstimate
Standard Error of Estimate S.4°e
4.2'C
100 ,3:..~~:
200
Temperature (OF) ~~ earnand Jones, 1984.
=;sure 5.7-~
300
32
100
200 Temperature (OF)
Examples of Temperature vs. Depth Plots from Two Areas in Southern Alberta
300
1
ESTIMATION OF VOLUMES OF HYDROCARBONS INPLACE
5.7.5
Data Quality
During the drilling and completion ofa well, there are a number of opportunities to obtain BHT data. It is important to plan ahead so that the best quality data is obtained at the most opportune time and at minimal cost. IfBHT data is required while drilling a well, a drillstem test may provide the most representative data. Temperature derivations from logs run in the open-hole wellbore have been observed to be consistently lower than the BHT measured from drillstem tests despite the use of the Horner plot method to extrapolate the temperature buildup (Hermanrud et aI., 1990). The preferred method of obtaining representative BHT data is to obtain measurements following the well completion and an appropriate shut-in period. An ideal opportunity to obtain BHT data is in conjunction with a static bottom-hole pressure measurement, or a pressure buildup following an oil or gas deliverability test. Even under these circumstances, two factors could influence the accuracy of a temperature measurement: a large drawdown during the flow period, and the depth at which the temperature is recorded. To improve the quality of data, the wellbore drawdown should be moderate, and the temperature sensor should be within the producing interval (Hermanrud et. aI., 1991). It is important to note that temperature gradients often vary from one region to another and, even within the same area, may deviate significantly from the mean average due to the proximity ofcertain geological features in the area. Prudence is required to recognize these deviations and not dismiss them as errors in measurement.
References Cronquist, C. 1990. Reserves Estimation, Petroleum Engineering Manual. IHRDC Publishers, Boston, MA, PE 508, pp. 53-56. Deming, D., and Chapman, D.S. 1988. "Heat Flow in the Utah-Wyoming Thrust Belt from Analysis of Bottom-hole Temperature Data Measured in Oil and Gas Wells," Jour. ofGeophys. Res., Vol. 93, Nov. 1988, pp. 13,667 - 13,672. Energy Resources Conservation Board. 1991. Alberta's Reserves ofCrude Oil, Oil Sands, Gas, Natural Gas Liquids, and Sulphur. Calgary, AB. Hermanrud, C., Cao, S., and Lerche, I. 1990. "Estimates of Virgin Rock Temperature Derived from BHT Measurements: Bias and Error." Geophysics, Vol. 55, Jul. 1990, pp. 924-931. Hermanrud, C., Lerch, I., and Meisingset, KK 1991. "Determination of Virgin Rock Temperature from Drillstem Tests." JPT, Vol. 43, Issue 9, Sep. 1991, pp. 1126-1131. Lam, H.L., and Jones, F.W. 1984."A Statistical Analysis of Bottom-hole Temperature Data in the Hinton Area of West-Central Alberta." Tectonophysics, Vol. 103, pp. 273-281. Lam, H.L., Jones, F.W., and Majorowicz, J.A. 1985. "A Statistical Analysis of Bottom-hole Temperature Data in Southern Alberta." Geophysics, Vol. 50, Apr. 1985, pp. 677-684.
Caution is recommended when taking a BHT measurement in shallow wells on hot summer days using a maximum reading thermometer. The maximum reading could be the surface temperature and not the BHT. In conclusion, BHT data is available from a number of sources and the quality of this data is often not questioned. Such acceptance stems from the fact that small variations in BHT when converted to absolute temperature result in a very small percentage error in the overall reserve estimate. On the other hand, to minimize this error and improve the overall quality of the reserve estimate, one should take advantage of the drilling and completion process to obtain data that best represents the true BHT conditions.
85
Ib-.
_
DETERMINATION OF OIL AND GASRESERVES
5.8
RESERVOIR PRESSURE
5.8.1
Introduction
Throughout the productive life of a reservoir, a record of its pressure is necessary in order to make a number of necessary calculations. Initial pressures obtained after the discovery of a pool are needed for the calculation of volumetric reserves, particularly for gas reservoirs. Reservoir pressure is needed to determine gas compressibility and formation volume factors for oil and natural gas, and to undertake PVT analysis. Material balance calculations for both oil and gas systems require initial reservoir pressures and subsequent pressure history after production has commenced. Fluids flow when a pressure difference is created between two points. When hydrocarbons are removed from a reservoir, a pressure drop is created in the wellbore. This causes the pressure within the formation to drop. When a flow of fluid is stopped or "shut in," the pressure will equilibrate until it reaches stable reservoir conditions. The time required to reach a stabilized pressure varies from reservoir to reservoir. Analysis of the pressure stabilization or "buildup" will reveal information about the permeability of the formation, the distance to reservoir boundaries, and any damage to the formation. If stable conditions are not reached, the pressure buildup data may be extrapolated to estimate the reservoir pressure.
5.8.2
Data Sources
Two types of pressure recorders are available to measure reservoir pressures: mechanical and electronic gauges. The mechanical gauge consists of a coiled bourdon tube pressure element, which spirals outward as pressure is increased inside it. A stylus on the end of the tube scribes a thinly coated metal chart, which is slowly rotated by a clock in the recorder. The distance the stylus moves is proportional to the pressure inside the bourdon tube. Electronic gauges use strain, capacitance transducer, and quartz gauges as pressure-sensing devices. These recorders offer the option of programmable sampling times, and are generally more accurate than mechanical recorders. All pressure gauges must be calibrated to ensure that correct pressures are being recorded. Generally, regulatory agencies are responsible for setting guidelines for gauge calibration. It is important that gauges used for pressure surveys be calibrated regularly. It is common practice to use at least two recorders during pressure surveys to ensure that representative data
86
is recovered even if one recorder should fail. Running tandem recorders also permits comparison between the two to verify the accuracy of the measurements. Pressure measurements are usually obtained by lowering these recorders or "bombs" down the wellbore. As discussed in Section 5.6.2, the first indication of the pressure of a formation may come from a drillstem test (DST), which is usually conducted before production casing is run into the wellbore (but cased-hole DSTs are not uncommon). DSTs are also run immediately upon penetration of a prospective formation in order to examine its potential before drilling fluids damage the .zone. After drilling operations are finished and the well has been completed for production, pressure measurements are usually obtained by lowering pressure recorders down the wellbore on a wire line. In some instances, the pressure recorders are left downhole for extended periods of time. In other circumstances recorders measure the reservoir pressure for only a few minutes. The latter type of pressure measurement is referred to as a static gradient. This measures wellbore pressures at different depths, and these pressures are plotted against the measurement depth. The resultant plot is used to identify density changes in the wellbore fluids. Pressure gradients at reservoir depth are also estimated. Production tests are commonly conducted when pressure recorders are left downhole. When left for a period of time, pressures are recorded vs. time. Figure 5.8-1 is an example of data from a static gradient, and Figure 5.8-2, from a flow and buildup test. Formation pressures may also be measured before the well is completed by running open-hole wireline tools. These tools push a probe into the formation and record the formation pressure. Bottom-hole pressures are also estimated by measuring surface pressures and adding the calculated pressure due to the hydrostatic head of fluid in the wellbore. In cases where gas and liquid are both present in the wellbore, acoustic level indicators are employed to determine the fluid level. The respective hydrostatic heads ofthe fluids are then calculated and added to the surface pressure to estimate the bottom-hole pressure.
5.8.3
Data Analysis
A major problem in recording pressure data is determining whether the reservoir pressure is actually stabilized. There are three widely accepted methods of obtaining a stabilized bottom-hole pressure. The first involves the gathering and extrapolation of pressure
. ESTIMATION OFVOLUMES OFHYDROCARBONS INPLACE
0 Depth
200 400
~
600 800 1000
g
~
= 1i.
•
(m)
(kPa)
0 300 600 900
11784
1400 1600 1800
12474 12819 13164
1500 1800
13509 14862
2100 2150
17532
2300 2388
Q
12129
1200
2200 2250
1200
Pressure
17977 18422 18867
19757 20095
~----- Slope = 1.15 kPaim -.\
\~ Fluid contact @
1670 mK8
2000
Slope = 8.9 kPaim/ 2200
produced fluid prior to being shut in, and the variable , dt, is the elapsed time since the well was shut in. Plotting the data on semi-log paper theoretically reveals a straight line when the infinite acting radial flow period is reached. Extrapolation ofthis data to the semilog value of I yields the theoretical static reservoir pressure. The semi-log value of I corresponds to a shut-in time of infinity. Example 1
This example shows how to extrapolate buildup pressure to obtain a static reservoir pressure. Pressure recorders were lowered into a new oil well. The well was flow-tested for 140 hours at a constant rate and then shut in to allow the reservoir pressure to build. The data shown in the following table was obtained from pressure recorders.
2400 2600 10000
Pressure Recorder Data 14000
18000
22000
Pressure (kPa)
Figure 5.8-1 Static Gradient
26,.------------------, 24 22
as 20
ll.
... 18 (t'jX 16
C14
~ 12 ~
10
=8 d:• 6 4 2
o +-~_~-_-~-~~-_--I 300 400 o 100 200 Time (h)
Figure 5.8-2 Pressure vs. Time
buildup data. The second utilizes a static pressure measurement, where the shut-in time to reach stabilization is determined from previous buildup tests. The third method is also a static pressure measurement, but the shut-in time of the well is arbitrarily set. The method most commonly used to extrapolate pressure data was first discussed by Horner (1951). The procedure involves the plotting of pressure data during buildup vs. a time function [(Hdt)/dt) on a semi-log plot. The time, t, is the time during which the well
Time (h)
0 1 2 4 8 24 60 90 140 140.11 140.25 140.50 140.75 141.2 142.25 144 148 152 156 160 170 180 190 210 230 350
Pressure (kPa)
Comments
20175 9830 8750 7290 5570 5050 5030 4670 4665 5577 5924 6435 6743 7656 9 163 11077 13338 14273 14878 15345 16146 16681 17 074 17623 17996 18989
begin flow continue
" " " "
" " stop flow
continue
" " " " " "
" " " "
" " " "
" end of test
As previously stated, the time function [(HLlt)/Llt) must be calculated. The time, t, is 140 hours, and since Llt is the elapsed time since the well was shut in, 140 must be subtracted from all the times once the buildup
87
b
_
-DETERMINATION OFOILAND GASRESERVES
portion ofthe test begins. The resultant data is shown in the following table:
Horner Time Data Time (h)
(t+4.t)/M
Pressure (kPa)
-
0 0.11 0.25 0.50 0.75 1.20 2.25 4 8 12 16 20 30 40 50 70 90 210
Once pressure data for a reservoir has been collected from two or more wells, the data should be corrected to a common datum depth. Many hydrocarbon reservoirs vary in elevation from one end to the other. In these situations, a common pool elevation or datum is generally established. When pressure data is recovered from wells that have different elevations, the pressure data must be corrected to this datum depth. The pressure gradient is multiplied by the difference in elevations, and the result is added to or subtracted from the uncorrected data. This procedure will correct all pressures collected for a given reservoir to a common datum depth.
4665 5577 5924 6435 6743 7656 9 163 II 077 13338 14273 14878 15345 16146 16681 17074 17623 17996 18989
1273.7 561.0 281.0 187.7 117.7 63.2 36.0 18.5 12.7 9.75 8.00 5.67 4.50 3.80 3.00 2.56 1.67
Example 2 This example illustrates how to determine the datum pressure for two wells.
When the data has been plotted on semi-log paper, a trend can be seen toward the end of the buildup (Figure 5.8-3). When the trend is extrapolated, the intersection of the line with a time value of I (which means infinite t.t) indicates the theoretical pressure the reservoir will reach. The extrapolated pressure was estimated to be 20 175 kPa. 26 24 "'
~~ ~ Extrapolated pressure of20 175kPa
~
18
"0
14
~
~
III
~
~
a
10
D
8 6 4 2
.
a
2 10
October 3, 1991
October 3, 1991
2388.0mKB (-1458.0 mss)
2453.0mKB (-I 470.0mss)
20095 kPa
20197 kPa
8.9 kPa/m -1458.0 mss -I 467.0 mss
8.8 kPa/m -1470.0 mss -1467.0 mss
9.0m
-3.0m
a
o -I---~-_ _~_ _~ 10
Date: Recorder run depth:
D
a a
1
Pool Datum is at 1467m subsea Well: 01-02-003-04 W5M 09-02-003-04 W5M Formation: Triassic Triassic KB· elevation: 930m 983 m Top of formation: 2397.0 mKB 2460.0 mKB 930.0m 983.0 m -2397.0 m -2460.0 m Formationtop: -I 467.0 mss -1477.0 mss
Pressure at run depth: Pressure gradient (obtained from static gradient) Pressure at pool datum:
x 16
::. 12
However, if the reservoir is not at stable conditions, or ifdepletion is thought to have occurred, a buildup analysis is very useful in the determination ofstable reservoir pressures.
103
-I 104
[(1 + 61)/61]
Figure 5.8-3 Horner Plot In this example, the first pressure point recorded matches the calculated pressure found by the Homer analysis. In cases where the initial reservoir pressure is at static conditions, a buildup analysis is not necessary.
• KB
=
9.0m x 8.9 kPa/m
-3.0m x 8.8 kPa/m
80.1 kPa
-26.4 kPa
20095.0 kPa +80.1 kPa
20197.0kPa -26.4 kPa
20 175.1 kPa
20170.6 kPa
The elevation of the drilling platform at the kelly bushing.
88
s
ESTIMATION OFVOLUMES OFHYDROCARBONS INPLACE
Datum pressures for the two wells are 20 175 kPa and 20 171 kPa at 1467 m below sea level.
Ai = area of common pressure
1\ = total reservoir area Vi = volume ofcommon pressure V, = total reservoir volume
Once a datum pressure has been determined for all wells surveyed in a pool, it may be determined that the pressures still vary from point to point. What must be found now is the average reservoir pressure; three methods are commonly used. The first is an arithmetic average. The second is an area-weighted average, where areas of the reservoir that have similar pressures are grouped together. The area-weighted average is the sum of the products of areas times pressures divided by the total area. The third method is the volume-weighted average. This method may utilize the rock volume, pore volume or hydrocarbon pore volume. The following equations summarize the three methods:
P = l:(P) / n P = l:(P i X AJ / A,
(2)
Volume-weighted average P = l:(Pi X VJ / V,
(3)
Arithmetic average Area-weighted average
Example 3 This example illustrates how to estimate the average reservoir pressure using the arithmetic, area-weighted and volume-weighted methods. The porosity volume map in Figure 5.8-4 was found to have the volumes for the four constant pressure areas as shown in the following table: Calculation of Average Reservoir Pressure
(I)
Pressure (kPa)
Area (ha)
Volume (ha.m)
20000 20050 20100 20150
115 179 155 90
1404 4425 I 930 435
where P = average reservoir pressure Pi pressure point n = total number of points
20050 kPa
20100 kPa
20000 kPa 20080 kPa •
20150 kPa
•
:20 120kPa
•
20040 kPa:
20000 kPa
• .:
•
20060 kPa
····:~20m
20170 kPa
20130 kPa
•
•
15 m
•
20175 kPa
20 140kPa: • 20 090
kPa~.:-:-·- ' - - - - - 5 m
•
•
20110 kPa
o
I ha(lOOmx 100m)
---
om
Area of 20 000 kPa pressure Area of20 050 kPa pressure Area of 20 100 kPa pressure Area of20 150 kPa pressure
liS ha 179 ha ISS ha 90ha
Volume of20 000 kPa pressure Volume of 20 050 kPa pressure Volume of 20 100 kPa pressure Volume of 20 ISO kPa pressure
1404 ha·m 4425 ha-m 1930 ha-m 435 ha·m
Total area of pool
539 ha
Total volume of pool
8194 ha-m
Figure 5.8-4 Porosity Volume Map 89
7
DETERMINATION OF OILAND GASRESERVES
Arithmetic Average ~(P,)
n
= 20 000 + 20 040 + 20 060 + 20 080 + 20 090 + 20 110 + 20 120 + 20 130 + 20 140 + 20 170 + 20 175 = 221 115 kPa = 11
P =
~(PYn = 221 115 kPa /11
= 20101 kPa Area-Weighted Average ~(Pj x Ai) = (20000 x liS) + (20 050 x 179) + (20 100 x ISS) + (20 ISO x 90) = 10 817 950 kl'a x ha A, = 539 ha p = ~(Pj x Ai)/A, = 10817 950 kPa x ha / 539 ha = 20070 kPa
Volume-Weighted Average ~(Pj
x V) = (20000 x 1404) + (20 050 x 4425) + (20100 x 1930) + (20 ISO x 435) = 164359500kPaxhaxm V, = 8194haxm=81.94x 106m3
p
= ~(Pj x Vj)N, = 164359500 kPa x ha x m / 8194 ha x m = 20058 kPa
It should be noted that in this case the arithmetic and area-weighted averages result in higher pressures than the most rigorous volume-weighted average.
References Homer, D.R. 1951. "Pressure Build-up in Wells." Proc., 3rd World Petroleum Congress, E. 1. Brill, Leiden, Netherlands, Vol. II, p. 503.
90
s
ESTIMATION OFVOLUMES OFHYDROCARBONS INPLACE
5.9
GAS FORMATION VOLUME FACTOR
5.9.1
Introduction
PXV=nxRxT
In order to determine the gas formation volume factor, Bg , which relates the volume of gas in the reservoir to the volume at the surface at standard conditions of'temperature and pressure, it is necessary to fully understand gas behaviour. This is explained in the four subsections that follow. Often the terrns in the B g determination (Equation 15 in Section 5.9.5) are used directly in the equation for calculating in-place gas volumes, but it is useful to have Bg as one term for hand calculations and simple material balance work.
5.9.2
Ideal Gas Law
Three properties affect the amount of gas in a reservoir: pressure, P, temperature, T, and volume, V. The 19th century chemists, Boyle and Charles, found that for a given amount of gas (see Example 1) the following relationship holds true: P, X V,
T,
P, xV, =
T,
(I)
where the subscript I signifies the first set of conditions, and the subscript 2, the second set of conditions. Equation (l) assumes that the amount of gas in the system does not change, and that the gas behaves as an ideal gas.
Example 1 A balloon has a volume of0.1 m 3 at 1000 kPa and 300°K. If the contents of the balloon are heated to a temperature of450 oK, either the pressure or the volume (or both) must change. In this case, it is assumed that both change, so that the pressure is now 1200 kPa and the volume is 0.125 m3• The pressures, volumes and temperatures at both conditions would be related as follows: 1000 kPa x 0.1 m'
1200 kPa x 0.125 m'
300'K
450'K
In order to calculate the amount of gas in a closed chamber of fixed volume, two more constants must be defined: the first is the moles of gas, which is essentially the number of molecules of gas; the second is the gas constant. The resultant equation is known as the Ideal Gas Law and expressed as follows:
where P V n R T
5.9.3
= = = = =
(2)
pressure of gas in container volume of gas in container moles of gas in container gas constant temperature 0 f gas in container
Gas Compressibility Factor
The Ideal Gas Law may be used to calculate the properties ofgases at moderate temperatures and pressures; however, the law does not hold true at high temperatures and pressures. To correct for the deviation, a term called the "gas compressibility factor" or "gas deviation factor," Z, must be included iri the equation. PXV=ZxnxRxT
(3 )
The gas compressibility factor is designed to correct the volume of a theoretical ideal gas to the volume occupied by a real gas. This factor can be determined in the laboratory by measuring the actual volumes of a given amount ofgas at prescribed pressures and temperatures and comparing these to the ideal volumes calculated by the Ideal Gas Law. It should be noted that the compressibility factor will change with temperature, pressure and gas composition. In cases where the gas compressibility factor is not obtained from detailed laboratory work, it can be closely estimated by first calculating two pseudo-critical properties that are used to determine the compressibility factor of a natural gas: pseudo-critical pressure and pseudo-critical temperature. Both of these can be calculated if the composition of the gas is known. Gas compositions are usually determined by gas chromatography on gas samples; each component is expressed as a mole fraction of the total. To calculate the pseudo-critical properties of a natural gas, the critical pressure and critical temperature of all the components are needed. These values are available in numerous publications, such as the Engineering Data Book (Gas Processors Suppliers Association, 1980). The pseudo-critical pressure of a gas is defined as the sum of the products ofthe mole fraction of each component times the critical pressure of that component. The pseudo-critical temperature is the sum of the products of the mole fraction of each component times the critical temperature of that component. The equations for calculating the pseudo-critical properties are as follows:
91
b
DETERMINATION OFOIL AND GAS RESERVES
P, = L(Xj x P;)
(4)
Gas Analysis
= L(Xj x T;)
(5)
Mole Fraction
T,
Component
where P, = pseudo-critical pressure T, = pseudo-critical temperature x j = mole fraction of component i P j = critical pressure of component i T j = critical temperature of component i Thomas et al. (1970) found that the pseudo-critical properties may also be estimated using the specific gravity of the gas. The specific gravity, SG, is the ratio of the gas density to the density of air. P, = 4892.547 - (404.846 x SG) (kPa)
(6)
T, = 94.717 + (170.747 x SG) (OK)
(7)
Once the pseudo-critical properties are found for a given gas, the pressure and temperature of the gas in the reservoir are needed to calculate the pseudoreduced properties of the mixture. The pseudo-reduced pressure is the ratio ofthe actual pressure to the pseudocritical pressure. The pseudo-reduced temperature is the ratio of the actual temperature to the pseudo-critical temperature.
P, = PIP,
(8)
T, = TIT,
(9)
where P, = pseudo-reduced pressure P = pressure of natural gas system T, = pseudo-reduced temperature T = temperature of natural gas system Once the reduced properties ofthe natural gas at a given pressure and temperature have been calculated, the gas compressibility factor can be determined by the use ofa gas compressibility factor chart (Figure 5.9-1) published by Standing and Katz (In Standing, 1977). The gas compressibility factor may also be determined by the use of a computer algorithm (Dranchuk et al., 1977).
Example 2 This example illustrates how to estimate compressibility factor of natural gas from a southern Alberta gas well. Well location: Formation: Initial formation pressure: Formation temperature:
92
02-03-004-05 W6M Bow Island 6790 kPa 296° K (23°C)
Helium (He) Nitrogen (N2) Methane (C1) Ethane (C2) Propane (C3) iso-Butane (iC4) n-Butane (nC4) iso-Pentane(iC5) n-Pentane (nC5) Hexane (C6)
0.0012 0.0469 0.9322 0.0129 0.0045 0.0007 0.0009 0.0003 0.0002 0.0002
Total
1.0000
Pc = L(xixPi) T, = L(xi x Ti) P, = 6790 I 4541.87 296 I 190.16
Critical Press. (kPa)
Critical Temp.
227.53 3399.00 4604.00 4880.00 4249.00 3648.00 3797.00 3381.00 3369.00 3012.00
5.23 126.10 190.55 305.43 369.82 408.13 425.16 460.39 469.60 507.40
(OK)
4541.87kPa 190.16°K 1.49 1.56
T,
=
Z
= 0.88 ( from Standing and Katz chart )
In compatison, if only the specific gravity were known, P, and T, would be estimated as follows: Specific gravity
= 0.587
= 4892.547 - (404.846 x 0.587) T, = 94.717 + (170.747 x 0.587) P, = 6790 I 4654.902 T, = 296/194.945
= = = =
P,
4654.902 kPa 194.945°K 1.46 1.52
Z = 0.87 ( from Standing and Katz chart)
5.9.4
Sour Gas
Calculating Z using the method described works well for sweet gases-natural gases that do not contain carbon dioxide (C02 ) or hydrogen sulphide (H 2S). Natural gases that do contain carbon dioxide and hydrogen sulphide are called sour gases. Estimation ofthe gas compressibility factor of sour gases was found to be incorrect when using the chart published by Standing and Katz, and several methods have been developed to estimate the correct compressibility factor for sour gases. Wichert and Aziz (1972) compared two of these methods to one that uses the Standing and Katz chart. It was found that the chart could be used if the pseudocritical temperature and pseudo-critical pressure were adjusted. An adjustment factor, e, was developed by Wichert and Aziz to correct the pseudo-critical
ESTIMATION OFVOLUMES OFHYDROCARBONS INPLACE
1.1
a
1
Pseudo-reduced Pressure 3 4 5
2
6
7
8
1.1 Pseudo-reduced Temperature
3,0 2,8 2.6 2.4 2,2
1.0
1.0 1.05 1.2
0.95
2.0 1.9
0.9
1.8 1,7 1,6
0.8
1.7
1,5 1.45 1.4
0.7
1.6 1.35
N ..:
-'"
N 1.3
a
o
LL >.
0.6
..:
..
1.5
1.25 ,~
~
1.2
en
,~
(fJ
~~ \~ '/.~
0.5
0.
E
a
1.4
0.
E
a
-»
U
'"
LL >.
"" :c '00
,~
:c '00
a ti
U
0.4
1.3
0.3
1.2
0.25 2.8
1.1
2. 6
2.4
1.1 2.'
2.2 '/.~
Compressibility of
Natural Gas Jan. 1, 1941
\.9
1.0
0.9
1.0
1,1
0.9
7
8
9
10 11 12 Pseudo-reduced Pressure
13
14
15
Source: Standing, 1977.
Figure 5.9-1 Compressibility Factors for Natural Gases
93 •
DETERMINATION OF OIL AND GASRESERVES
properties. The factor may be determined graphically or by using the following equation: e = 120 x (AO. 9 - A1. 6) + 15 X (B°.5 - B4 )
(10)
= :E(xi X P) = 6056.39 kPa T, = :E(xi X T) = 273.18 OK P,
e
where A = combined mole fraction of COz and HzS B = mole fraction of HzS
= 32.77 T,' = 273.18 - 5/9 (32.77) = 254.98 OK 6056.39 X 254.98 Po' = 273.18 + 0.3003 (1-0.3003) X 5/9 (32.77)
(11)
=
c
where T c'
r, x Te' T, + B x (1-B)
(12)
x (5e/9)
adjusted pseudo-critical temperature P,' = adjusted pseudo-critical pressure =
Example 3 The adjustments described are detailed in this example of a sour natural gas well from the Foothills area of Alberta. Well location: Formation: Initial formation pressure: Formation temperature:
03-02-004-05 W6~ Rundle 34300 kPa 359 OK (86°C)
Gas Analysis Component
Mole Fraction
x (0.3762°·9 - 0.37621.6) +
15 X (0.3003°.5 - 0.3003 4)
The pseudo-critical temperature and pseudo-critical pressure are estimated in the normal manner and then adjusted as follows:
P'
= 120
Critical Press. (kPa)
Critical Temp. (OK)
Nitrogen (N,) Sulphide (HzS) Carbon dioxide
0.0104 0.3003
3399.00 9005.00
126.10 373.50
(CO z)
Methane (Cl) Ethane (C2) Propane (C3) iso-Butane(iC4) n-Butane (nC4) iso-Pentane(iC5) n-Pentane (nC5) Hexcane (C6) Heptane + (C7+)
0.0759 0.5277 0.0358 0.0079 0.0018 0.0041 0.0020 0.0022 0.0059 0.0260
7382.33 4604.00 4880.00 4249.00 3648.00 3797.00 3381.00 3369.00 3012.00 2486.00
304.19 190.55 305.43 369.82 408.13 425.16 460.39 469.60 507.40 568.76
Total
1.0000
= 5574.83 kPa
P, = 34300/5574.83 = 6.15 T, = 359/254.98 = 1.41 Z = 0.87 ( from Standing and Katz chart) In comparison, if the critical properties had not been adjusted, the compressibility factor would have been estimated to be 0.77, a difference of 11.5 percent. Use of the incorrect compressibility factor in estimating reserves would result in large errors.
It should be noted that the compressibility factor estimated is only correct at the pressure and temperature used in calculating the pseudo-reduced pressure and pseudo-reduced temperature. When the compressibility factor of a gas is required for material balance calculations, each pressure point requires that a corresponding compressibility factor be estimated.
5.9.5
Derivation of Gas Formation Volume Factor
Gas formation volume factor, Bg• relates the volume of gas in the reservoir to the volume on the surface at standard conditions of temperature and pressure and is often used to simplify hand calculations ofgas reserves. For the purpose of estimating reserves, B g is generally expressed as the amount of space occupied at standard conditions by a unit volume of gas under reservoir conditions. The dimensions ofBg are unit volume at standard conditions per unit volume at reservoir conditions, and therefore, Bg is dimensionless. Derivation of Bg begins with the following equation for nonideal gases:
P,x V,
P; x V;
Z,T,
Z;T;
(13)
where P" V" Z, and T s are all measured at standard surface conditions, and Pi' Vi' Z, and T, are all measured at initial reservoir conditions. Transposing
94
r
ESTIMATION OF VOLUMES OFHYDROCARBONS IN PLACE
tenus, Bg (as defined in this section) is derived as follows:
v
vo Iume 0 f gas under
= - ' = standard surface conditions
B g
Vi
per unit volume ofreservoir space
(14)
Assuming that Vi is one unit volume ofreservoir space, and that the gas compressibility factor, Z" at standard surface conditions is unity, B g can be reduced to:
(15)
References Dranchuk, P.M., Purvis, RA., and Robinson, D.E. 1977. "Computer Calculation of Natural Gas Compressibility Factors using the Standing and Katz Correlations." Institute of Petroleum Technology, IP-74-008, Vol. I. Gas Processors Suppliers Association. 1980. Engineering Data Book, Tulsa, OK. Standing, M.B. 1977. "Volumetric and Phase Behavior of Oil Field Hydrocarbon Systems." SPE of AIME, Dallas, TX. Thomas, H.K., Hankinson, RW., and Phillips, K.A. 1970. "Determination of Acoustic Velocities of Natural Gas." JPT, Vol. 22, pp. 889-895. Wichert, E., and Aziz, K. 1972. "Calculate Z's for Sour Gases." Hydrocarbon Processing, May 1972.
B g may now be substituted into the equation for calculating in-place volumes of nonassociated and gas cap gas.
95
DETERMINATION OF OILANDGASRESERVES
5.10
OIL FORMATION VOLUME FACTOR
2.10
B
the volume in cubic metres (barrels) that one stock tank cubic metre (barrel) occupies in the formation at the prevailing reservoir temperature and pressure. A stock tankcubicmetre (barrel) is defined as the volume occupied by one cubic metre (barrel) of crude oil at standard pressure and temperature, which are 101.325 kPa and 15°C (14.65 psi and 60°F) respectively. Crude oil in the ground always contains varying amounts of dissolved gas (solution gas). Because both the temperature and the solution gas increase the volume of stock tank oil in the formation, the FVF will always be greater than one. The symbol B, is used in equations to refer to the formation volume factor. The reciprocal of the FVF is called the shrinkage factor. Just as the formation volume factor is multiplied by the stock tank volume to find the reservoir volume, the shrinkage factor is multiplied by the reservoir volume to find the stock tank volume. Although both terms are in use, most petroleum engineers use formation volume factor.
5.10.2 Data Sources A laboratoryanalysis of fluid properties is the best source ofdata to estimate the FVF. It is preferable that the laboratory analysis be made on a sample obtained during the completion ofthe discovery well, and that the sample represent as nearly as possible the original reservoir fluid. This will ensure that the original FVF is accurately determined with respect to the bubble point and a decline in the bottom-hole pressure. Figure 5.10·1 shows the oil FVF in a typical high. gravity undersaturated oil reservoir when the reservoir pressure declines from the initial pressure to stock tank condition as determined from a laboratory analysis (the symbols shown in this figure are defined in Example 2, Section 5.10.5.) The oil FVF can also be estimated from empirical methods and correlations available from the technical literature (Amyx, 1960). A correlation prepared by Katz (1942) from data on mid-continent crudes requires the reservoir temperature, and the pressure, gas-oil ratio, and API gravity ofthe crude. A second empirical correlation developed by Standing (1947) for California fluids requires the total gas-oil ratio, the gravity of the stock tank oil and produced gas, and the reservoir temperature.
96
4
1.90
fi ~
~ l5
I
1.80
1.70
/A
~ 1.50
..
~
t 40
/
I
4 _ _.11.
I
/'
Q)
Differential liberation
~I.I.-1.1._
Bo1b ( 1 . 7 2 3 ) - / -
~ 1.60 5 ~
t..__
1:'0...)
2.00
5.10.1 Introduction The formation volumefactor (FVF) for oil is defined as
'j
_
od b(2.074)
Flash Liberation
) :g
I
:§l
/
A
;-
~
1.30
!
a.
1.20
e
1'"
1.10
1.00 - ¥ - - - - - , r - - - r - - , - - - - , - - . - - - - - , 1000
3000 4000 Pressure (psig) Source: Chevron CanadaResources.
Figure 5.10·1
2000
5000
6000
Comparison of Formation Volume Factor by Differential and Flash Liberation
5.10.3 Data Acquisition Before representative samples of the reservoir fluid are collected, it is important that the well be properly conditioned. A complete well conditioning and sampling procedure is described in Chapter 5.6. In most reservoirs, the variations in reservoir fluid properties among samples taken from different parts of the reservoir are not large, and lie within the margin of error inherent in the techniques of fluid sampling and analysis. On the other hand, some reservoirs, particularly those with large closures, have large variations in fluid properties, which may be explained by a combination of the temperature gradients, gravitational segregation, and lack of equilibrium between the oil and the solution gas. Methods for handling reservoir calculations where there are significant variations in the fluid properties have been documented in the literature (Cook et al., 1955; McCord, 1953).
5.10.4 Data Analysis The composition of the stock tank oil will be quite different from the composition ofthe original reservoir fluid. Most of the methane and ethane will have been released from solution, and sizeable fractions of the propanes, butanes and pentanes will have vapourized as the oil moves from the reservoir to the stock tank and the pressure is reduced. The change in liquid composition is not a single nor a well-defined process,
1 ESTIMATION OF VOLUMES OF HYDROCARBONS IN PLACE
but is a series of flash and differential liberation processes. The difference between these two processes is that in the flash liberation process, all of the gas remains in contact with the liquid, while in the differential process, some ofthe gas is released (removed from contact with the liquid phase). For this to be so, the volume of the system in the flash process must increase as the pressure declines. Thus the flash process is one of constant composition and changing volume, and the differential process is one of constant volume and changing composition. In the case of reservoir fluids which are at their bubble point when the pressure declines as a result of production, the gas liberated from the oil does not flow to the well, but accumulates until a critical gas saturation is reached. This critical saturation will be reached sooner in the vicinity of the well where the pressure is lower than at more distant points, particularly for wells producing under large pressure drawdowns. With gas saturations greater than critical near the well, the gas will move more rapidly than the oil (differential liberation), whereas in the remainder of the reservoir the liberated gas will remain in contact with the oil (flash liberation). The volume ofthe reservoir surrounding the producing wellbore, where the gas is highly mobile, is usually only a small part ofthe total drainage area. Thus, where there is a more moderate pressure decline below the bubble point in the larger part of the reservoir, the flash liberation is more representative. On the other hand, when the gas saturation exceeds the critical value in most of the drainage area, the gas will flow much faster than the oil. This situation would be characterized by producing gas-oil ratios considerably in excess of the initial solution gas-oil ratio. With the removal of gas from contact with the oil, differential liberation more closely represents the reservoir behaviour. Consequently, differential liberation data should be applied to the reservoir fluid when the reservoir pressure has dropped considerably below the bubble-point pressure and the critical gas saturation has been exceeded in most of the reservoir. Flow through tubing and a choke is a decliningpressure flash liberation in which the gas remains in equilibrium contact with the oil throughout the process. There are, however, important differences between tubing flash and laboratory flash liberation. Tubing flash liberation is accompanied by declining temperatures and, where the gas-oil ratio exceeds the initial dissolved ratio, the oil is in contact not only with its
own liberated gas but with additional gas produced from either the oil zone or a gas cap. In contrast the laboratory flash liberation is isothermal, and only the gas liberated from the sample is in contact with the liquid. When the volume of dissolved gas in the crude oil is low (indicating low volatility), there are only slight differences between the flash and differential liberation data. Under these circumstances the residual barrel by the differential process may be identified with the stock tank barrel, and differential liberation data may be used directly in the oil-in-place equation. Experience indicates that low volatility conditions may exist where the following are present: the stock tank gravity is below 30° API; the solution gas-oil ratio is less than 70 m 3/m3 (400 scf/stb); and the reservoir temperature is below 54°C (130°F). These are, of course, only approximate limits. When the volatility of the crude is high, as in the example shown in Figure 5.10-1, more consideration should be given to the predominant gas liberation mechanism occurring in the reservoir, in the wellbore, and in surface separation facilities. The FVF used may more closely approach that of a flash liberation (Craft and Hawkins, 1959). In a simple analysis, where only one FVF vs. pressure relationship is used, industry practice tends toward using an estimated flash liberation relationship. It may be argued that the flash process in the wellbore is the final equilibrium that the oil and gas phases must adjust to. Also, the pressure drop from bottom-hole pressure to separator pressure often amounts to a large fraction of the total pressure decline from formation pressure. In a common industry compromise, the differential liberation FVF curve is shifted by the ratio between the flash and differential liberation FVFs at the bubble-point pressure (refer to Example 2 in Section 5.10.5). In reality, however, each producing system is different and each should be examined closely to determine where the major pressure drops are occurring. In some cases, further testing and facility modelling may be warranted in order to maximize liquid production. Figure 5.10-1 shows the FVF calculated by both the flash and differential processes for a more volatile crude oil having a gravity of 46.6° API. The difference between these two curves is significant. This example helps to illustrate the variability of FVFs, and the importance of understanding the impact of each liberation process in the producing system.
97
.iII?
_
... DETERMINAnON OF OIL AND GAS RESERVES
5.10.5 Data Adjustment In the calculation of a flash FVF, it is often necessary to adjust the data from a laboratory fluid analysis to more appropriately represent the true producing conditions. Normally, flash separation data in a laboratory analysis is referenced to reservoir fluid volumes at the saturation pressure (bubble point). The following two examples show how the data from a reservoir fluid study is used to calculate the flash FVF for producing reservoir conditions. The data is provided in Tables 5.10-1, 5.10-2 and 5.10-3.
'Table 5.10-1
adjusted B,
=
Relative Volume VN••t
6000 5500 5000 4500 4000 3500 3300 3200 3100 3029 3010 2993 2982 2948 2773 2571 2333 2098 1849 1622 1390 1298
0.9398 0.9473 0.9556 0.9648 0.9750 0.9867 0.9921 0.9949 0.9979 1.0000 1.0027 1.0051 1.0067 1.0117 1.0408 1.0805 1.1404 1.2206 1.3309 1.4717 1.6732 1.7740
B ofbx VIV s• t
(I)
where adjusted B, = flash formation volume factor for pressures above saturation pressure
= formation volume factor from flash at saturation pressure VIV s. ,
= relative volume from pressure
volume relations at pressure above saturation pressure
Pressure Volume Relations
Pressure (psig)
*
Source: PVT Analysis by Core Laboratories Canada Ltd., on Chevron Pembina 1-9-50-12 W5M. Chevron Canada Resources, File 7013-795. "Relative volume is in barrels at the indicated pressure and temperature per barrel of saturated oil.
Example 1:
for oil compressibility. This is done by adjusting the measured volume at saturation pressure using the following equation:
For example, if the reservoir pressure is 27 600 kPa (4000 psig), then and
= 0.9750 (Table 5.10-1)
B ofb
= 1.723 (Table 5.10-2)
adjusted B, = 1.723 (0.9750) = 1.680
Example 2:
At Pressures Below Bubble Point
Because oil shrinkage occurs due to gas liberation at pressures below the saturation pressure, flash FVFs that are referenced to a volume at saturation pressure must be corrected to reflect this shrinkage. The adjustment is made using the following equation: adjusted n, = Bod ,
BOlli
B
(2)
cdb
where adjusted B, = flash formation volume factor for pressure below saturation pressure =
relative oil volume from differential liberation at pressure below saturation pressure
= formation volume factor from
flash at saturation pressure =
relative oil volume from differential liberation at saturation pressure
For example, if reservoir pressure is 14500 kPa (2100 psig), then
Bod
= 1.767 (Table 5.10-3)
B ofb
= 1.723 (Table 5.10-2) = 2.074 (Table 5.10-3)
B odb
At Pressures Above Bubble Point
Flash FVFs are normally referenced to a volume at saturation pressure. At pressures above the saturation pressure, the flash FVF must be corrected to account
VIV sat
and
. 1.723 ad[usted B o = 1.767 x - - = 1.468 2.074
98
5
ESTIMATIONOF VOLUMES OFHYDROCARBONS IN PLACE
Table 5.10-2
Separator Tests of Reservoir Fluid Sample
Separator Separator Gas-Oil Pressnre Temperature Ratio! (OF) (psig)
Ratlo!
Stock Tank Gravity (OAPI60°F)
Formation Volume FactorJ
Separator Volume Factor"
Specific Gravity of Flashed Gas
Gas-Oil
reservoir to 320
74
795
891
-
-
1.121
0.725
to 0
74
288
290
46.6
1.723
1.007
1.226
Source: PVT Analysis by Core Laboratories - Canada Ltd., on Chevron Pembina 1-9-50-12 W5M. Chevron Canada Resources, File 7013-795. 'Cubic feet of gas @ 60°F and 14.65 psia per barrel of oil @ indicated pressure and temperature. 2Cubic feet of gas @ 60°F and 14.65 psia per barrel of stock tank oil @ 60°F. 3Barrels of saturated oil @ 3029 psig and 228°F per barrel of stock tank oil @ 60°F. "Barrels of oil @ indicated pressure and temperature per barrel of stock tank oil @ 60°F.
Table 5.10-3 Pressure (psig)
Differential Vapourization Relative Oil Volume!
Relative Total Volume-
Solution Gas-Oil
Ratio!
Gas Formation Volume Factor"
Gas Expansion FactorS
3029
2.074
2.074
1634
2700
1.947
2.184
1406
0.00568
170.65
2403
1.852
2.326
1231
0.00660
151.52
2100
1.767
2.533
1071
0.00764
130.89
1801
1.695
2.818
934
0.00901
110.99
1502
1.627
3.249
805
0.01098
91.07
1202
1.565
3.905
685
0.01384
72.25
900
1.504
5.080
568
0.01884
53.08
598
1.442
7.481
452
0.02882
34.70
300
1.363
14.802
318
0.05791
17.27
0
1.088
0
1.32043
-
0.757
Source: PVT Analysis by Core Laboratories - Canada Ltd., on Chevron Pembina 1-9-50-12 W5M. Chevron Canada Resources, File 7013-795. I Barrels of oil at indicated pressure and temperature per barrel of residual oil at 60°F (Bg), 2Barrels of oil plus liberated gas at indicated pressure and temperature per barrel of residual oil at 60°F (BJ. 3Cubic feet of gas at 14.65 psia and 60°F per barrel of residual oil at 60°F
CR,).
4Cubic feet of gas at indicated pressure and temperature per cubic foot at 14.65 psia and 60°F (B g) . sCubic feet of gas at 14.65 psia and 60°F per cubic foot at indicated pressure and temperature: \lgas FVF (\lB g) .
99
DETERMINATION OFOILAND GASRESERVES
5.10.6 Summary
References
Production from the reservoir rock to the stock tank usually involves a combination of flash and differential liberation processes. In determining a value for the oil formation volume factor, the overall flow process of the oil stream should be analyzed to determine where the major pressure drops occur and what weighting should be given to the flash and differential FVFs. If the volatility of the crude oil is high, there may be a significant difference between the values of the FVF determined by the flash and differential processes (Figure 5.10-1). In such cases, the true FVF may more closely approach the flash liberation process. If the volatility ofthe crude oil is low, only slight differences between the flash and differential data are likely, and use of the differential liberation data may be feasible. Future changes in producing procedures should also be considered in making any assessment of the oil formation volume factor.
Amyx, J.W., Bass, D.M., and Whiting, R.L. 1960. Petroleum Reservoir Engineering. McGraw-Hill, New York, NY, pp. 429-435. Cook, A.B., Spencer, G.B., Bobrowski, F.P., and Chin, T. 1955. "A New Method of Determining Variations in Physical Properties of Oil in a Reservoir, with Application to the Scurry Reef Field, Scurry County, Texas." US Bureau of Mines Report, Feb. 1955. Craft, B.C., and Hawkins, M.F. 1959. Applied Petroleum Reservoir Engineering. Prentice-Hall, Inc., Englewood Cliffs, NJ, pp. 86-181. Katz,D.L. 1942. "Prediction of the Shrinkage of Crude Oils." Drilling and Production Practice, American Petroleum Institute, Vol. 137. McCord,D.R. 1953. "Performance Predictions Incorporating Gravity Drainage and Gas Cap Pressure Maintenance - LL-370 Area, Bolivar Coastal Field." Trans. AIME, Vol. 198, No. 232. Standing, M.B. 1947. "A Pressure-VolumeTemperature Correlation for Mixtures of California Oils and Gasses." Drilling and Production Practice, American Petroleum Institute, Vol. 275.
100
ESTIMATION OFVOLUMES OF HYDROCARBONS INPLACE
5.11
QUALITY AND RELIABILITY OF DATA AND RESULTS
5.11.1 INTRODUCTION The quality and reliability of reservoir data reflect directly on the results obtained in preparing reserves estimates. As indicated in preceding sections, the conditions under which basic data are obtained and the laboratory methods used to generate additional data are both important elements that must be taken into consideration. Elementary reasoning and common sense are also important elements in the process of preparing reserves estimates. In the determination of reservoir rock properties and fluid saturations, it is common practice to rely on core data as the reference point and to fit log analysis data to it. Consequently, the core data must reflect to the greatest degree possible the in situ conditions ofthe reservoir. Because of cost considerations, it is usually not possible to obtain cores under preferred conditions, such as specifically prepared oil base muds, lease crudes, and oriented, pressurized, or sponge coring techniques. Thus, in most circumstances, conventional cores comprise the best data available.
5.11.2 Permeability from Cores Permeability is a particularly important measurement obtained from cores because it provides an indication of whether hydrocarbons may be effectively produced from intervals of interest. The reliability of the permeability measurements can be influenced by the coring procedure (induced fractures or scale formation), weathering and storage effects, plug sample selection, preparation in the laboratory, and the measurement techniques applied. Conventional core analyses are performed without the application ofsimulated overburden pressure, and horizontal permeabilities are measured in two directions at 90° to each other. The highest measured permeability is designated as "k.n,," and the other as "~o'" A practical approach in most situations is to assume that the ~o measurements more closely represent the in situ reservoir permeability than the k.nax readings for the following reasons: I. Small fractures induced during the coring procedure may result in an excessively high k m ax' particularly in limestones and dolomites. 2. Lack of overburden pressure usually results in high k.nax and ~o readings, particularly in poorly consolidated sandstones.
3. Core plugs represent extremely small samples of the reservoir rock and may provide higher or lower permeabilities than might be obtained if it were possible to use much larger samples, particularly in heterogeneous reservoirs or reservoirs characterized by vuggy porosity. In such cases, permeabilities derived from well production characteristics and pressure measurements may be more representative of in situ reservoir conditions. As a general rule, the larger the sample, the better the reservoir representation. As previously mentioned, however, if the core was oriented when it was originally cut, the k.nax and k"o permeabilities have greater importance and can be related to actual directions in the reservoir. In cases where clays contained in sandstone core samples have been dehydrated during the cleaning procedure, erroneously high values may be measured for both k.nax and ~o permeabilities. Glaze may be created by core bit action, particularly on limestone cores, and may obscure variations in core characteristics during visual inspection and result in unrepresentative sampling and in permeabilities that are much too low. Sandblasting is commonly used to remove the glaze during sample preparation. Logs should be consulted during the sample selection procedure.
5.11.3 Porosity from Cores Porosity measurements made on core samples are less subject to error than permeability measurements. However, incomplete cleaning during laboratory plug preparation may result in erroneously low porosity measurements. The laboratory techniques used to measure core porosities may affect the accuracy of the results. Table 5.4-I in Section 5.4.2 provides a comparison of techniques. Since most laboratory porosity measurements are obtained at surface conditions, the porosities are generally higher than in the reservoir, particularly in poorly consolidated sandstones, unless compressibility tests are conducted to provide reduction factors to allow for overburden pressures. It is worth repeating that the quality of the results obtained from core analyses is directly related to the quality and condition of the core when it reaches the laboratory. Therefore, in cutting and retrieving the core, precautions must be taken to preserve, as much as possible, the conditions that exist downhole in the reservoir. Cutting and retrieval of core to surface results in removal of overburden pressure, introduction of drilling
101
2
DETERMINATION OF OIL AND GASRESERVES
fines, and modification of clays, all of which can affect porosity measurements.
5.11.4 Saturations from Cores Because ofthe coring and retrieval procedures used for conventional cores, most laboratory saturation measurements obtained are unreliable. At best, the oil saturations obtained may provide a preliminary indication of what residual oil saturations might be after waterflooding. The water saturation measurements are usually meaningless. On the other hand, saturations measured on properly preserved core obtained under well-controlled field and reservoir conditions can give reliable results. .
about pore structure that can be related to connate water saturations. The results should be used with caution since the studies are generally performed on weathered core that has been cleaned and resaturated with other fluids. The results may differ considerably from those obtained from work on oil core using the Dean Stark extraction technique.
2. Keep coring fluid in a closed system to prevent water from being introduced inadvertently.
The wettability of reservoir rock is another important consideration that is often difficult to assess and may be altered by coring procedures, surface handling, and laboratory techniques. Generally speaking, the majority of reservoir rock encountered in the western Canada sedimentary basin is considered to be preferentially water-wet. However, the term "wettability" is the subject of much debate and is somewhat misleading in that it implies that it is a property ofthe reservoir rock that determines whether it is water-wet or oil-wet. Some parties hold that wettability of a reservoir rock depends on which fluid saturated the rock first. Others contend that wettability is a function of the rock, water, and hydrocarbon properties and their associated oxygen-carbon chains. In fact, most ofthe hydrocarbon reservoirs were initially deposited under marine environments where the initial saturating fluid, water, was subsequently partially displaced by hydrocarbons.
3. Analyze samples ofthe drilling fluid for water content at regular intervals during the coring procedure.
5.11.5 Effective Porous Zone and Net Pay from Cores
4. Avoid penetration of any underlying aquifer. 5. Preserve recovered core in lease crude, or suitably protect it from exposure to the atmosphere until it is ready for analysis in the laboratory.
Effective porous zone and effective net pay refer to that portion of the porous reservoir rock that has sufficient permeability to permit the flow of reservoir fluids. Porous rock, with permeability below a certain minimum level in conjunction with capillarity and relative permeability, will not allow the flow ofreservoir fluids, at least at rates significant in terms of production economics. These minimum permeability levels may differ depending on the production mechanism under consideration. In other words, effective net oil pay under a depletion drive mechanism, where expansion of reservoir fluids is the driving force, may be greater than under a waterflood, where the water displacing the oil tends to follow the path of least resistance and may by-pass low permeability reservoir rock. In practice, the connate water saturations related to minimum permeability levels tend to be in the range of 50 to 60 percent. Special core studies and conventional core analyses should be used to establish cutoffs of permeability and porosity below which the reservoir is considered noneffective for specified depletion and production regimes such as solution gas drive, water displacement, gas displacement, or miscible displacement. Once
Oil-base cores normally provide reservoir samples that should provide accurate measurements of connate water saturations. The coring fluid is usually lease crude or a specially prepared drilling mud that contains weighting material and only very small amounts of water in emulsion form. The following steps are necessary to ensure that uncontaminated core is obtained: I.
Set casing and cement to a point immediately above the target interval to ensure that the drilling fluid does not become contaminated with uphole fluids.
In Alberta, the Energy Resources Conservation Board (1993) publishes each year an updated PVT and Core Studies Index, which is useful in identifying the reservoirs that have had special core studies performed. However, caution should be used in selecting core analyses and special studies from this index since some cores identified as being obtained with oil-base muds were not obtained under properly controlled conditions. Frequently the muds used were oil-water emulsion muds containing high proportions of water. Another useful source of connate water saturation data and relationships derived from oil-base core analyses relating to reservoirs in the western Canada sedimentary basin is available in a paper published by Buckles (1965). In instances where suitable oil-base cores are unavailable, capillary pressure studies are performed on samples of conventional core to determine information
102
s
ESTIMATION OF VOLUMES OFHYDROCARBONS IN PLACE
these cutoffs have been established from core data, correlations (cross-plots of water saturation vs. permeability, water saturation vs. porosity, and permeability vs. porosity) with porosity logs may be developed, thereby permitting a relatively uniform and consistent approach to the selection of effective porous intervals and effective net pays throughout a given reservoir. This approach provides a logical and reliable basis for the creation of effective porous zone maps and net pay maps. Cutoffs that have been selected arbitrarily and applied inconsistently throughout a reservoir can lead to unrealistic estimates of in-place hydrocarbons, ultimate recoveries, and costly mistakes in reservoir management. In reservoirs where core information is unavailable, log porosity cutoff values may be determined from core information using nearby reservoirs as models or by applying generally accepted porosity cutoffvalues from the reservoir rock under consideration (2 to 4 percent for carbonates; 7 to 10 percent for sandstones). These porosity cutoffs are based on experience that shows that the corresponding horizontal permeability cutoffs are in the range of 0.5 to 1.0 mD, with the lower end of the range used for gas reservoirs and the higher end for oil reservoirs. It should be noted that these cutoffs are related to measurements made on unstressed core.
5.11.6 Porosity from Well Logs It is important to calibrate porosity logs using core data wherever possible since good core-derived data are generally considered to provide the best benchmarks. Cross-plots of log information should be used extensively in order to better characterize porosity.
Uncertainty in the detailed mineral composition of the reservoir rock and borehole rugosity can result in apparent log porosities that are much different than the true effective porosities. Bitumen and pyrobitumen contained in reservoir rock can greatly reduce effective porosity with contents ranging up to 30 percent of pore space in some reservoirs. The analyst should be attentive to all indications of the presence of bitumen and pyrobitumen and should make it a point to scrutinize well cuttings and core descriptions for evidence oftheir presence.
5.11.7 Water Saturations from Well Logs Water saturations determined from well logs may be influenced by the following: • Thin beds (averaging effect ofresistivity tools due to lack of vertical resolution)
• Conductive minerals in the reservoir rock • Selection of coefficient "a", cementation exponent " ' exponent "n" m " an d saturation • Resistivity of the formation water With regard to formation water resistivities, care should be taken to use measurements from samples considered to be representative of reservoir water uncontaminated by drilling mud filtrate. Water resistivities determined from log calculations over known wet intervals should be compared to those determined from water analyses wherever possible.
5.11.8 Effective Porous Zone and Net Pay from Well Logs As earlier discussions have stressed, the determination of realistic cutoff levels for permeabilities and porosities are critical to the selection oftruly effective porous zones and net pay from well logs. Errors in determining effective net pay may introduce order-of-magnitude errors in estimates of effective hydrocarbons in place and thus lead to costly consequences, particularly in the planning and implementation of enhanced recovery schemes. It is important to use all the data available to ensure the reliability of results. For instance, in situations where core information is lacking, correlation of mud cake build-up on caliper logs can give a good indication of the sections of the reservoir that have effective permeability and can be used in conjunction with porosity logs to select effective net pay. In summary, it cannot be emphasized too strongly that realistic cutoffs ofhorizontal permeability and porosity should be used in determining effective porous zone and, in conjunction with water saturations, net pays from cores and well logs. Unless unusual reservoir conditions exist, such as the presence of large concentrations of clays, horizontal permeability and porosity cutoffs generally correspond to connate water saturations greater than 50 to 60 percent. Based on empirical determinations derived from unstressed core measurements, generally accepted (minimum) cutoffs are as follows: Horizontal air 1.0 mD (medium to high gravityoils) permeability 0.5 mD (wet gas) 0.1 mD (dry gas under special circumstances) Porosity 2 to 4 percent (carbonates) 7 to 10percent (sandstones) 26 to 28 percent (heavyoil in sandstones) Watersaturation 50 . 60 percent
103
7
DETERMINATION OFOIL AND GAS RESERVES
The cutoffs selected for a given oil reservoir may vary widely depending on the type of depletion mechanism being contemplated. Fracture porosity in certain gas reservoirs may justify the use of lower porosity cutoffs (0.25 to 2 percent).
5.11.9 Drillstem Tests The drillstem test provides an indication of the fluids contained in a reservoir, a measure of the flow characteristics, and a reading of the reservoir pressure. The reservoir pressures measured during a drillstem test may be used to assess the quality and reliability ofthe test as follows:
1. The hydrostatic pressures measured should be checked to ensure that no significantpressure change took place during the test. Any significant change in hydrostatic pressure would indicate that a proper packer seal had not been obtained and that the zone being tested was not properly isolated from the wellbore fluid column. 2. The pressures measured on different gauges should be checked for consistency, particularly when more than one gauge is run at any position. Charts should be checked for evidence of tool plugging by comparing inside and outside gauge pressures.
5.11.10 Production Tests Good production tests are a function of the objectives of the test and the test design, as well as the reliability of the equipment used. From the design viewpoint, it is wise to use two downhole pressure recorders so that comparisons can be made to check for consistency. Surface and bottom-hole pressures should be compared, and any fluid level changes should be noted during buildups. Flow rates should be of sufficient duration to permit stable conditions to be reached. Generally, when the objective is to determine static reservoir pressure, buildup times should be at least twice as long as the preceding flow rates.
5.11.11 Reservoir Fluid Samples The sampling equipment and procedures are of utmost importance in obtaining representative reservoir fluid samples. Care must be taken to prevent contamination of the samples by ensuring that the sample containers are properly purged prior to sampling. Special containers should be used when collecting gas samples containing hydrogen sulphide. Proper well conditioning prior to obtaining subsurface oil samples is important, especially when more than one fluid phase is present in the wellbore.
5.11.12 Reservoir Temperature The most reliable source of reservoir temperature data is a bottom-hole temperature taken with a continuous recording subsurface temperature gauge under stabilized bottom-hole conditions, preferably in conjunction with a static bottom-hole pressure measurement. Other methods, such as using maximum reading thermometers during testing or logging operations, are considered to be less reliable. Small variations in bottom-hole temperature, when converted to absolute temperature, generally result in only a very small percentage error in the overall reserves estimates; nonetheless, care should be taken to obtain the best measurements possible. As a general rule, temperatures measured in wells will tend to understate true reservoir temperature because temperature equilibrium has not yet been reached in the wellbore. However, in certain situations such as in shallow wells on warm days, maximum reading thermometers may reflect the high atmospheric temperature on the day the measurement was made.
5.11.13 Reservoir Pressure Accurate static reservoir pressures are extremely important in the determination ofhydrocarbon reserves. The accuracyof the measurements is a function of the following: • The type of measurements being made: surface or bottom-hole • The reliability of the recorders used for pressure measurements • The duration of the shut-in period Bottom-hole measurements are considered to provide more reliable measurements of reservoir pressure than surface measurements, which must then be extrapolated to bottom-hole conditions. Usually tandem recorders are run to allow comparisons and verification of the accuracy of the measurements as well as to ensure that at least one pressure is obtained should one of the recorders fail. The duration of the shut-in period is critical in obtaining reliable pressure information and should be a function of the quality (permeability) of the reservoir rock and the fluids that occupy it. The poorer the reservoir quality and the higher the viscosities of the occupying fluids, the longer the shut-in period.
104
s
ESTIMATION OFVOLUMES OF HYDROCARBONS INPLACE
5.11.14 Gas Compressibility Factor
5.11.17 Interrelationships
Reliable gas compressibility factors are dependent on the quality ofthe gas analyses being used and how representative they are of the reservoir fluids. Since a compressibility factor is only correct at the pressure and temperature used in the estimation, it is important to use reservoir pressure and temperature data of acceptable quality. With gases containing carbon dioxide or hydrogen sulphide, large errors can be introduced into the reserves estimates if the appropriate sour gas corrections are not made in estimating gas compressibility factors.
The interrelationships and interdependencies of the various parameters are important in arriving at reliable estimates of in-place volumes. Ifthe sources ofthe data are reliable, then the quality of the resulting estimates can be improved by a consistent approach in the selection ofparameters. For example, in a particular reservoir where reliable oil core data are available, water saturations can be plotted vs. porosities and vs. horizontal permeabilities. In tum, porosities can be plotted vs. horizontal permeabilities. Minimum porosity and horizontal permeability cutoff values can then be selected that are consistent with a selected water saturation cutoff value, say 50 percent. The porosity and horizontal permeability cutoffs established from the oil core data can then be applied to conventional core data. The combined corederived information can then be used to calibrate well logs and additional interrelationships established through the use of further cross-plots. This approach ensures that the data derived from one source is consistent in its application with data from another source. Not only does interrelating the various parameters introduce consistency to the estimates of in-place volumes, it also provides a sound basis for the application of recovery factors. Care should be taken to ensure that the applied recovery factors are consistent with the inplace volumes and cutoff values used. For example, negotiated in-place volumes, such as those resulting from unitization negotiations, should be used with great caution since they are not necessarily an accurate and consistent representation ofthe reservoir.
5.11.15 Formation Volume Factor The quality and reliability of formation volume factor data are dependent upon whether or not the reservoir fluid samples from which the data were obtained are truly representative of reservoir conditions. Proper selection of flash vs. differential formation volume factors is required to best represent reservoir mechanisms.
5.11.16 Material Balance Errors in material balance calculations generally fall into the following categories: I. Thermodynamic equilibrium not attained in actual field conditions 2. PVT data obtained using liberation processes that do not represent reservoir condition mechanisms 3. Inappropriate average pressures 4. Uncertainty in the material balance "m" ratio 5. Inaccurate production data 6. Inability to recognize the presence of an exterior source of energy, such as an aquifer The amount of pressure decline covered by the production history is one of the best criteria in gauging potential errors. The material balance is a comparison of voidage to expansion. It concentrates on evaluating fluid expansion. Large pressure declines produce large expansions, making inaccuracies in production volumes relatively less significant. Similarly, pressure errors are less critical with more pressure depletion. In general, a pressure decline of 10 percent of the original reservoir pressure is needed before the material balance becomes reliable. This critical depletion level is highly dependent upon the quality of pressure, production, and PVT data.
References Buckles, R.S. 1965. "Correlating and Averaging Connate Water Saturation Data." JCPT, Jan. Mar. 1965, pp. 42-52. Energy Resources Conservation Board. 1993. PVT and Core Studies Index. Guide G 14, Calgary, AB.
105
7
Chapter 6
PROBABILITY ANALYSIS FOR ESTIMATES OF HYDROCARBONS IN PLACE
6.1
INTRODUCTION
Chapter 5 has addressed the importance of, and the challenges involved in, obtaining accurate and reliable measurements from samples. This in itself is difficult enough, but there is a "fact of life" in the petroleum industry that further complicates the volumetric estimation procedure: petroleum reservoirs are heterogenous, so parameter values vary from sample to sample even when they are correctly measured. The variation might be handled by using many sample measurements and statistical techniques, but the cost of obtaining samples is so expensive that there can never be even close to a sufficient number of samples. To gain an appreciation for the magnitude of the shortfall, imagine trying to predict the outcomes of political elections from the opinion of one voter. If this sounds ludicrous, it is; yet a one voter sample is a larger percentage of the total population than the reserve estimator can realistically hope for. Several approaches to this dilemma have been tried over the years and each has its shortcomings. This chapter is about searching for a better way. Industry's historical approach has been to "guesstimate" a single best value for each parameter, resulting in a single value for the volumetric estimate. This sounds easy until it is tried. There is never enough information to justify the selected value. Getting a second opinion does not help because no two people will calculate the same volumetric number, and some data will exist to support and contradict both. Thus two technically competent people can have very different opinions and either or both could be right or wrong. The industry's current practice of multi-disciplinary group or team decision-making compounds the problem because the multiple opinions will almost certainly be different. Achieving group consensus is rarely possible because no "right" answer can be determined, and even consensus does not guarantee truth; in fact, it can provide a false sense of security that may collectively lead all those involved into subsequent contradictions
between the agreed-to perceptions and reality, and thus introduce the potential for unfortunate consequences. The single-value approach presents some organizational disadvantages as well. In order to arrive at the "right" number there are only three ways to handle differences of opinion:
I. Some people have to concede they are wrong, despite evidence that suggests they might be right. 2. The group goes into an endless analysis mode and never determines the "right" answer. 3. Dissenting opinions are overruled and ignored. This could hardly be called good team buildingimagine the confrontations generated and the feelings of the participants! Is it appropriate that, after a certain amount of discussion, the debate is often adjudicated at a higher level of the organizational heirachy? The potential for bias in the assessment likely increases as the debate proceeds to higher-authority levels, because each successive level is less familiar with the technical details, but more cognizant of the impact that a particular number will have on current plans and operations. Sometimes there may even be personal implications, as with management whose performance assessment is directly linked to the reserves booked for the year, or the consultant whose opportunities for further work may depend directly on the magnitude of the reserve estimate. A final criticism of the single-value approach is that at the conclusion of the exercise the participants are expected to be fully committed to the resulting decisions and to work together to implement them. This is not a reasonable expectation for a process that is essentially adversarial as winners and losers seldom work well together. In spite of the high organizational costs and the low probability of achieving a reliable, accurate estimate, the single-value approach is still used. A plausible explanation is that the industry is unaware of a better alternative. Other approaches have been tried over the years, including those listed in Table 6.1-1. Of the
106
S'
PROBABILITY ANALYSIS FOR ESTIMATES OFHYDROCARBONS IN PLACE
approaches listed, the Warren Method is the most workable, but it is not widely used at the present time. The method was developed by a pioneer in the application of probabilistic methods to the oil industry, Dr. Joseph E. Warren (1988). It works because it: • Is applicable to volumetric theory • Provides a means to deal effectively with varying amounts of indirect data that may, at times, seem overwhelming in volume but are always incomplete and insufficient to support a purely statistical analysis, or justify a single number as the right answer Quantifies the uncertainty in the estimate by separating the range ofprobable values from the much larger range of possible values • Is applicable to all stages of evaluation, from initial assessment of basin potential to individual pool development • Is sufficiently flexible to incorporate all the available data, which can differ for every pool and at each evaluation stage • Is quick, easy and inexpensive to use
The Warren Method is simple enough that it can be used on a personal computer or even a hand-held calculator, and successive iterations are actually easier to perform than the initial calculation. These features enable the focus to be on the quality of the input data rather than on the arithmetic, and they encourage its use. In addition, the method can be extended to provide reserve and net present value estimates, while dry-hole risk can be accommodated in the pre-drilling evaluations.
6.2
WARREN METHOD THEORY
The Warren Method is based on the following combination of theory an~ assumptions: I.
It has been proven that the product of unimodal random variables is log-normally distributed as the number ofvariables approaches infinity (Aitcheson and Brown, 1966, Theorems 2.8 and 2.9), and that the product oftwo or more log-normal distributions is a log-normal distribution (Theorems 2.2 and 2.3). This theory and its application in analogous situations, plus the tests on artificial samples, suggest that the volumetric hydrocarbon-in-place estimate may be approximated as a log-normal
Table 6.1-1 In-Place Volumetric Estimation Techniques Methodology
Comments
Single-Value Estimate
No satisfactory way to select the "right"value for each parameter in the volumetric equation. No way to resolve differing opinions on prospect potential. Cannot quantify uncertainty in in-place estimate nor the probability of occurrence.
Absolute Minimum! Maximum Value Approach
Consistent use of minimum!maximum parameter values to calculate absolute minimum! maximum hydrocarbonsin place yields a minimum value that is uneconomic and a maximum value that is too good to be true. Range is too large to be of practical use. No way to separate range of probable values worth considering from the much larger possible value range.
Statistical Analysis
Generally insufficient samples to develop in-place distribution for the total population from the sample population. Drilling best prospects first biases sampling, yielding optimistic predictions if sample results are extrapolatedto total population. Sufficient samples for a play are usually available once the explorationist has run out of prospects. For a given pool, they are available after the pool has been developed. The timing is unacceptable for both.
Monte Carlo Computer Simulation
Developmentof in-place probability distribution is a significant advancement over previous methods. The dilemma is how to model parameter distributions. Development of in-place distribution from multiple single value calculations is computer-intensive and time-consuming. The method tends to be too cumbersome to accommodate iteration requests and time constraints.
Warren's Probability Analysis
Yields similar solutions as Monte Carlo computer simulation in less time and at lower costs.
107
--DETERMINATION OFOILANDGASRESERVES
distribution because it is the product of successive multiplications. The characteristics of the hydrocarbon-in-place (HCIP) distribution may be calculated from the moments of the parameter distributions as follows:
maximum error of 2.05 percent in the tests, while average absolute and maximum errors for the variance were 2.2 percent and -7.5 percent respectively. The recommendations balance the need for accuracy with the need for simplicity in the estimation procedure. In particular, a formula utilizing the mode rather than the median of the distribution was chosen to estimate the mean (it is easier to estimate the mode than the median). The distribution moments are calculated from the minimum, most likely and maximum values for the distribution using the following equations:
M, (HCIP) = m, (x.) x m, (x.) x m, (x.) x ... (I) M, (HCIP) = m, (x,) x m, (x,) x m, (x.) x ... (2)
"
Rio = M, (HCIP) ~,.
(3) m, (x)
I
= --
2.95
(4) where
(x mi, + .95 x pm b + xm,,)
m, (x) = m~ (x) + z
a
M, (HCIP)
= In ----=--'--------'M~ (HCIP)
M,(HCIP) = first moment of hydrocarbon in-place distribution MiHCIP) = second moment of hydrocarbon in-place distribution m.Ix,...) = first moment of the nth parameter distribution mixn ••• ) = second moment of the nth parameter distribution x i- •• = parameter distribution (<\>, h, A ...) R so = median value of the in-place distribution; plotted at the 50th percentile on log-probability paper R s4., = median value plus one standard deviation; plotted at the 84.1th percentile 2. A three-point approximation can be used to estimate parameter distribution moments in the absence of complete knowledge of the continuous distribution. This assumption, which is also integral to Monte Carlo simulation, is necessary because a complete knowledge is seldom, if ever, available. It is supported by the work of Keefer and Bodily (1983), who compared the accuracy ofseveral threepoint approximations in estimating the means and variances for a set of beta distributions. The recommended approximation for the mean had an average absolute error of 0.37 percent and a
where
xmin
x
[
m"
-x
3.25
' J
mi,
(5)
(6)
= minimum parameter value (probability = 0.05)
xprob xmax =
most likely parameter value (mode) maximum parameter value (probability = 0.95)
3. Consistent, reliable, unbiased 3-point estimates can be developed. This assumption, which is also necessary to Monte Carlo simulation, may appear daunting. Capen's (i 976) hypothesis that SPE members will "miss" an average 68 percent of the questions and the results that support the hypothesis are a sobering assessment of the industry's present inability to deal with uncertainty. However, Capen suggests that the skills necessary to provide reliable estimates may be developed with practice, and he offers some practice techniques. He notes that some meteorologists have apparently mastered this skill and suggests that oil industry personnel can eventually attain a similar proficiency. In practice the assumptions are applied in the reverse order listed.
6.3
APPLICATION
The first step in estimating the in-place hydrocarbons of a pool is the development of value ranges for each of the parameters in the volumetric equation. The minimum and maximum values establish the range for the pool average value by distinguishing between what is and what is not within the realm of possibility. It is crucial that the true average value for the pool be
108
s
PROBABILITY ANALYSIS FOR ESTIMATES OF HYDROCARBONS IN PLACE
greater than the minimum value and less than the maximum value. However, if unrealistic minimum or maximum values are used, the variation in the in-place distribution will be so large that the estimate will have no practical use. For example, the minimum average pool porosity value must be slightly greater than the cutoffvalue for the rock type or the discovery well could not have produced hydrocarbons on the drillstem test. Using the cutoff value as the minimum average pool value is probably acceptable, but using zero as the minimum average value is not. Similarly, assuming the well flowed gas on the test, the residual gas saturation value might be ~n acceptable approximation for the minimum pool average gas saturation value. Values of zero and one are always too extreme when estimating the volume of hydrocarbons in a pool because they imply that no hydrocarbons exist, contrary to the production from the pool. Warren's methodology does permit an evaluation of "dry hole risk," but the topic is beyond the scope of this discussion. The most likely value or modal value is the "pool parameter average value with the highest frequency of occurrence." By definition, it is greater than the minimum value and less than the maximum value. A suggested interpretation is the "best guess" for the pool average value. Several iterations with different best guesses usually demonstrate that the in-place distribution is relatively insensitive to variation in the most probable value. Because the in-place distribution can be calculated so easily, iteration using all the potential probable values is often the quickest and easiest way to resolve which value should be used for this parameter. In the absence of sufficient measurements, the source for parameter values is the combined training and experience ofa company's earth science personnel. The multi-discipline team approach to in-place estimates provides some desirable features. It brings a higher level of competency to the parameter estimates than can be supplied by anyone discipline working in isolation, and the inter-discipline discussion tends to highlight any individual bias or inconsistency that may exist in the evaluation. For a given prospect, the objective is to identify the models that do not apply, based on the available information, and then develop unbiased parameter value ranges encompassing all the models that may apply to this particular prospect. A multi-discipline team that appreciates the unique viewpoints of its individual members and works to include all views in a consistent explanation has a definite advantage in accomplishing this task.
Possibly the most challenging part of the evaluation is incorporating parameter interdependence into the volumetric equation. Team members may agree that a dependence exists, but that the relationship is vague or unknown. An apparent impasse in the discussion usually signals that the team is grappling with a dependency. This is especially obvious when individuals are basing their estimates of one parameter on their estimates of a previous parameter. Because each case is unique, a single solution applicable to all cases does not exist. Resolution requires flexibility and at least one team member with the ability to postulate the mathematics ofthe dependence from the discussion. The guiding principles are as follows:
1. Deal with only one geologic process at a time. 2.
Prevent the team from estimating parameters for which they have no direct measurements.
3.
Ifa parameter can be identified as a product of other parameters, estimate the primary parameters and substitute them into the volumetric equation.
4.
When one parameter is clearly dependent upon another, substitute the dependency into the volumetric equation to minimize the number ofparameter estimates required from the team, and thus reduce the chance ofinconsistencies creeping into the estimate.
An example ofthe application ofthese guidelines is the estimation of pool rock volumes. The rock volume should not be guessed at directly because the pool rock volume is never measured. Teams often find it easier to approximate the rock volume as a combination of geometric shapes and estimate the dimensions for each shape. For example, a rectangle-triangle combination might be used to approximate a reef cross-section (Figure 6.3-1). The rectangle represents the reef crest while the two triangles represent the reeffront and back slopes. Now the team's expertise can be used to provide estimates for gross thickness, H, crest width, W, reeflength, L, and slope angles, X. Angle ofrepose controls the front slope, while regional dip is the primary influence on the back slope. This information, plus the equations for triangular and rectangular areas, yields the cross-sectional area of the reef. Multiplication by both the ratio ofnet pay to gross thickness and the reeflength yields the volume of the reef considered to contain hydrocarbons. The dependency between cutoff values and pool average values for porosity and net pay deserves mention. Increasing the cutoff value decreases the net pay value, but increases the average porosity value. The
109
DETERMINATION OFOIL AND GAS RESERVES
Reef Back Slcpe /'
Reef Front Slope
/' /' /'
/' /' /' / / / / /
/'
/
/
/'
/'
-W-
Underlying Water Hydrocarbon Bearing Rock Volume
= (Area Back Slope + Area Crest + Area Front Slope) (Length) (Net/Gross Pay Ratio) 2
2
= (0.5 H + HW + 0.5 H ) (L) (Net/Gross Pay Ratio) tan X, tan Xb
Figure 6.3-1 Estimation of Reef Volume recommended method of addressing this issue is to calculate in-place distributions for each of the parameter value combinations corresponding to the different cutoffvalues. Iteration usually demonstrates that the different combinations yield essentially the same in-place distribution.
6.4
TYPICAL SITUATION: CONVENTIONAL GAS
The example described in the next few pages is typical ofmany ofthe situations encountered. It serves to show just how far afield one can go if insufficient attention is paid to the uncertainty in the in-place estimate. To give this example some reality and illustrate the economic utility of the method, typical recovery and economic factors are assumed; however, in practice, equal attention is paid to developing the range of all parameters. A recent carbonate discovery well flowed gas at a rate of 225 x 103 m3/d from 10 m oflogged pay following completion. Log-derived porosity and water saturation values are 0.13 and 0.205 respectively. Movable water was not interpreted to exist in the pay interval. The formation temperature during logging was 74°C. Core is not available from this well. Based on the interpretation of the single rate flow and buildup data, the well is completed in a single porosity reservoir, with 300 mD-m conductivity and a -2 skin factor. Bottom-hole formation temperature recorded during the buildup stabilized at 81°C. The Homer plot extrapolation yielded a value of 24 731 kPa(abs).
110
The gas deviation factor is calculated from the gas composition as 0.88 at a temperature of81 °C and a pressure of 24 731 kPa. Radius of investigation calculations yield an investigation area of 56 hectares. A single boundary is interpreted to exist at a distance of 266 m from the well. This correlates with the seismic interpretation, which located the eastern edge of the structure 200 to 350 m from the wellbore. No insight on the location of the other edges is available from the well test or the seismic interpretation. Geological interpretation provides the location of the remaining edges, which are inferred from the depositional model, dip angle, and offset well data. Post-depositional erosion results in a very steep-sided structure. Subsequent infilling of these erosional channels with impermeable material provides the trapping mechanism for the structure. From these interpretations, the maximum areal extent of the pool is four sections (Figure 6.4-1). The geologist has also interpolated a most likely value, covering about 2.25 sections. The basis for this contour is solely the assumption that the true value is likely nearer the mid-point than the extremes. The Exploration Department is rumored to be contemplating a bid in excess of $3000/ha for the offsetting acreage. Justification seems to be the four-section upside potential ofup to 4900 x 106 m3 ofreserves. The Exploration Department's request for review of their numbers has just been received. The sale will take place at the end of the week.
PROBABILITY ANALYSIS FOR ESTIMATES OFHYDROCARBONS IN PLACE
2000 x 106 m3 and an economic hurdle volume of 120 x 106 m3 •
Optimizing porosity. area, recovery factor, etc., 3 indicates up to 4900 X 10' m of recoverable gas. Geologist's most likeiy contour
.' -"---V/\ p
.,. ..
-------\--.
,
,,
,, ,,
p:, ,, ,, , , ,,,
"*-
.
, ,,,
,,
~
-- -------- --- ..
How will you proceed?
,
, ,,, I ,
P/'
,
,
;
, .: ' . /
j;/ .
Your boss just "volunteered" you to resolve the situation. In addition, he advises that senior management would like to review the corporate reserves booked against this well, plus production and cash flow forecasts at the upcoming quarterly review. The press reports described the well in glowing terms, "possibly the best discovery ofthe year" and-you agree-it can't just sit there.
Behind the Numbers
'\
\ \\
A completion test has shown excellentreservoir with one adjacent boundary.
Figure 6.4-1 Typical Situation: Gas Pool Map The Production Department apparently has no plans to tie in this well at the present time. Volumetric estimates using minimum parameter values yield in-place hydrocarbons of 44 x 106 m'. The economic hurdle volume for the tie-in is 250 x 106 m30freserves. Environmental concerns preclude flaring additional gas volumes to perform an economic limits test. Review of the four analogous pools reveals various stages of depletion, with pool reserves estimated at 55, 120,250, and 550 x 106 m'. Cumulative production from the seven wells in these pools ranges from 30 x ]06 m3 to 300 x 106 m3 • This statistical review did not persuade the Production Department to tie in the subject well, but raised questions concerning the size of facilities required. In addition, the Production Department advises that they recently abandoned the lone well in the 55 x 106 m3 pool, due to reservoir depletion. Several reviews with increasingly senior levels of management have resulted because the well tie-in costs were not recouped, and Production's personnel are anxious to avoid any future recurrence. They note that the recently abandoned well also demonstrated a commercial flow capability following completion and had an upside potential of
The situation may seem tense but it is not hopeless! Although some sabres are rattling, your boss's insight ge~s you in while the majority are still willing to listen. Pnvately, both departments confide that their numbers are not absolute but ... The ultimate purposes of reserve estimates are as follows: 1. To assess whether the uncertainty in the reserves of a given prospect is of sufficient magnitude to justify the expenditures required to reduce the uncertainty 2. To assess the safety of the prospect and of the aggregate from an investment viewpoint 3. To provide an indicator of aggregate performance Ea~h of these different purposes requires a unique estimate for the prospect. Additionally, there are times when the prospect estimate is less important than its impact on an aggregate ofreserve estimates, such as the company reserve profile. An understanding of the responsibilities and COncerns of the different groups and their relation to the prospect or the aggregate is vital to resolving this situation. Erroneous conclusions with potentially disastrous consequences can result when an estimate intended for one purpose is misused for another.
In this case, the Production Department is charged with the responsibility for tying in the well. The concerns that relate directly to the prospect estimate are the size and design of the surface facilities, the type of sales contract to negotiate, and the likelihood that the tie-in will be economic. Budget allocation requires that the economic potential of this prospect be compared and ranked relative to the other financial opportunities available to the department. This is an aggregate-related issue because the focus now is on the cumulative outcome for the budget period and the effect on the
111
7
DETERMINATION OF OIL AND GAS RESERVES
overall performance of both the department and the company. The company's future depends on continued access to economic sources ofproduction which, in this example, is the responsibility ofthe Exploration Department. One way to access new production sources is via the bidding process. To be successful, the bid price must exceed all competitive bids, but it must also be less than the net economic value of the reserves acquired. The consequences of bidding too low or acquiring the prospect at an uneconomic price are equally undesirable. This prospect-related issue can be addressed by comparing the likelihood of exceeding the prospect economic value for a given bid price to the likelihood of acquiring the prospect at that price. The aggregate issue is again budget allocation and the impact of this opportunity on overall performance. The issues at the corporate level tend to be aggregaterelated. Both internal and external comparisons to established criteria are performed at this level, underscoring the need for aggregate reliability and consistency throughout the industry. Reliability is required to establish trust in future projections, and is achieved when past performance essentially agrees with past projections over some time period. Consistency is necessary to permit comparison. Comparisons may be between producing horizons, between geographic areas, between departments within a company, between companies, or even between industries. Comparison criteria are usually economic and incorporate required or desired objectives. An example of a required objective might be the time component in sales contracts, security of supply issues, or possibly safety and environmental issues. This section demonstrates that calculating the in-place hydrocarbon distribution cannot satisfy these concerns directly. The calculation is only a necessary first step in developing solutions that avoid disaster while achieving acceptable results for at least the majority of the probable reserve outcomes. This approach is based on the concept that a solution that avoids disaster under all probable outcomes is preferable to one that performs nearly ideally under a narrow range of conditions, but provides unacceptable results the majority of the time. The optimal solution is the one that avoids disaster for all probable outcomes while maximizing the desired results over the widest range ofprobable outcomes. For an individual prospect, this requires consideration ofthe probable range of outcomes available to the prospect, while an aggregate question requires consideration of the probable variation in the aggregate.
'1!
Pool Parameter Values Congratulations! You've established sufficient trust that representatives from both departments have agreed to meet with you for the purpose of establishing parameter value ranges. An unexpected break is the attendance of two people whom you've successfully worked with before. Several intense discussions prove fruitful and produce group consensus on the following parameters. Areal Extent
By group consensus, the pool areal extent must be greater than 56 hectares. The most conservative guess is 64 hectares, which is deemed to be the minimum possible value. The maximum possible value is quickly set at 1024 hectares, but opinions on the most likely value range from 1.5 to 2.75 sections. Resolution is reached when you offer to run three cases, using 384, 576, and 704 hectares as the most likely value, to demonstrate the impact on the in-place distribution. Discussion of deposition and erosional processes, seismic interpretation plus several pictures of badlands terrain produces consensus that the area ofthe top ofthe pool is less than the area of the base. Opinions range from 60 to 95 percent ofthe basal area, with 80 percent as the most likely value. The inter-relationship is handled using an average area, which is equal to 0.5 (base area + top area). Substituting these percentages into the equation yields the average area equal to 80, 90 and 97.5 percent ofthe base area respectively (Figure 6.4-2). These percentages and the basal area estimates are substituted for the average area in the in-place distribution calculation. Net Pay
Discussion quickly identifies that for this case the pool's net pay is the product of two geological processes. The gross thickness of the rock is controlled by deposition and erosion, but not all of the rock is reservoir quality. Only the portion whose porosity and permeability has been enhanced via post-depositional processes is considered to contain hydrocarbons. This interrelationship is handled by first estimating the gross pay of the pool and then the percentage conversion to reservoir rock. Net pay is the product of the two parameters. The gross pay thickness is controlled by topographical variation on the upper erosional surface and the elevation ofthe gas-water contact, ifone exists, on the bottom. In the worst case, free water exists just below the bottom of the discovery well and, in the best case, it is not present. Free water is known to exist in two of the analogous pools, but at different elevations. This is
112
<
PROBABILITY ANALYSIS FOR ESTIMATES OFHYDROCARBONS IN PLACE
For the minimum case, Atop = 0.6 Abo" AOVg = 0.5 (A,op + Abo,,)
the most likely at 15 percent. Water saturation estimates are 18, 20 and 22 percent respectively. (7) (8)
substituting (7) in (8) Aovg = 0.5 (0.6 Abos• + Abos.) 0.5 (1.6 Abos.) 0.8 Abos• Similarly for 0.8 Abos• Atop = Aavg = 0.9 Abas• and when Atop
Aavg
= =
0.95 Abos• 0.975 Abos•
Figure 6.4-2 Conversion of Base Area to Average Pool Area
consistent with the theory that the hydrocarbons were locally sourced. From this and the 0.98 degree regional dip angle, the gross pay thickness estimates are 6.5 m as the minimum case, 19 m as the maximum, and 15 m as the most likely value. At 13 m, the gross pay thickness for the discovery well is slightly below the pool average and came in about 2 m lower than expected. A number of possible mechanisms are discussed for conversion of limestone to dolomite, none of which are definitive. In the end the estimates are based on the group's experience with the region, gained from the examination of logs and core from this formation over the entire geological basin. Based on that experience, the rock encountered by the discovery well is about average in terms of converting gross pay to net pay. The conversion efficiency for the pool is estimated at 65, 80, and 90 percent, respectively. Porosity and Gas Saturation
Regional experience again comes to the forefront in the estimation of these parameters. The question of bitumen infilling of the available porosity arises but is considered remote, based on the group's experience with this formation. The group also considers the possibility that porosity and water saturation are interrelated, but postulated correlations prove inconclusive. However it is agreed that the greatest variation in the in-place estimate results from the independent treatment of the two parameters, so value ranges are developed accordingly. The minimum possible pool porosity is estimated at 12 percent, the maximum at 17 percent and
Pressure
The initial reservoir pressure is uncertain. The Homer plot gives an extrapolated pressure of 24 731 kPa (abs) from the buildup, but this is not the initial reservoir pressure because the boundary's presence violates the requirement for infinite acting radial flow. Regional pressure gradients suggest an initial pressure of 22 000 to 26000 kPa (abs). The group agrees that the minimum possible pressure is 24 000 kPa (abs) because the pressure was still building at the end of the buildup period, with a final value of 23 966 kPa (abs). A maximum value of26 000 kPa (abs) is assumed, with a most likely value of24 700 kPa (abs). Temperature and Gas Deviation Factor
Recorded bottom-hole temperatures during the buildup ranged from 80.97 to 81.25°C. This variation is very small relative to the uncertainty in the other parameters. Perhaps parameters with less than I percent difference between the minimum and maximum values can be treated as a constant without significantly affecting the in-place distribution? The effect can be observed by first considering temperature as constant at 81°C, and then as a parameter, with values of 80.97, 81 and 81.25°C. The gas deviation factor varies from 0.87 to 0.89 over the 24 000 to 26 000 kPa (abs) pressure range. The variation between the minimum and maximum value is less than 3 percent, so perhaps it too can be treated as a constant? The incentive for doing so is that the increased accuracy achieved by incorporating the gas deviation factor's dependency on temperature, pressure and gas composition into the calculation(s) may not be worth the effort. Since the gas deviation factor is actually something between a constant and a random variable, the validity ofthe assumption might be confirmed by considering the impact of the two extremes on the in-place distribution. Values of 0.87, 0.88 and 0.89 were used for the parameter range, while 0.88 was selected when the gas deviation factor was considered constant. Gas In Place
In-place distribution calculations using 384, 576 and 704 hectares as the most likely value for areal extent are presented in Tables 6.4-1, 6.4-2, and 6.4-3. For a constant, m, (x) = constant and m2 (x) = constant squared. The in-place distribution is obtained using the calculated R so and Rs4.t values to establish a straight line on log-probability paper (Figure 6.4-3). For
113
, DETERMINAnON OF OIL AND GAS RESERVES
Table 6.4-1
Gas-in-Place Distribution for Most Likely Area of 384 Hectares Most Likely Value xprob
Minimum Possible Value xmin
Pool Parameters
Basalarea, Abo" (ha) Correction to avg. area, C, Grosspay, H(m) Net/gross pay ratio, N/G Porosity,
64 0.80 6.5 0.65 0.12 0.78 24000
(l - Sw)
Pi(kPa abs) Constants
384 0.90 15 0.80 0.15 0.80 24700
Maximum Possible Value xMs x
m. (x)
m, (x)
1024 0.975 19 0.90 0.17 0.82 26000
492.5 0.8915 13.47 0.7831 0.1466 0.80 24903
329783 0.7977 196.4 0.6191 0.02173 0.6402 620557523
0.000091
8.327 x 10-'
10 000 (288.16)
10000(288.16)
10' (101.325) r,z,
10' (l 01.325) (354.16) (.88)
M1 (OGIP)= 1234.7
OGIP=
I
10000 (288.16) Ab", (C,) H (N/GH (l-Sw) Pi 10' (101.325) Ti Z, 1
m, (x) = - - (x mio + .95 x pmb + xm,,) 2.95
m, (x) = m; (x) +
[X
max -
J
x min
3.25
M 1(OGIP) = m, (x,) X m, (x,) X m, (x.) X ...
a' = In
M,(OGIP) = 2 298 761
M, (OGIP) 0.4107
M; (OGIP)
-"-
,
R,o=M,(OGIP)e' = 1005.5 x 10 m' a , , R 84 . , = R so e = 1908.6 x 10 m
M, (OGIP) = m, (x.) X m, (x,) X rn, (x,) X .•.
Note: Pi. Ti• and Zj are respectively initial reservoir pressure, temperature, and gas formation factor.
prospect issues the question is: How much of the distribution should be considered? The suggested range is all values from the R, to R,s values, which can be read from the graph. For this example the range is 400 to 3250 X 106 m 3 using the 576 hectare most likely value distribution. This encompasses 90 percent of the probable outcomes and is consistent with developing solutions that work the vast majority of the time. Since human nature is inclined to over-estimate the extent of knowledge, an initial reaction might be disbelief at the magnitude of the range. However, an order-of-magnitude variation in the range is common, especially for new discoveries. For situations where a single number is desired to describe the distribution, the mean value (M1(HCIP» is recommended. For the 576 hectare distribution M 1 (HCIP) = 1389.7 X 10 6 m 3 • This value has no significance to prospect issues, only to aggregate questions. Misuse it at your own peril!
The effect of varying the most likely value of the distribution can be seen on Figure 6.4-3. In this example this variation is insignificant compared to the Rs to R,s range in the distribution. Group consensus on which value to use is usually easy to obtain following the team's inspection ofthe graph because it does not really matter which distribution is used. However, if consensus does not exist, a further compromise is to draw a line through the smallest R, value and the largest R,s value to establish a composite in-place distribution. The characteristics ofthis distribution can be calculated by reading the R so and R 84. 1 values from the graph and using the equations to calculate M,(HCIP) and M 2(HCIP). Alternatively, one can carry the two extreme distributions through the decision-making process until everyone agrees that "it does not matter."
114
d
PROBABILITY ANALYSIS FOR ESTIMATES OFHYDROCARBONS INPLACE
Table 6.4-2 Gas-in-Place Distribution for Most Likely Area of 576 Hectares Pool Parameters
Basal area, Ab,,,, (ha) Correction to avg. area, C f Gross pay. H(m) Net/gross pay ratio, NIG Porosity, (I - Sw) P, (kPa abs) Constants
Minimum Possible Value
Most Likely Value
Maximum Possible Value
X m1n
x pr ob
XMax
64 0.80 6.5 0.65 0.12 0.78 24000
576 0.90 15 0.80 0.15 0.80 24700
m, (x)
m, (x)
554.3 0.8915 13.47 0.7831 0.1466 0.80 24903
394506 0.7977 196.4 0.6191 0.02173 0.6402 620557523
0.000091
8.327 x 10'·
1024 0.975 19 0.90 0.17 0.82 26000
10000 (288.16)
10 000 (288.16)
10' (10 1.325) T, Z,
10' (101.325) (354.16) (.88) M, (OGIP) = 1389.7
OG1P=
10000 (288.16) Ab" , (C,) H (NIGH (I-Sw) P, 10' (101.325) T, Z,
a'
- -
6
R so = M, (OGIP) e' = 1164.7 x 10 m
a' =In
M, (OGIP) = 0.3533 M: (OGIP)
a
3
MlOGIP) = 2 749 913
, ,
R 84.l =Rsoe =2110.4x 10 m
Note: Pi. Til and Z, are respectively initialreservoirpressure, temperature, and gas formation factor.
Table 6.4-3 Gas-in-Place Distribution for Most Likely Area of 704 Hectares Pool Parameters
Basal area, Ab,,,,(ha) Correction to avg. area, C, Gross pay, H(m) Net/gross pay ratio, NIG Porosity, (1 - Sw) P, (kPa abs) Constants
Minimum Possible Value
Most Likely Value
Maximum Possible Value
x min
x prob
x Max
m. (x)
m, (x)
64 0.80 6.5 0.65 0.12 0.78 24000
704 0.90 15 0.80 0.15 0.80 24700
1024 0.975 19 0.90 0.17 0.82 26000
595.5 0.8915 13.47 0.7831 0.1466 0.80 24903
441 902.6 0.7977 196.4 0.6191 0.02173 0.6402 620557523
0.000091
8.327 x 10'·
10 000 (288.16)
10 000 (288.16)
10' (101.325) T, Z,
10' (101.325) (354.16) (.88)
=
M, (OGIP) = 1493.1
OG1P=
10000 (288.16) Ab" , (C,) H (NIGH (l-Sw) P, 10' (101.325) T, Z,
R so
-,
=M, (OGIP) e " = 1270.2 X 10' m'
a'= In R 84. 1
M,(OGIP) = 3 080 291
M, (OGIP) = 0.3233 M: (OGIP)
=R so ea =2243.0 x 10, m,
Note: Pi' Ti • and ~ are respectively initial reservoir pressure, temperature, and gas formation factor.
115
DETERMINATION OFOIL AND GAS RESERVES
2
5
10
20 3040506070 80
90 95
98
10'
Il
384 ha
/
£
, 10
L576 ha
i:
.~ (!J
102
10
10 2
5
10
20 3040508070 80 Percentage
90 95
98
Figure 6.4-3 Typical Situation: Gas-in-Place Distribution The previous distributions were calculated assuming that reservoir temperature and gas deviation factor are constants. For comparison, in-place distributions were calculated using 384, 596 and 704 hectares as the most likely value and the following temperature and gas deviation factor assumptions: I. Variable temperature, constant gas deviation factor 2. Constant temperature, variable gas deviation factor 3. Variable temperature and gas deviation factor In all cases, the calculated values for M 1 (HClP), R so and RS4. 1 agree with the previously calculated values to four significant figures: In-place distribution calculations for a 576 hectare most likely value with variable reservoir temperature and gas deviation factor are presented as Table 6.4-4. The inverse (liT, I/Z) of the denominator parameters is used to conform to theory. Calculations for the other combinations are not presented, but left as an exercise for the reader. For comparison purposes, the time required to prepare all twelve distributions was approximately 3 hours using a programmable calculator.
Observations The purpose of performing the calculations is to show the ease with which the in-place distribution can be updated. In the working world, this feature translates into more rigorous estimates that are updated more frequently and with less time and effort than is achieved with any other method. This statement becomes truer as the team gains familiarity with the methodology, the prospect, and each other. Gradually the emphasis on the reasons for performing the calculation shifts from a reactive postevent exercise to more of a planning and evaluation activity. Production's 44 X 106 mJ minimum pool volume and Exploration's 4900 x 106 mJ upside number do not appear on the probability distribution. The 44 x 106 mJ value is the product of all the minimum possible parameter values, while the product of the maximum parameter values and an optimistic 87 percent recovery factor yields the 4900 x 106 m' upside number. Consistently using the worst or best parameter values for the in-place estimate always results in a number which is less than or greater than 99.5 percent of the cumulative probability distribution and is even more extreme for the cumulative reserve distribution. The question for both groups is why they are basing their decisions on such improbable numbers. Some insight on what numbers should be used can be gained by preparing a reserve distribution (Table 6.4-5, Figure 6.4-4) and a discounted net profit before investment (DNPBI) distribution (Table 6.4-6, Figure 6.4-5) for the pool. Both distributions are prepared analogous to the in-place distribution. The reserve distribution uses the in-place distribution moments and recovery factor estimates of 65, 75 and 87 percent respectively as input, while the DNPBI distribution requires the reserve distribution moments and a unit value for the gas of $8.00, $11.00 and $15.00 per thousand cubic metres. The unit value for the gas is the estimated present value ofthe future profit from the future production, accounting for prices, production profiles, effluent composition, royalties, operating costs, inflation and discounti?g. Multiplying by the prospect reserves and subtracting the present value of the capital investment yields an estimated net present value for the prospect." From the reserve distribution shown in Figure 6.4-4, pool reserves are between 320 and 2350 x 106 m'. With • An understanding of Warren's theory governing the unit value parameter is necessary to attempt this procedure (Warren, 1988).
116
_ _ _ _ _ _ _ _ _C1
PROBABILITY ANALYSIS FOR ESTIMATES OFHYDROCARBONS INPLACE
Table 6.4-4 Gas-in-Place Distribution for Most Likely Area of 576 Hectares. Variable Temperature and Gas Deviation Factor Pool Parameters
Minimum Possible Value
xmin Basal area, Ab. " (ha) Correction to avg. area, Cr Gross pay, H(m) Net/gross pay ratio, NIG Porosity, (I - Sw) Pi(kPa abs)
64 0.80 6.5 0.65 0.12 0.78 24000
1 273.16+ 81.25
Temperature Gas Deviation Factor
11.89
Most Likely Value xprob
Maximum Possible Value xMax
m, (x)
m, (x)
1024 0.975 19 0.90 0.17 0.82 26000
554.3 0.8915 13.47 0.7831 0.1466 0.80 24903
394506 0.7977 196.4 0.6191 0.02173 0.6402 620557523
576 0.90 15 0.80 0.15 0.80 24700 1 273.16+ 81
1 273.16 + 80.97 11.87
1/.88
10000 (288.16)
Constants
0.002823
7.9693 X 10"
1.136463
1.291612
0.028439
0.000809
10' (101.325) M1 (OGlP) = 1389.6
OGIP=
10000 (288.16) Ab", (C,) H (NIGH (I-Sw) Pi 10' (101.325) r,z,
a' = In
M,(OGlP) ~ 2 749 372
M, (OGIP) = 0.3534 M: (OGIP)
a , , R 84. 1 = R so e = 2110.2 x 10 m
"
R so = M 1 (OGIP) ~,-- = 1164.5 X 10' m'
Note: Pi. Ti• and Z, are respectively initial reservoir pressure, temperature, and gas formation factor.
a 98 percent probability of exceeding the 250 x 10' m3 tie-in hurdle volume, development of this pool should be a sufficiently safe bet for even the Production Department. Once pool deliverability, pressure, temperature and effluent composition information have been supplied, the central production facilities, such as the gathering line to the gas plant, can be sized. The number of wells required to deplete the pool and interwell spacing can be estimated by comparing well deliverability to pool deliverabi!ity. Sizing of the individual wellsite facilities can also be determined from the well deliverability estimates. One way of obtaining an estimate for pool deliverabi!ity is to divide the reserve distribution by a desired rate of take. For this case a 1/3650 rate of take yields an initial deliverability range of88 to 644 x 103 m3/d. Since the discovery well flowed at 225 x 103 m3/d, it is not necessary to design the central facilities to handle the entire 88-644 x 103 m3/d range. Using the discovery well's capability as the minimum throughput, with 644
103 m3/d as the maximum, is technically acceptable and more economical than designing to cover the larger range. Completion of the equipment sizing exercise in this fashion provides the input required for sales contract negotiation, and simplifies matching contracted deliverabi!ity to facility capability. X
The purpose of equipment sizing at this stage is twofold. The present value cost of both present and future capital is required to evaluate the economic attractiveness of developing the prospect. However, only those facilities, such as the gathering line to the gas plant, that are required immediately to initiate production will be constructed on the basis of this initial estimate. Sizing of future facilities, such as field compression, can be confirmed prior to their construction because significantly more information will be available by that time. At this stage the optimal design is the one which provides the largest probability of achieving a positive net present value over the prospect reserve distribution. The optimal design does not have to provide the capability
117
DETERMINATION OF OIL AND GASRESERVES
Table 6.4-5 Reserve Distribution for Most Likely Area of 576 Hectares Pool Parameters
Minimum Possible Value xm1n
Most Likely Value x prob
Maximum Possible Value xMax
aGIP (10· m')
Recovery Factor
0.65
0.75
0.87
m, (x)
m, (x)
1389.7
2749913
0.7568
0.5773
M, (RIG) = 1051.7 a
a =In
M, (RIG) =0.3613
" = 877.9 x 10 m
-,
R so = M, ( RIG ) e
M: (RIG)
,
6
6
R"., =Rsoe = 1601.4 x 10 m
2
5
10
20 3040506070 80
90 95
98
4
10'
10
10
10 2
5
10
20 304050.6070 80 Percentage
M,(RIG) = I 587 519
90 95
98
Figure 6.4-4 Typical Situation: Reserve Distribution to operate at all the rates specified by the rate of take deliverability distribution, and probably would not when its magnitude is very large. In this case the present value tie-in cost is estimated at $2.6 million, including future field compression. The present value of future development drilling, including dry hole and wellsite facility costs, is estimated at $3.5
J
J
miUion, while sunk costs are $2.5 million. Now the effect ofbid price on profitability can be observed. The cumulate exploration and development cost of $8.6 million" ($2.6 miUion + $3.5 million + $2.5 million) intersects the discounted net profit before investment curve at a probability of42 percent (Figure 6.4-5). Thus, ifthe remaining four sections ofland could be acquired at no cost, there is a 58 percent probability ofachieving a positive net present value (NPV) through development of this pool. At the rumoured bid price of $3000/hectare, the cost for the remaining four sections is approximately $3 million, which reduces the probability of achieving a positive NPV to 39 percent on a total cost of$II.6 million (Figure 6.4-5). Is this a good gamble? Unless one is unusually lucky, probably not. A wiser course might be a minimal bid price and anticipating that the rewards (and risks) of development will likely be shared with others. Then the sharing options can be identified and their economic merits evaluated. The example illustrates one way of turninga promising exploration prospect into a probable money-losing venture. Of course there are many other ways. The key to consistent financial success is staying true to the purpose ofexploration and development, which is profitable investment, not production at any cost. Warren's method ultimately provides a means to do just that, and it starts with the in-place estimate.
Summary The example illustrates the use of the Warren Method to estimate hydrocarbons in place, and some
*Although variablecapital costs can be accommodated, single-value costshave been used to simplify the example.
118
s
I
PROBABILITY ANALYSIS FOR ESTIMATES OFHYDROCARBONS IN PLACE
Table 6.4-6 Discounted Net Profit Before Investment Distribution for Most Likely Area of 576 Hectares
Pool Parameters
Minimum Possible Value xml n
RIG (10' rrr') Unit Value ($/m3)
0.008
Most Likely Maximum Possible Value Value xMax Xprob 0.011
0.015
mt (x)
m, (x)
1051.7 0.01134
I 587519 0.0001332
M, (DNPBI) = 11.926 M,(DNPBI) = 211.4759 z
a = In
a'
M, (DNPBI) M~(DNPBI)
=
R50 = M, (DNPBI) e- 2" = $9.78 X 10'
0.3967 .
,
,
R"., = R50 e = $18.36 x 10
2
5
10
20 30 40 50 60 70 80
90 95
98
10'
10'
§ 10'
2 10
w
0 ~
x ~
C Q) ~ Q)
>
E !!!
~ ID
e'"
0 0
Q.
;; z 10
,, , 0
~BidPrice
'0
10
/ " Sunk Capital
Q)
C ~
Development Drilling Capital
0 0
'" is
..:.:...-
Tie-In Capital
1
1 2
5
10
20 30 40 50 60 70 80 Percentage
90 95
98
Figure 6.4-5 Typical Situation: Discounted Net Profit Before Investment
applications of the in-place estimate in economic evaluations. For those who accept that a probabilistic answer is the limit ofhuman capability, when assessing the future it is an extremely powerful and flexible, yet deceptively simple, tool for dealing with the uncertainties of reserves estimation. But it is not infallible. It cannot compensate for unrealistic input, it cannot warn
when the input is unrealistic, and it cannot identify the reasons for the discrepancies. These limitations restrict its use to knowledgeable, conscientious evaluators and evaluation teams that are comfortable with the method's assumptions and theory and willing to expend the effort required to attain realistic input. The payoff for these individuals is an analysis that faithfully summarizes their thoughts and their earth science expertise in a mathematical form and that can be extended to any desired depth and variables. Despite this caveat, the Warren Method will undoubtedlybe attemptedby the unthinking and the unqualified; the output, if accepted unquestioningly, will prove costly. The only safeguard is a careful examination of the evaluators' competence and the supportingevidence for the input. If both survive scrutiny, the predictions from the output are worth testing.
References Aitchison 1., and Brown J.A.C. 1966. The Lognormal Distribution. Cambridge University Press, New York,NY. Capen, E.C. 1976. "The Difficulty of Assessing Uncertainty." JPT, Vol. 28, Aug. 1976. Keefer, D. L., and Bodily, S. E. 1983. "Three Point Approximations for Continuous Random Variables." Management Science, No. 29, pp. 595-609. Warren, J.E. 1988. "Exploration and Production Decisions: Risk, Uncertainty and Economics," Course, OGGl, Houston, TX, Sep. 1988.
119
... Chapter 7
MATERIAL BALANCE DETERMINATION OF
HYDROCARBONS IN PLACE 7.1
INTRODUCTION
One of the fundamental principles used in engineering is the Law of Conservation of Matter. The application of this law to petroleum reservoirs is known as the "material balance equation" which has proven to be an invaluable supplement to direct volumetric calculation of reservoir parameters. Numerous articles and papers describe all aspects ofits use in the analysis ofreservoir performance. The material balance equation is being widely used today, aided by access to computers and the increasing knowledge base in the literature. The results from material balance calculations are significant because they are largely independent of the factors that contribute to volumetric estimates. As databases for production, reservoir pressure, and fluid properties improve, the usefulness of the material balance equation increases. When oil, gas or water is removed from a reservoir, the pressure in the reservoir tends to fall, and the remaining fluids expand to fill the vacated space. The hydrocarbon system is also affected by fluids and energy sources that are in pressure communication with it. Examples of these include connected natural aquifers, nearby injection or production activities, and other oil or gas reservoirs. The material balance is the application of the Law of Conservation of Matter to a petroleum reservoir during its depletion history. It is important for the reservoir engineer to understand the system at hand and apply the material balance realistically. Simply stated, the material balance says that the initial mass, plus the mass added, less the mass removed, must equal the mass remaining in the system. In reservoir engineering usage, mass is often replaced by volume. Thus the bulk volume, plus fluid entry volumes, plus expansion, must equal the bulk volume remaining plus voidage. If the bulk volume is considered constant, then at reservoir pressure and temperature, expansion equals voidage. The writing of a volumetric material balance is an exercise in describing the expansion of
oil, gas, water and rock with changes in pressure and temperature over discrete time periods. These time periods are chosen to extend from initial production to various later dates when both reservoir pressures and voidage cumulatives are known. The pressure-volume-temperature (PVT) properties described in Chapter 5 provide the basis for relating expansion to voidage. In material balance usage, rock and fluid volumes are normally considered at two conditions: (I) reservoir pressure and temperature, and (2) surface reference conditions. The PVT data is usually presented in a format that conveniently bridges these conditions. Since changes in reservoir temperature are relatively insignificant except for thermal projects, experimental PVT data is generally based on a constant reservoir temperature, and pressure is treated as the primary independent variable. The material balance equation has been used extensively to determine initial fluids in place, calculate water influx, estimate fluid recovery, and predict reservoir pressures. The use of the equation in defining initial fluids in place is the focus of this chapter. Applications of the equation to gas reservoirs, oil pools, and mixed drives will be discussed.
7.2
UNDERLYING ASSUMPTIONS
In terms of normal physics, the material balance equation itselfis devoid ofconditions and assumptions, but in regular oilfield usage, a number of underlying assumptions arise. These may result from the way in which the input data is derived or from computational simplifications. The material balance calculation is based on changes in reservoir conditions over discrete periods of time during the production history. The calculation is most vulnerable to many of its underlying assumptions early in the depletion sequence when fluid movements are limited and pressure changes are small. Uneven depletion and partial reservoir development compound the accuracy problem.
120
s
MATERIALBALANCE DETERMINATION OFHYDROCARBONS INPLACE
The basic assumptions in the material balance method are as follows: Constant Temperature. Pressure-volume changes within the reservoir are assumed to occur without related changes in temperature. The pressure changes happen slowly in most of the reservoir, and the mass of adjacent rock volumes is such that the reservoir system very closely approaches constant temperature performance. Pressure Equilibrium. A uniform pressure is assumed to apply across the pool. The model is considered as a tank, with infinite permeability. This is a critical assumptiou, since the expansion properties ofthe rock and fluids are stated in terms of prevailing pressure. Local pressure variations around producing or injection wellbores may generally be disregarded. However, regional trends must be recognized and included in the pressure averages. Constant Reservoir Volume. Reservoir volume is assumed to be constant except for those conditions ofrock and water expansion or water influx that are specifically considered in the equation. The formation is considered to be sufficiently competent that no significant volume change will occur through movement or reworking of the formation due to overburden pressure as the internal reservoir pressure is reduced. The constant volume assumption also relates to an area of interest to which the equation is applied. If the focus is on some part of a reservoir system, except for specific exterior flow terms it is assumed that the particular portion is encased in no-flow boundaries. Reliable Production Data. As measurement technology has improved and regulatory authorities have consolidated the data-gathering process, the reliability ofproduction and injection data has improved substantially. Good well rate data is critical to the material balance, as net voidage figures directly in the calculated oil in place. Representative PVT Data. The PVT information is the other main ingredient of the material balance equation. It is assumed that the PVT samples or datasets represent the actual fluid compositions and that reliable and representative laboratory procedures have been used. Notably, the vast majority of material balances assume that differential depletion data represent reservoir flow and that separator flash data may be used to correct for the wellbore transition to surface conditions. Such "black oil" PVT treatments relate volume changes to temperature and pressure only. They lose validity in cases of volatile oil or gas condensate reservoirs where
compositions are also important. Special laboratory procedures may be used to improve PVT data for volatile fluid situations.
7.3
EXPLANATION OF TERMS
As previously indicated, the material balance equation relates net reservoir voidage to expansion of reservoir fluids. This section describes the various components of voidage and expansion used in the conventional black oil material balance. . Table 7.2-1 lists reservoir voidage terms. In addition to wellbore flow streams, water influx-efflux acts as a pseudo production quantity. Various independent water influx formulations are discussed in Section 7.7.3. Table 7.2-' Reservoir Voidage Terms
Fluid Gas cap gas
Surface Volumes
Reservoir Volumes
Liberated gas
Gpe Gps-N pRs
GpeBge (Gp-N R)B sp sgs
Injected gas
-0-I
Oil
Np
-GiBgi NpBo
Water
Wp
Water injected
-w I
where G = gas subscripts B = formation volume factor N = oil W = water R = gas-oil ratio
WpBw -WB I w c = gas cap g = gas = injected fluids a = oil p = produced fluids s = solution gas w = water
In Table 7.2-1, the formation volume factor, B, is the volume at reservoir temperature and pressure per unit of surface reference volume. The change in formation volume factor for the various fluids is proportional to their compressibilities. Rock compressibility usually ranges from 0.4 x 10,6 to 1.5 X 10-6 vol./pore volume/ kPa (kPa,I), and is primarily dependent upon porosity. Water compressibility is linear with pressure, and ranges from 0.3 to 0.6 kPa,l. Oil compressibility shows some nonlinearity with pressure. It varies from 0.4 to 3.0 kPa,l, relating to its gravity. Gas at 14000 kPa has a compressibility in the order of 60 x 10-6 kPa,l. The behaviour of the material balance calculation follows directly from the relative compressibilities as manifested by the formation volume factors. 121
7
qiit-
DETERMINATION OFOIL AND GASRESERVES
As pressure is reduced in an oil-gas-water system, liquid volumes increase in the undersaturated fluid region. When the oil reaches its saturation pressure, gas is released and a vapour phase begins to form. Further pressure depletion results in diminishing liquid volumes and rapidly expanding gas volumes. Both total fluid volume and system compressibility then increase. Table 7.2-2 provides various expansion terms that occur in a material balance equation. These terms offset the various voidage quantities in the material balance equation.
Table 7.2-2 Reservoir Expansion Terms Material
Expansion
Gas Cap Liberated Gas Oil
N(B,-B,;)
Water
NB (Hm) S c dP 01 ww _ 1 8w
Rock
NB,j(Hm) --'-----crdP I-S w
where c
compressibility (volumechange/ volume/pressure unit)
f dP m
N B R
7.4
N(R,;-R,)B g,
formation change in pressure gas cap reservoir volume/ oil zone reservoir volume oil in place formation volume / surfacereference volume ratio of gas content/ oil volume (surface reference conditi'ons) connate water saturation(fraction of pore space)
GENERAL MATERIAL BALANCE EQUATION
The general material balance equation equates reservoir voidage to reservoir fluids expansion. If the voidage terms of Table 7.2-1 are equated to the expansion terms of Table 7.2-2 and N is factored out from the expansion terms, rearrangement yields the general material balance, Equation (I). This form contains
all of the factors that could be applied to routine determinations of oil and gas in place. The fifth term in the numerator, We' is water influx and is defined in Equation (13) in Section 7.7.
(1)
In Equation (1) the formation volume factors reflect the reservoir volume per unit of stock tank or surface volume. They are dimensionless, i.e., reservoir m 3/surface m3 • The terms of the equation represent volumes or changes at reservoir conditions. Reservoir engineers commonly use the same formation volume factors for gas cap gas, solution gas and injected gas, the degree of error inherent in such a simplification depending upon the circumstances. If such a shortcut is taken, Equation (1) is reduced to the form ofEquation (2), which will be used to illustrate adaptations ofthe material balance for particular conditions. The engineer is free to re-insert the distinction between diverse gas compositions when it is worthwhile to do so.
(2)
7.5
SPECIAL CASES OF THE MATERIAL BALANCE EQUATION
7.5.1
Undersaturated Oil Reservoirs
Several terms of the material balance equation may disappear when reservoir conditions negate their use. This is particularly true for the volumetric undersaturated oil reservoir. For this case there is no gas cap, and since reservoir pressure is above the bubble-point pressure of the oil, there is no free gas in the oil zone. Production depends largely upon liquid expansion ifreservoir pressure is being depleted. Therefore, rock and connate water expansions are significant and should be included. Equation (3) provides a material balance for an undersaturated pool with water injection, production, and influx. N
BOi
(3)
(B, -B,;) + I-S (Swcw+c r) dP w
122
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _1
MATERIAL BALANCE DETERMINATION OFHYDROCARBONS IN PLACE
7.5.2
Saturated Oil Reservoirs
The saturated oil reservoir, either with or without a gas cap, exhibits a much greater compressibility than the liquid-filled undersaturated system. Even a small gas saturation is noteworthy, due to the relatively high compressibility of gas. In such cases rock and water compressibility are often neglected in the interest of minimizing the calculations. Equation (4) is the material balance equation for a saturated reservoir, initially at the bubble-point pressure. The terms for gas and water injection, water influx and water production may be added as required.
With the usual assumption of an isothermal reservoir, Equation (6) becomes: (J,,((Z/P),,-(Z/P),;)
(13,-13,) -l- (Ft,,-Ft,) 13,1fm13'i
(J"
7.5.3
Gas reservoirs are also amenable to the material balance treatment. Starting with Equation (1), it is assumed that water production, influx and injection are zero. Since gas has a very high compressibility, rock and water expansion in the gas cap may be safely neglected. Oil production and expansion terms are not applicable. Cross-multiplying Equation (I) and making substitution gives Equation (5): m~13"
=--
(5)
13 gci
Eliminating the terms for net water voidage, rock and water expansion, and those relating to oil zone production and expansion gives Equation (6): (6)
The gas formation volume factor, 13g' may be replaced according to Equation (7): (7)
(9)
Equation (9) can also be transformed to the form shown in Equation (10): P
= (P/Z)"-(Jp,(--),,
(10)
(J"Z
(4)
Gas Reservoirs
(J "
= o, C_(P/Z~,,(Z/P)J
(P/Z)"
s,
(8)
Rearranging, Equation (8) becomes:
~p13,1f((Jp-~pFt,)13,
(13,-13,)
= (Jp,(Z/P)"
Equation (10) is in the form of a straight line, y = mx -tb. Hence, plotting P/Z vs. (J and extrapolating the line to P/Z = 0 yields the initial gas in place. This is a traditional method ofcalculating gas reserves for a volumetric pool. Fluid entry or exit from the system is indicated by upward or downward plot curvature, respectively. Such performance may be seen in cases of water influx from an aquifer, interference with other pools, or interference with a portion of the reservoir outside of the area of interest. Formation compaction also may cause a nonlinear PIZ curve. In this case the historical trend will run above the gas-defined slope in early years and then tum sharply down to the true gas in place.
7.6
LIMITATIONS OF MATERIAL BALANCE METHODS
The basis of the material balance is firm, and the equation can be made to encompass most ofthe factors relevant in hydrocarbon production. However, in practical application, several sources of errors limit the accuracy of material balance methods. The gravity of these errors varies with circumstances. 1. Thermodynamic equilibrium is not attained in actual field conditions.
2. PVT data is obtained using liberation processes that do not represent reservoir conditions. 3. Inappropriate average pressures are used.
where P sc = Z = T = T sc = P =
standard or reference pressure gas compressibility factor reservoir temperature reference temperature formation pressure
4. There is uncertainty in the "m" ratio. 5. The production data used is inaccurate. The amount ofpressure decline covered by the production history is one ofthe best criteria in gauging potential errors. The material balance is a comparison ofvoidage to expansion and concentrates on evaluating fluid
123
• DETERMINATION OF OIL ANDGASRESERVES
expansion. Large pressure declines produce large expansions, making inaccuracies in production volumes relatively less significant. Similarly, pressure errors are less critical with more pressure depletion. In general, a pressure decline of 10 percent of the original reservoir pressure is needed before the material balance becomes reliable. This critical depletion level is highly dependent upon the quality of the pressure, production and PVT data. Pressure errors originate from several sources. Gauge and sonic survey errors can be compounded during processing and conversion to a common datum. True static pressures may be difficult to derive in low transmissibility pools with high viscosity fluids. Areally imbalanced withdrawal or injection may create regional pressure gradients in the pool. It is important that such areal pressure variations be properly reflected in the averages applied to material balance equations. Volumetric averaging of measured values is the preferred technique. Multiple layers ofdiffering permeability and severe lateral changes in permeability within the formation may complicate the gathering of representative pressures. A study by Hutchinson (1951) presents the quantitative effect of pressure errors on material balance determinations of hydrocarbons in place.
7.7 7.7.1
SUPPLEMENTAL CALCULATIONS Gas Caps and Aquifers
Most ofthe material balance parameters are defined by pressures, PVT measurements, and production-injection data. Original oil or gas in place can be calculated in some circumstances, but in cases where gas caps or aquifers exist, the material balance equation contains more than one unknown. Supplementary calculations must then be utilized for a solution. Gas caps can often be estimated by volumetric means. Core and log data from upstructure wells can be used with conventional volumetric mapping techniques to estimate the amount of associated gas that is in contact with the oil zone. The gas cap volume enters the material balance equation through the parameter "rn" in Equation (1). As gas is a high mobility fluid, the gas cap can often be represented as having the same reservoir pressure history as the adjacent oil zone. However, when the gas zone is large relative to the oil zone or when the gas zone is geographically widespread, the areal pressure variation within the gas cap should be considered. Small errors in gas cap average pressure can produce large changes in calculated oil in place, because gas is much more compressible than oil.
If an aquifer is large enough to impact the pressure performance ofthe hydrocarbon zones significantly,part of the water is likely to be substantially removed from the hydrocarbons, due to its low compressibility. Water also has much less mobility than gas. Therefore, the assumption of common pressure used for oil zones and their gas caps is usually not applicable to hydrocarbon zones and their aquifers.
7.7.2
Water Influx Measurements
The simplest method of externally determining water influx for use in the material balance equation is to measure it directly. In pools where water influx is anticipated, the operators may periodically log selected wellbores to determine water saturations. The advance ofthe oil-water or gas-water contact can be defined with a selection of logged wellbores distributed across the area of the hydrocarbon-water interface. The engineer must have reliable data for reservoir porosity and water saturation adjacent to the water contact. It is also very helpful to have an independent source of residual hydrocarbon saturation in the water-invaded zone. Such data may be obtained from relative permeability measurements in special core analyses. The accuracy of water influx volumes from periodic water contact elevation maps varies with the circumstances. The reliability of water saturation and porosity values is important. There is also an element of doubt in the reservoir stratification. Rock capillarity variations and transmissibility barriers may cause undulations in the water contact as influx occurs. Areal variations in reservoir pressure can also lead to nonuniform water advance. The user should be aware of the potential for error when working with a limited number of water contact measurements.
7.7.3
Analytical Water Influx Models
Water influx may be calculated from the material balance equation as a function oftime using a volumetric estimate of oil in place. The engineer can then endeavour to match this influx vs. time trend with an analytical "model." Ifa reasonable match ofan extended historical period is achieved with a single set of coefficients, the analytical relationship is plausible and provides a basis for estimating future influx for use in the material balance. Schilthuis (1936) provided the simplest aquifer influx model. His model assumes that constant pressure is maintained somewhere in the aquifer and that flow to the oil zone is proportional to the pressure differential,
124
n
MATERIAL BALANCE DETERMINATION OFHYDROCARBONS INPLACE
with the remaining factors in D'Arcy' s Law constant. Equation (II) shows the Schilthuis steady state formulation:
~p
= pressure differential, aquifer limit to oil-water contact (kPa) (psi) Q(t) = dimensionless water influx; function of to to = dimensionless time lJ. = constant, 0.0863 (6.323 x 10-3) k = aquifer permeability (Ilm2) (mD) t = time (days) = porosity, fractional Il = water viscosity (mPa's)(cp) c = effective rock, water compressibility, kPa-1 (psi") rw = equivalent oil zone radius (m) (ft) ~ = constant, 6.2792 (1.119) h = equivalent aquifer thickness (m) (ft) El = azimuth angle of aquifer inflow (degrees)
(I I)
where k = water influx constant (m3/d/kPa) P, = aquifer boundary (initial) pressure (kPa) p = oil-water contact pressure (kPa) Hurst (1943) proposed a modification of Equation (II) wherein the influx constant is altered and a denominator term, log (at), is added. The denominator compensates for the gradually lengthening flow path of the water through the aquifer as depletion progresses. Hurst's modification is shown in Equation (12): dW,= c(p;-p) dT
log (at)
(12)
where c = water influx constant (m3/d/kPa) a = time conversion constant that depends on units of time t = elapsed time from start of influx (h) Van Everdingen and Hurst (1949) produced an unsteady state water influx solution which can deal with infinite or limited aquifers. This model is based on radial flow from a concentric aquifer to an interior oil zone, but it can be adapted to situations where the aquifer underlies or extends primarily in one direction from the oil pool. Van Everdingen and Hurst overcame the site-specific nature of the solutions to the radial form of the diffusivity equation by providing their data in terms ofdimensionless time and dimensionless influx. Briefly, their . formulation is as follows: W, = BL [~p. Q (t)]
(13)
(14)
B=
El
Amcr ' h t''Y w 360
where We = water influx (m") (bbl) B = water influx constant (m3/kPa) (bbllpsi)
The superposition theorem is applied to calculate water influx, We' The pressure history at the water contact is divided into a series oftime intervals for which average contact pressures can be estimated. These average pressures define decrements between the initial aquifer pressure and the hydrocarbon interface pressure that are assumed to be constant within each time interval. The superposition theorem holds that the aggregate effect of all these pressure differentials is equivalent to the summation of their individual effects, each operating over its respective time interval. In practice, reservoir parameters are chosen to calculate to as a function of the time intervals. Tables and figures ofQ(t) have been supplied by Van Everdingen and Hurst (1949) and in the summary by Craft and Hawkins (1964). Craft and Hawkins also provide a good description ofhow to calculate the summation of ~PQ(t) to get W as a function of time. e Carter and Tracy (1960) developed a method based on Hurst's (1958) simplification of the Van Everdingen and Hurst unsteady state influx calculation. The CarterTracy method gives answers similar to those of Van Everdingen and Hurst without the iterative solution involving the conventional material balance equation and the water influx summation equation.
7.8 (15)
MULTIPLE UNKNOWN MATERIAL BALANCE SITUATIONS
The solution methods outlined rely on separately determining relationships for secondary unknowns in the material balance equation, namely the gas cap to oil zone ratio, m, or the water influx term, We' A second technique utilizes a simultaneous solution for oil in place and a secondary parameter. Theoretically, given
125
DETERMINATION OF OIL AND GASRESERVES
multiple pressure and production combinations, the material balance equation could be simultaneously solved for multiple unknowns. In practice, transient effects, data errors and unrepresentative averages make the simplistic simultaneous solution unreliable. Havlena and Odeh (1963) presented an algebraic rearrangement of the material balance. Their technique involves calculating production and expansion entities that are interrelated as terms of a linear equation. Since the pressure-production-time points plot as a straight line, graphical methods can more easily be used to determine the best solution for the dataset. Havlena and Odeh emphasized the idea of examining multiple values of a parameter by means of a statistical variation factor. In some circumstances, this approach provides a useful supplementary measure of how well the entire pressure-production history is satisfied by a particular reservoir solution. The straight-line method involves the use of variable groups. The reservoir circumstances determine which variable groups are plotted against each other. This method attaches a significance to the sequence and direction of the plotted points and the shape of the resulting plot. The variable groups can be effectively computed and plotted with a spreadsheet program, particularly if the derivation of PVT data is automated through macros. The analyst must then examine the sequence and configuration of the plot points to assess their meaning.
by Eo, s, and Eg, respectively. The s, components may be deleted, except in the case of undersaturated oil pools. The final right term, We' is calculated by the unsteadystate water influx equation, Equation (13). Alternatively, the Carter-Tracy influx formulation could be used.
With minor rearrangement, Equation (2) may be rewritten as:
Btl E,+rnsE g
There are many different formulations of the straightline material balance method. The reader is encouraged to reference the comprehensive and readable presentation by Havlena and Odeh (1963). Figure 7.7-1 shows the form of the straight-line plot for a pool with unknown oil zone and gas cap size, and Figure 7.7-2 portrays one with unknown oil zone and water drive. McKibbon et al. (1963) provide an excellent example of the application of the straight-line material balance to an oil reservoir with active water influx.
N
u,
N
=oil in place
g' Source: Havlena and Odeh, 1963.
N, [B, + B,(R,-R,,)] + (W,-W,) a, - G,B" Figure 7.7-1
J
mB (B.-B,,) + W, + --"
n,
(16)
where cf = formation compressibility Cw = water compressibility B, = formation volume of oil and originally dissolved gas Using Havlena and Odeh terminology, the left side of Equation (16), denoted by F, represents the net reservoir volume ofproduction. The expansion terms for oil, rock and water, and gas on the right side, are denoted
Straight Line Plot for Oil Zone and Gas Cap Case
The usual criteria for a successful material balance solution are consistency of the results and agreement with volumetric calculations. The consistency aspect is often left as a rather nebulous, unquantified factor, but Havlena and Odeh offer a method to systematize it. Agreement with volumetric oil in place estimates can be overemphasized. Volumetric calculations tend to focus on total oil in place due to their reliance on geologic interpretations and petrophysical data. Material balance oil-in-place is the active oil that takes part in the depletion history. The similarity of volumetric and material balance oil-in-place values should not be overrated as a measure of the accuracy of either.
126
$
MATERIAL BALANCE DETERMINATION OFHYDROCARBONS INPLACE
Formal computer programs are available to perform many material balance calculations. They handle much of the repetitive computation and greatly speed the solution process. However, users must be sure that they understand how such programs work. The methods must fit the problem and be compatible with the available dataset.
N
B= N = oil in place
o0
IdpQ(A1 D)
Eo Source: Havlena and Odeh. 1963.
Figure 7.7-2
Straight Line Plot for Oil Zone and Water Influx Case
Although it is theoretically possible to solve for multiple unknowns with the straight-line method, in practice, difficulty is met in some cases. Highly accurate data are needed to solve simultaneously for a notable gas cap and an oil zone, or for a gas cap, oil zone and water influx. The difficulty in these two cases relates to the high compressibility of gas and its large potential impact on the pressure response. In conclusion, the straight-line requirement does not prove the uniqueness of the solution, but is one of the conditions that a satisfactory solution should meet. As always, the quality of the solution will depend on the quality and quantity ofthe input data and on the ability and thoroughness of the analyst. The straight-line method is recommended as being robust and effective. Its dynamic nature is a valuable supplement to traditional methods.
7.9
COMPUTER SOLUTIONS
Computer spreadsheets are valuable tools in material balance work. They greatly reduce laborious calculations and allow easy sensitivity analyses with varied data. A noteworthy advantage of spreadsheets is that the user retains complete knowledge and control of the computation method.
Havlena and Odeh (1963) caution against total automation ofthe straight-line material balance, because the sequence and direction of successive points provide information as to the nature ofthe solution. The engineer should take care to scrutinize this aspect ifmachine plots are utilized in the straight-line method.
References Carter, R.D., and Tracy, G.W. 1960. "An Improved Method for Calculating Water Influx." Trans., AIME, Vol. 219, p. 415. Craft, B.C., and Hawkins, M.F. 1964. Applied Petroleum Reservoir Engineering. Prentice Hall, Inc., Englewood Cliffs, NJ, p. 205. Havlena, D., and Odeh, A.S. 1963. "The Material Balance as an Equation of a Straight Line." Trans., AIME, Vol. 228, p. 896. Hurst, W. 1943. "Water Influx Into a Reservoir and its Application to the Equation of Volumetric Balance." Trans., AIME, Vol. 151, p. 57. - - -.. 1958. "The Simplification of the Material Balance Formulas by the Laplace Transformation." Trans., AIME, Vol. 213, p. 292. Hutchinson, C.A. 195I. "Effect of Data Errors on Typical Engineering Calculations." Paper presented at SPE of AIME meeting, Oklahoma City, OK. McKibbon, lH., Paxman, D.S. and Havlena, D. 1963. "A Reservoir Study ofthe Sturgeon Lake South D-3 Pool." JePT, Vol. 2, No.3, Fall 1963, p. 142. Schilthuis, R.l 1936. "Active Oil and Reservoir Energy." Trans., AIME, Vol. 118, p. 33. Van Everdingen, A.F., and Hurst, W. 1949. "The Application of the Laplace Transformation to Flow Problems in Reservoirs." Trans., AIME, Vol. 186, p. 305.
127
i
g
PART THREE ESTIMATION OF RECOVERY FACTORS AND FORECASTING OF RECOVERABLE HYDROCARBONS
R
---
Chapter 8
OVERVIEW OF PART THREE
8.1
INTRODUCTION
Part Two focuses on in-place hydrocarbons or resources; Part Three addresses reserves, which are the portion of the resource, or the quantities of oil and gas and related substances that are economically recoverable under known technologies and a generally acceptable forecast of future economic conditions. Forecasting of recoverable hydrocarbons may be approached from several standpoints: recovery factor as a percentage of original in-place resources; statistical analogies, reservoir simulations, and material balance techniques; or methods such as decline analysis, where the determination of in-place hydrocarbons is not a requirement. Many factors may affect the recovery of hydrocarbons: • Depletion mechanisms and the timing of the implementation of various recovery methods • Reservoir and hydrocarbon characteristics • Well spacing, completion techniques, mechanical conditions, and production equipment The natural depletion mechanisms for oil include, but are not limited to, primary production mechanisms in which reservoir fluids are produced as a result of the energy of fluid expansion, solution gas drive, water drive, gas cap drive, compaction drive, and combination drive. These primary production mechanisms are described in Chapter 9. Production of natural gas generally involves primary depletion using surface compression, but recovery of liquid- and sulphur-rich gases often utilizes re-injection of dry gas or cycling to maximize recovery. The depletion methods for natural gas recovery are covered in Chapter 10. Primary oil recovery can be improved by secondary and tertiary recovery schemes referred to as "enhanced recovery." Chapters II through 15 describe the various enhanced recovery methods used in oil reservoirs: waterflooding, hydrocarbon miscible flooding, immiscible
gas injection, thermal stimulation, and carbon dioxide flooding. Another method of improving recovery from oil reservoirs is by the use ofhorizontal wells, which allow drainage from larger areas than vertical wells. Chapter 16 discusses horizontal wells. Reservoir characteristics that may affect hydrocarbon recovery include heterogeneity and reservoir discontinuities, both vertical and lateral; the structural characteristics of the reservoir; the presence of natural fractures, both open and closed; pore size geometry and distribution; permeabilities; in situ stresses and fracture orientation; parting pressures (injecting fluids); and reservoir pressure. Hydrocarbon characteristics that may affect recoveries include viscosity, composition, and the pressurevolume-temperature relationships of the hydrocarbons in the reservoir. The interrelationship of fluids and reservoir rock, expressed in terms such as interfacial tensions and wettability, control fluid movement in a reservoir. The overall contrast between the mobility of fluids in a reservoir significantly affects recovery. The well spacing, completion intervals within wells, completion techniques such as fracturing, and proximity of wells to underlying water or a gas cap are all factors to consider when analyzing recoveries. Mechanical equipment such as compressors can also significantly affect recoveries as well as the abandonment of wells.
8.2
PURPOSE OF DEPLETION STRATEGY
The purpose of a depletion strategy is to maximize project economics and the recovery of hydrocarbons. While this may sound obvious, the current focus on quarterly earnings by most North American shareholders, coupled with a tough economic climate, often results in the need for immediate cash flow, which sometimes overrides longer term business strategies. However, it should not preclude companies from investigating other
131
DETERMINATION OFOILAND GASRESERVES
development options and addressing those that meet their financial constraints. The development of a depletion strategy should ultimately result in the identification of all potential recoverable reserves and the establishment ofa framework that can maximize revenues from the project. Developing a depletion strategy early in a project is very important because the timing of the implementation of various production strategies could be critical. It may not be prudent to continue primary production without fully addressing a depletion strategy for a pool. The following are examples of what could happen:
1. Depleting a gas cap could cause a disastrous decrease in the recovery factor of an oil pool. 2. Production from an oil pool to the extent that the pressure drops below the critical gas saturation in the reservoir prior to commencement of a waterflood could have a detrimental effect on recovery. 3. Gas production with the pressure declining significantly below the dewpoint in a retrograde gascondensate reservoir before implementing a dry gas cycling scheme could result in a dramatic decrease in liquid recovery. Planning the depletion strategy during the initial development stages of a pool will also identify the appropriate data that should be gathered and accumulated through both the drilling and the production stages of development. The availability of this information will assist in identifying the most economically feasible depletion mechanism.
8.3
TECHNIQUES FOR RESERVES AND PRODUCTION FORECASTING
The techniques used for reserves estimation and production forecasting vary depending upon several criteria: • The reservoir depletion strategy The type of depletion mechanism, both existing and future • The stage of reservoir development and depletion • The extent of the production history • The constraints that have been imposed on production by regulation, markets, or the physical nature of production facilities .
The reliability of techniques to forecast reserves and production improves during the life of the pool as more options become available. In the very early stages, with little more than geophysical, geological, and wellbore data and test information available, it is common practice to rely on analogy and statistical data for preliminary reserves estimates. During subsequent phases of reservoir depletion, the availability of increasing volumes of information may lead to the use of two more sophisticated techniques of reserves estimation: numerical simulation and decline curve analysis. These are the techniques most commonly used for reserves estimation and production forecasting. The use of numerical simulation is not restricted to reservoirs with significant producing histories, but the ability to calibrate the reservoir model developed by matching historical performance offers far more reliable results although the technique is often expensive. This technique is of particular value where decisions are necessary regarding the feasibility of some form of enhanced recovery mechanism. Numerical simulation is discussed in Chapter 17. Decline curve analysis is both used and misused in reserves and production forecasting, and it has widespread use in every aspect ofreservoir depletion. Clearly, the more established a decline trend becomes, the more reliable the extrapolation ofthat trend, provided the underlying reservoir or production mechanism that is causing the decline does not change. Decline curve analysis is discussed in Chapter 18. In Part Two, the techniques for determining the most likely in-place hydrocarbon volumes are discussed. The assignment of recovery factors to these volumes at this stage, particularly in the case of oil, requires an assessment ofthe reservoir environment and the recovery mechanism in order to determine likely performance by analogy to similar, and preferably nearby, pools. In westem Canada, a wealth of statistical data is available from the Alberta Energy Resources Conservation Board (ERCB); the B.C. Ministry of Energy, Mines, and Petroleum Resources; and the Saskatchewan Department of Energy and Mines. Some ERCB data is presented in Chapter I 9.
132
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _. .d
Chapter 9
NATURAL DEPLETION MECHANISMS FOR OIL RESERVOIRS
9.1
INTRODUCTION
Pressure drops in a reservoir caused by the withdrawal of some of the fluids initiate the expansion of the remaining fluids. Oil, gas, and water are then produced as a result of their expansion and the expansion of the surrounding reservoir rock. This recovery process is called a natural depletion mechanism. The names for the various natural depletion mechanisms-fluid expansion, solution gas drive, water drive, combination drive, and gas cap drive-are associated with the major contributing source of expansion energy. When more than one major source of expansion energy contributes to the depletion process, it is referred to as a combination drive. This chapter discusses the natural depletion mechanisms, the types of predictive tools and their applicability at the different stages of development of a reservoir, and the factors affecting recovery.
9.1.1
Fluid Expansion
Fluid expansion exists as a natural depletion process when only one mobile fluid exists in the reservoir. (Fluid may refer to either gas or oil.) The withdrawal of some of this fluid will cause a pressure drop. The remaining fluid will expand and displace itself toward the pressure drop. Because ofthe highly compressible nature of gas, fluid expansion is generally the dominant depletion mechanism in gas reservoirs. Conversely, because ofthe low compressibility ofliquids, fluid expansion is not a good source of depletion energy in oil-filled reservoirs. Fluid expansion in oil reservoirs exists by itself only at pressures above the bubble point. At the bubble point, the gas dissolved in the oil breaks out of solution, and the expansion energy associated with the compressive nature of this gas becomes the dominant depletion mechanism. Only oil deposits containing very undersaturated oil will be produced with fluid expansion as their dominant depletion mechanism.
9.1.2 Solution Gas Drive The predominant source ofenergy for solution gas drive comes from the expansion of gas released from the oil. As the pressure drops in a reservoir, the ability of the oil to keep gas dissolved is reduced, and free gas is released. With further pressure reduction, the free gas expands and displaces oil towards the producing wells. Because of its highly compressible nature, the gas will expand and displace significantly more oil than an initially equal volume of liquid. In an undersaturated oil reservoir, that is, one without any initial free gas, the initial depletion mechanism will be due to the expansion ofoil. Generally, there will be a direct relationship between the volume and rate at which the oil is produced and the pressure reduction, as shown in Stage I in Figure 9.1-\. When the pressure drops below the bubble point, free gas is released and becomes the major source of expansion energy. Gas-oil ratio does not significantly increase during this stage until the critical gas saturation is reached. Because of the compressible nature of the gas, with continued oil production, the pressure drop is significantly reduced and the oil rate will be fairly constant, as shown in Stage II of Figure 9.1-\. Expansion
SolullonGas Drive
II
III
,. ....
Pressure
~I
/
I
-- .....
"\ \
q!!'1
rY/
Oil Production
~I
o
\
1 I
I I
----_ ....
/ /
Cumulative Oil
Figure 9.1-1
Solution Gas Drive Reservoir
133
, DETERMINATION OF OIL AND GAS RESERVES
As the pressure continues to drop, the evolved free gas will reach the critical saturation; at this point, gas will start to move and will be produced in conjunction with the oil. As the gas saturation increases, the ease with which gas moves within the reservoir relative to oil increases, and the gas is then produced preferentially over the oil. With continued production and the associated pressure drops, the gas continues to be evolved, increasing its saturation level. The production of gas increases and the production ofoil decreases. This complicated procedure, represented by Stage III in Figure 9.1-1, continues until the rate at which gas is being evolved from the oil is less than the rate of gas being produced. At this point, the pressure and production rates drop quickly, as shown in Stage IV.
9.1.3
Water Drive
An oil deposit is considered to be produced by water drive when the predominant source of expansion energy comes from the water-filled portion of the reservoir. Since water has a lower compressibility than oil, the volume of water needs to be significantly larger than the oil-filled portion of the reservoir. The pressure in the oil deposit will drop as production is initiated. As the pressure gradient reaches the aquifer, the water starts to expand, displacing the oil toward the producing wells. If the aquifer is large enough and thus has sufficient expansion energy, all the mobile oil will be produced without any further pressure drops. The oil rate will remain constant until the aquifer contacts the producing well, after which the water production will increase as the oil rate drops.
If the aquifer is not large enough to provide full pressure support, the pressure drops. When the bubblepoint pressure is reached, free gas will be released, and this gas will start to contribute significantly to the depletion energy. This type of depletion mechanism is referred to as a combination drive because there is more than one significant source of depletion energy. Figure 9.1-2 shows the relative difference between solution gas drive, full water drive, and a partial water drive. In many situations, at a localized area around the producing wells, the water contact will rise dramatically and effectively water out the wells. This phenomenon is called "water coning." The consequence of water coning is that large volumes of oil will be trapped and thus become unrecoverable. In reservoirs that are subject to coning, recovery factors tend to be very low. The more viscous the oil and/or the greater the vertical permeability, the more dramatic the effect of coning on recovery.
Partial Water Drive
., ty0t\ »>
.
'\et '5\l.9......
'!\QI.1\
... '"
'ncteas\~9_ ~-
-
~ -- -~-
Solution Gas Drive
Full
~~~~
Cumulative Oil
Figure 9.1-2 Comparison of Solution Gas Drive and Water Drive Reservoirs
9.1.4
Gas Cap Drive
A reservoir that initially contains free gas as well as the gas dissolved in the oil will benefit from the additional expansion energy of the free gas. If the volume of free gas is large enough so that this source of expansion energy overshadows the effect of other sources of energy such as solution gas drive, the primary depletion mechanism is called a gas cap drive. As in water drive reservoirs, the oil undergoes an initial pressure drop until the pressure gradient reaches the gas cap. The gas then expands and displaces the oil toward the producing wells. If the gas cap is large enough, the oil deposit will undergo only minimal pressure drop, and the oil production rate will remain constant until the gas cap reaches the producing well interval. Due to relative permeability effects, the gas production rate will then increase quickly as the oil rate drops off. If the gas cap is not large enough to give complete or nearly complete pressure support, then as the pressure drops, solution gas drive will be contributing free gas energy. The resultant drive mechanism is also referred to as combination drive. Figure 9.1-3 shows the response ofa gas cap drive reservoir that becomes a combination drive reservoir. As in water drive reservoirs, many gas cap drive reservoirs are also subject to coning effects. Because of the inherent differences in viscosity of gas and oil, coning is often more serious in gas cap reservoirs than water drive reservoirs. In the presence of gas coning, recovery factors tend to be relatively low.
9.1.5
Compaction Drive
In weak, unconsolidated reservoirs, the pressure drop due to the production of fluids causes an imbalance III
134
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _FZ8
NATURAL DEPLETION MECHANISMS FOR OILRESERVOIRS
this "sandwich" effect. The recovery factor in this situation would be fairly low. ,
___ Pressure Oil Production Rate
,, ,,
,
, ,, , ,
,
Cumulative Oil
Figure 9.1-3 Gas Cap Drive Reservoir
the stress within the bulk rock, and the weight of the overburden causes the bulk rock to compact. The compacting rock squeezes the internal fluids, thus maintaining the pressure. The resultant drive mechanism is referred to as compaction drive. Compaction drives are found in heavy oil reservoirs and some natural1yfractured reservoirs where fractures tend to close as the reservoir is being depleted. Compaction drives can increase the recovery due to solution gas drive by more than 10 percent of the original oil in place.
9.1.6
Combination Drive
Often recovery from oil reservoirs is the result ofmore than one drive mechanism. A reservoir with combination drive poses a difficult problem for reserve estimation. General1y one depletion mechanism is dominant at any stage of depletion or geographic area of the reservoir. In a reservoir that has a smal1 gas cap, initial1y the dominant drive mechanism is solution gas drive. When significant volumes of gas have evolved out of solution, the dominant drive mechanism becomes gas cap drive. For example, in the presence of both a gas cap and an aquifer, the dominant mechanism at the gas-oil interface would be gas cap drive, and the dominant drive mechanism at the water-oil interface would be water drive. It is critical for the evaluator to understand the reservoir and which drive mechanism is dominant. In a combination drive reservoir that has both a water leg and a gas cap, coning has a double effect in that the gas cones downward and the water cones upward. Thus, significant volumes of oil will be by-passed by
9.2
FORECASTING OF RECOVERABLE OIL
Throughout the productive life of a reservoir, there is always a need to establish the reserves. Recovery estimates are used to justify capital spending, predict future cash flow generation and, ultimately, estimate shareholder value. Because of the importance of reserve estimates, al1 available data should be used when determining the size of the oil deposit and the amount of oil that can be recovered economical1y. The amount and accuracy ofthe available information increase as an oil deposit passes through the various phases of the production life cycle. Thus, the recommended methodologies used to estimate recoverable oil change as the quantity of information increases. Two basic approaches are used to establish reserves for an accumulation. In the first approach, the ultimate economic recovery factor is established through analogous or analytical methods, and then applied to volumetric estimates based on geological interpretations (as discussed in Part Two). The second approach predicts future production rates, with reserves calculated as the summation of the volume produced above the economic limit. Table 9.2-1 and Figure 9.2-1 show the recommended methodologies according to stage ofproduction life and whether recovery factor or reserves are predicted. Sometimes material balance and numerical simulation are useful in the development stage The purpose of establishing a reserves estimate, the size and value of the reserves to the corporation, and the
*
a:
MaterialBalance
is
Decline Analysis-)-
~>
-,
-<-' Numerical Simulation )Analytical Methods
k-c7""c-_-
Analogous Methods
"-.
~
)-
-~
Time
Figure 9.2-1 Recommended Methods for the Stages of Exploitation
135
• DETERMINATION OF OIL AND GAS RESERVES
Table 9.2-1 Recommended Reserves Forecasting Methods
Stage
Forecast Method
What is Forecasted
Exploration
Analogous Analytical methods
Recovery factor Recovery factor
Delineation/ development
Analogous Analytical methods
Recovery factor Recovery factor
Early life
Analytical methods
Recovery factor
Middle/late life
Numerical simulation Decline analysis Material balance
Reserves Reserves OOIP
Abandonment
Actual production
Reserves
amount and reliability of the data should dictate the degree of effort put into calculating an estimate. Often comparing two or more methods of evaluation is recommended. For example, an estimate determined from decline analysis could be compared with one calculated using an analytical method. Information on a particular reservoir can be obtained by techniques such as drilling, coring, logging, production testing, pressure testing, and fluid analysis. Prior to obtaining any ofthis information through the drilling of the first well, the evaluator must resort to the use of established information from analogous fields. Analogies can be used to estimate recovery factors, initial production rates and decline rates that are applied to the geological interpretation. The more similar the analogous field is in size, depth, fluid properties and formation, and the closer its proximity to the prospect, the better the estimate of recovery factor will be. In analytical methods, the mathematical equations that represent material balance calculations have been simplified by making certain assumptions about particular parameters. By measurement of some and "guessing" at the remainder, the evaluator can establish the recoveries. Analytical methods have been developed for the more complicated processes such as solution gas drive, water drive and gas cap drive. Fluid expansion is a fairly simple process, and therefore production forecasting and recovery estimates are generally solved directly from the material balance equation. Analytically predicted recovery factors along with either early life production history or rates based on analogous fields are applied to the geological interpretation in order to establish recoverable volumes of hydrocarbons. Material balance, whether done graphically or numerically, attempts to establish initial in-place
136
volumes of oil, gas and water. In order to establish a recovery estimate, the results of the material balance analysis must be combined with another prediction technique or assumptions applied to the depletion of the reservoir. For example, assumptions on abandonment conditions define the pressure or production rate at which the field would be abandoned; thus the difference between volumes in place and the volumes remaining at abandonment establishes the reserves. The material balance method is discussed in detail in Chapter 7. Decline analysis is the prediction of future rates based on observed behaviors seen in actual production histories. Typically, reservoir engineers forecast the future well flow behaviours by extrapolating production history using a straight line. A single straight line will represent the entire life ofa reservoir only when there is one source of reserve energy in a simple homogeneous reservoir with all wells producing at a similar rate. In other words, a single straight line would represent the entire production life for only a few oil reservoirs. In using decline analysis, it is important to know what stage of the natural depletion is represented by the production history and is being represented by the prediction. More than one straight-line segment may be necessary. Other factors that can invalidate the use ofthe straightline method are the existence of dual porosity systems, layered reservoirs with each layer having different properties, and geographic areas of an accumulation with each area having different properties. These phenomena, when incorporated into the prediction, change what would have been a straight-line segment in a homogeneous reservoir into a curved line. A technique to handle geographic differences is to subdivide the reservoir i?to areas of similar characteristics and perform dechne
NATURAL DEPLETION MECHANISMS FOR OIL RESERVOIRS
analyses on each area. Summing the various areas will give a more accurate picture of the entire reservoir. Numerical simulation, material balance and decline analysis are the methods most commonly used in the middle and late stages of depletion. These methods require a sufficient amount ofreliable data to be effective predictors of recoveries. The following subsections present general comments on the use ofthese methods for the specific drive mechanisms. Numerical simulation and decline analysis are discussed in more detail in Chapters 17 and 18, respectively.
9.2.1
Solution Gas Drive
Oil recovery as a result of solution gas drive typically ranges between 2 and 30 percent. The lower recoveries generally occur in low API, shallow, and low pressure oil reservoirs, whereas the higher recoveries occur in high API oil, deep, and high pressure reservoirs.
Analytical Methods The most common analytical methods for estimating recovery in solution gas drive reservoirs are based on material balance concepts. Four methods are applicable below the bubble point. The most common analytical approach used is the Tracy Method, followed by the Muskat Method. The following are the most commonly used analysis methods: The Tracy or Tarner Method (Tracy, 1955) is a rearrangement of the basic material balance equation so that pressure-dependent variables are grouped. Tamer extended the method by incorporating the gas-oil equation based on gas-oil relative permeability curves, resembling the Pirson and Muskat methods. The Muskat Method (Muskat, 1949) uses the material balance equation, written in differential form, in conjunction with the gas-oil relative permeability curves. Because of the importance of these curves, some degree of confidence in the data is crucial. The Pirson Method (Pirson, 1950) is based on the Schilthuis material balance equation written in finite difference form. This is essentially a material balance equation that predicts oil recovery as a fraction of oil in place at the bubble point as the pressure declines over a time period. The gas-oil relative permeability curve is required to define the producing gas-oil ratio. The Humble (Schilthuis) Method (Schilthuis, 1936) is based on the Schilthuis material balance equation. In the forecasting of future production, the equation is applied to reservoir conditions at the beginning and end
of specified periods, and the interim production or pressure change is obtained by difference. Short-cut methods are used when there is little data or when a recovery estimate is desired quickly. These are not recommended if a high degree of confidence is desired. Two short-cut methods are as follows: Wahl et al, (1958) created various nomographs based on the Muskat Method using varying fluid properties and relative permeability characteristics. The Roberts and Ellis (1962) Method uses the early GaR data to predict future production. Using oil gravity and solution gas-oil ratios, the trend of producing gas-oil ratio is matched to the published predictions.
Decline Analysis The productive life for a solution gas reservoir that initially was above the bubble point is made up of four distinct stages as shown in Figure 9.1- I. In a decline analysis, the analyst must know what stage of depletion is represented by the production and must predict when the reservoir will enter future stages. Because ofthe difficulty of predicting when these future stages will occur, production decline analysis is generally not used as a predictive tool until the production data reaches Stage III.
Reservoir Simulation In solution gas drive reservoirs, generally analytical and decline techniquesare sufficient to estimate reserves. In special situations typically dictated by geological discontinuities or heterogeneity and in naturally fractured reservoirs, simulation may be warranted to establish reservoir flow and resultant recoveries.
9.2.2
Water Drive
Oil recoveries in a water drive reservoir can typically range from 2 to 50 percent depending on factors inherent in the reservoir. A common method for evaluating recovery efficiency of water drive reservoirs uses the observed rise of the water-oil contact due to the water influx from the aquifer. This requires sufficient production history for the water-oil contact to rise noticeably and a method of measuring the rise. The relationship over time between the fraction of the reservoir invaded by water and the initial oil-filled reservoir compared to the oil produced allows the prediction of the total oil recovery. Additional factors that affect reserves include coning, fractional flow, and economic limit.
137
• DETERMINATION OFOILANDGAS RESERVES
Analytical Methods If the observation of the advance of the water-oil contact is insufficient to directly predict recovery efficiency, or direct measurement of the advance is not possible, theoretical methods based on material balance are recommended with preference given to the Welge Method. Assuming near-constant pressure at any time, the reservoir recovery, ER , can be calculated using the relationship:
W,· w,e, E = --'---'---'R
where
HCV e
(I)
W, = water influx Wp = cumulative water production Bw water formation volume factor net water influx at reservoir We - Wp Bw conditions HCV, = cumulative water-invaded hydrocarbon volume
Ultimate recovery is then determined from reservoir production vs, cumulative water encroachment. The following are the most commonly used analysis methods: The Welge Method (Welge, 1952) is recommended if production history data is insufficient to determine the efficiency of the water drive. Fractional flow of water, fw ' as a function of water saturation, is used to predict oil recovery. I . _A_kk-,,"::..:(_.1.,:-:,pg::...s_in_CJ....:.) q,/.lo I + /.lw k ," J.Lo k.;
(2)
throughput area = formation permeability !c"o relative permeability to oil .1.p = density difference, water densityoil density acceleration due to gravity g CJ. = formation dip total throughput rate q, Ilw = water viscosity Jlo ::;:;: oil viscosity k.w= relative permeability to water
where A k
138
The last term on the right-hand side of Equation (2) represents the effect of gravity on fractional flow. For a nontilted reservoir, this term becomes zero. The relative permeability vs. saturation relationship must be reliable in order for this method to result in a reasonable recovery estimate. The Dietz Method (Dietz, 1953) predicts oil recovery in reservoirs where the waterfront flows updip along the base of the formation, causing the front to assume a tilted position. This is especially noticeable in reservoirs with a water influx rate that exceeds the critical rate and in reservoirs containing viscous oil. The MarshaI'Method(Marshal, 1957) uses BuckleyLeverett theory to predict recovery in a stratified reservoir. From production history the time required for a given water cut to move between two rows ofwells in a field is obtained, and the velocity of the water front determined. Field-measured water cuts are used to describe oil-water relative permeability curves. If enough water-cut ranges are available, the fractional flow curve vs. distance in the reservoir, as defined by BuckleyLeverett, can be predicted. The Schilthuis Method (Schilthuis, 1936) determines water influx by calculating the water flow from the aquifer to the reservoirin a series ofsteady-state steps. Water influx is assumed to be proportional to the pressure difference between the aquifer and the reservoir. Since aquifer pressure is assumed equivalent to the initial reservoir pressure, this method is valid only for infinite-acting aquifers. The weakness in this method is due to calculation of an aquifer constant from production history. The Modified Hurst Method (Hurst, 1943) is similar to the Schilthuis material balance method in that it also . predicts water influx. The Hurst equation extends the Schilthuis Method by accounting for the increase in the drainage radius in the aquifer. Correlations have been identified and should only be used for quick evaluations or where data is minimal: Khan and Caudle (1968) for thin oil columns, Caudle and Silberberg (1965) for edge-water drive, Hutchinson and Kemp (1956), and Henley et al. (1961). The analytical methods discussed in this subsection assume that the water-oil contact rises as a flat surface, either from the flank or from the bottom. If the reservoir is subject to coning, these analytical methods will overestimate oil production rates and ultimate
NATURAL DEPLETION MECHANISMS FOR OIL RESERVOIRS
recoveries. In the early life of the reservoir, i.e., prior to water break-through, empirical correlations exist to identify the susceptibility of the wells to coning. These methods forecast recoveries by estimating break-through time and the water-oil curve forecast. Using the wateroil forecast, oil production can be estimated. Although reservoir simulation is recommended for evaluating coning situations, the following correlations are available for quick evaluation:
I. Kuo (1989) combines various correlations that determine critical rate calculations, break-through time calculations, and water-cut performance predictions on a PC spreadsheet for rapid analysis of comng. 2. Boumazel and Jeanson (1971) combine experimental correlations with a simplified analytical approach based on the assumption that the front shape behaves like a straight line. This method may be applied to thick homogeneous reservoirs that are horizontally fed. 3. Sobocinski and Cornelius (1965) developed a correlation based on laboratory data for predicting water coning time as it builds from static to break-through conditions. This method involves correlating dimensionless cone height against dimensionless time. 4. Kuo and DesBrisay (1983) developed correlations based on numerical simulation to determine the sensitivity ofwater coning behaviour to various reservoir parameters, including the ratio of vertical to horizontal permeability, the ratio of perforated interval to oil thickness, the production rate, and the mobility ratio. 5. Numerous correlations have been developed based on the theoretical curves by Muskat for homogeneous reservoirs. The best known correlations include Muskat and Wuckoff(l935), Chaney et al. (1956), and Chierici et al. (1964). All these methods use the theoretical curves to obtain a critical production rate, the maximum production rate at which oil can be produced without coning. In order to estimate recoveries, a way of forecasting water-oil ratio and oil production must be incorporated. Therefore, these correlations in themselves will not forecast recoveries.
Decline Analysis In some water drive reservoirs, the production forecast might be represented by two straight-line segments, pre- and post-water break-through. Due to
the difficulty of predicting the timing of water breakthrough using decline analysis, this method is generally used after break-through has occurred. Many approaches are available in analyzing production after water break-through. Table 9.2-2 outlines the more commonly used combinations of production data plots. In the analysis of any data set, it is recommended that a number of these combinations be used, selecting the combination that gives the best match. Table 9.2-2 Decline Analysis Plots Used after Water Break-through 1. Logoil rate vs, time (exponential decline) 2. Oil rate vs. cumulative oil (exponential decline) 3. Logoil rate vs. cumulative oil (harmonic decline) 4. Logcumulative oil vs. log cumulative oil plus water
5. 6. 7. 8.
Oil and waterratesvs. cumulative oil Logoil and waterrates vs. cumulative oil Log water-oil ratio vs, cumulative oil Logwater-Coil + water) ratio vs. cumulative oil
Material Balance Material balance methods for estimating reserves in water-drive reservoirs frequently result in erroneous estimates. A detailed understanding of the supporting aquifer is required for any degree ofreliability. Often information about the aquifer is extremely difficult to obtain. Knowledge that is critical includes the size of the aquifer, the strength or pressure support provided by the aquifer, and the areas of the oil reservoir that receive pressure support. In addition to an estimate of original oil in place, the parameters defining the aquifer must be solved from the production and pressure history data. With the addition of these unknowns, the material balance method has a greater number ofvariabies to solve than it has equations. Because of this, material balance generally results in multiple estimates of original oil in place. Early in the production history of a reservoir, material balance methods may give erratic results for water influx due to inaccurate pressure measurements or because well pressure measurements may not be an accurate representation of actual average reservoir pressure. In the early life of depletion, an erroneous negative water influx may be calculated.
139
F
DETERMINATION OF OIL AND GASRESERVES
Reservoir Simulation
In reservoirs where coning is a key issue, a reservoir simulation radial coning model is recommended. Reservoir simulation is discussed in more detail in Chapter 17.
and pro-rated back to the individual wells, sometimes result in erroneous amounts and allocation of gas production. The accuracy ofgas production depends on the frequency and method of measurement and the variation between wells in the reservoir.
9.2.3
Reservoir Simulation
Gas Cap Drive
Oil recoveries in a gas cap drive reservoir can be as high as 60 percent depending on factors inherent to the reservoir. Three dominant factors influence recovery: I. Since the gas cap provides the recovery energy, it must be of sufficient size to displace oil to the producing wells. In general, the longer the gas cap can maintain the pressure, the greater the recovery. 2. High vertical permeability allows the liberated solution gas and oil to segregate, adding additional energy to the gas cap. 3. Early gas break-through increases the gas-oil ratio significantly, thus removing the main source ofdrive energy. Analytical Methods
Because the drive mechanism in gas cap drive reservoirs is frequently combination drive, generally in conjunction with solution gas drive, the recommended methods for prediction of recoverable oil are decline analysis, material balance and reservoir simulation, all of which take into account the complicated nature of the reservoir. Short-cut methods include the following: The Welge Method (Welge, 1952), as previously described for water drive reservoirs, may be used for low viscosity oil reservoirs.
The Dietz Method (Dietz, 1953), as previously described for water drive reservoirs, may be used in reservoirs where the gas cap overruns the oil along the reservoir flank. In this case, the rate ofadvance must be below the critical rate for the method to be valid.
Because ofthe relatively higher mobility ofgas, careful planning is critical if a reservoir simulation model is to be used. The grid blocks and time steps should be small enough that the movement ofgas can be physically represented by the simulator. If coning is an issue, a radial model is recommended. Reservoir simulation is discussed in more detail in Chapter 17.
9.2.4
Combination Drive
In a combination drive reservoir, generally one depletion drive mechanism is dominant at a particular time or in a particular area of the reservoir. Therefore, in generating production forecasts, it is necessary to identify the predominant sources of energy throughout the life of the reservoir and to identify the predominant sources of energy affecting a particular geographic area of the reservoir. Because of the complexity of predicting the start and shape of the future production affected by different dominant depletion mechanisms, decline analysis techniques generally are not attempted until the last stage. Techniques appropriate to the specific depletion mechanism dominant during the last stage of depletion should be used.
9.3
FACTORS AFFECTING OIL RECOVERY
Although the drive mechanism is the primary factor influencing recoveries, numerous other factors, either inherent to the reservoir or resulting from human intervention, influence ultimate recovery. The following subsections address some of these other major factors.
These analytical methods assume the gas-oil contact will advance as a flat interface. If the reservoir is subject to severe coning, these methods will overestimate both the production rate and the recovery. The correlations describing water coning can also be modified to estimate gas corung,
9.3.1
Material Balance
where k = k,o = h = Pc = Pw =
Since gas is an important fluid in the recovery of oil in gas cap drive reservoirs, a word of caution is advised when using the material balance method. Oil field measurement practices, where gas is measured periodically
140
Production Rate
The production rate, qo' of a well is defined by the radial flow equation: q = o
21tkk"h (Pr- Pw) IJ)n(r,lrw )
permeability relative permeability to oil net pay in situ pressure of accumulation wellbore pressure 110 = viscosity of oil
(3)
NATURAL DEPLETION MECHANISMS FOR OILRESERVOIRS
r, = external boundary radius rw = wellbore radius For natural depletion mechanisms, the only parameters that can be altered due to human intervention are nearwellbore permeability and producing pressure. The near-wellbore permeability can be enhanced through stimulation techniques such as acidizing and fracturing. The producing wellbore pressure can be reduced by the installation and optimization of artificial lift equipment. For a given oil deposit, adjusting the production capability of the wells will not alter the theoretical quantity ofmoveable oil, but will affect the recoverable . resource through economic limit, as demonstrated in Figure 9.3-1. Ifthe only difference between the two cases shown is the production capacity of the well, the cumulative production at the economic limit will be larger for the high rate case.
Reserves
Theoretical
Recovery
Economic Limit
Cumulative Recovery
Figure 9.3-1 Relationship Between Production Rate and Reserves
9.3.2 Oil Quality The type of oil in the reservoir directly affects reserves through the volume of gas in solution and through oil viscosity. Oils that have less gas dissolved in solution have less reservoir energy for oil recovery under solution gas drive; these are generally lower gravity oils. Oil viscosity influences recovery in two ways. First, if there are two fluids in a reservoir with significantly different viscosities, oil production would decline quite rapidly because ofconing or fingering ofthe other fluid. Second, productivity of a well is inversely proportional to viscosity (Equation 3). All things being equal, a more viscous oil would have a lower production rate and would reach its economic limit sooner.
In general, lower API oil receives a lower price at the refinery. Since the price directly impacts the economic limit, the limit would be reached sooner for lower priced crude.
9.3.3 Reservoir Characteristics Reservoircharacteristics can affect recovery factors from theoretical calculations primarily because of heterogeneities in the reservoir. Generally, heterogeneities cause a reduction in reserves either by (I) decreasing the amount of oil in place that can be effectively tapped by the wells, or (2) causing uneven depletion of portions ofthe reservoir, in turn resulting in a greater amount of oil being left in the ground because it is uneconomic to produce. Some ofthese reservoir characteristics include permeability variations, dual porosity systems, naturally fractured reservoirs with cemented fractures, and low permeability stringers. Although a heterogeneous reservoir generally has a lower recovery than a homogeneous reservoir, some heterogeneities can assist the drive mechanism, and thus increase reserves. For example, in bottom-water-drive reservoirs where coning is of concern, shale stringers can restrict the advance of water, allowing higher oil production for a longer period of time. Also, open uncemented, or partially cemented natural fractures can help improve recoveries from low permeability reservoirs that otherwise would be uneconomic to produce. In general, the more heterogeneous the reservoir, the larger the difference in the actual reserves as compared to the theoretical calculations.
9.3.4 Reservoir Geometry Many factors associated with the reservoir geometry influence the amount of oil produced under primary depletion. Some of these are the shape of the reservoir, the continuity ofthe formation, the layering ofmultiple sands, faulting, structure, and dip. These factors can affect both the drive mechanism and the economic viability of developing the accumulation. Depending on the predominant drive mechanism, the geometric configuration will have varying degrees of effect. For example, in a solution gas drive reservoir, vertical relief could allow the formation of a secondary gas cap, which would maintain the evolved gas as an energy source. In general, the less continuous reservoirs would result in a lower recovery because some parts of the reservoir might not be in communication with the producing
141
•, DETERMINAnON OFOILANDGASRESERVES
wells. In this case, infill drilling to reach untapped oil would result in an increase in reserves. Also, due to discontinuities in the reservoir, gas-oil and water-oil contacts might not advance as a flat interface, and thus oil would be by-passed. A layered reservoir poses a different type of problem, especially if the multiple zones have significantly different reservoir characteristics. If one zone were more prolific due to considerably higher permeability, it would have a higher recovery factor than the less prolific zone. In this case, it is often beneficial to estimate the recovery factor separately for the multiple zones. Because ofthe different behaviours of the various zones, a layered reservoir manifests itself as a hyperbolic or harmonic decline if decline analysis is being used.
9.3.5
Effects of Economic Limit
Whether a recovery factor is rigorously established through detailed techniques like numerical modelling or estimated through engineering judgement, innate assumptions are made about the economic limit of the reservoir. In some cases the economic limit is established in the current economic environment using known technology. The key factors affecting the economic limit are the prices for the hydrocarbons, the operating cost, the current fiscal regime, and encumbrances such as overriding royalties and net profit interests. These factors are discussed in Part Four. The following subsections discuss some of the other factors that influence the economic limit. Well Spacing
A single well in a large deposit of oil will theoretically produce all of the moveable oil, but this would take a very large number of years and would not provide the optimum economic recovery. As the well is produced, a pressure gradient is established in the reservoir. With continued production, the pressure gradient moves further out into the reservoir, effectively reducing the average reservoir pressure. As the average pressure drops, the production rate of the well will drop proportionately. When the radius of the area affected by the pressure gradient becomes sufficiently large, a pseudoequilibrium is established in which the flow at the furthest boundary reached by the pressure gradient is equivalent to the production rate of the well. The pressure gradient will continue to move further out into the deposit, minimally affecting the production rate, until the physical limits of the deposit are encountered.
If other wells are drilled into the same reservoir, but are far enough apart that their respective pressure gradients will not interact until after the economic limit has been reached, each will behave as if it were the only well in the reservoir. If the densities of the wells are such that their respective pressure gradients interact at the economic production limit, the reservoir pressure would be at the original level at the point of interaction, resulting in an overall high average reservoir pressure at abandonment. Inserting a well midway between the two original wells will result in a lower average reservoir pressure at abandonment, and thus a higher economic oil recovery. However, the oil recovered per well will be less. With continued reduction in spacing, the average reservoir pressure at abandonment will continue to drop, but in diminishing increments. The result will be a typical relationship between the oil recovered above the economic limit and the number of wells in the pool. The intersection of the oil recovery forecast and the economic limit establishes the reserves for this reservoir. The relationship between well spacing and abandonment pressure is depicted in Figure 9.3-2. The point at which increasing the number of wells will no longer markedly increase the oil recovered when producing above the economic limit is generally referred to as the optimum spacing (Figure 9.3-3). This assumes that the revenue benefit from the additional recoverable oil in reducing spacing while moving from point a to point b offsets the cost ofdrilling, completing, and equipping the necessary additional wells, and provides the required return on investment. Increasing the density of wells beyond point b may be economic through the effects of rate acceleration. However, the volume of oil recovered above the economic limit will remain the same unless by having more wells and thus larger volumes, the economy-of-scale factors will reduce the average economic limit per well. The optimum well spacing will be unique for each deposit and should be established by a combined technical and economic assessment. Facility Sizing and Constraints
Facilities must be installed in order to separate the produced oil, gas, and water. The size ofthe facility and the resulting capital and operating costs (the economics of the project) have an impact on the ultimate reserves. Very simply, if the capital cost of the required production facility is greater than the potential revenue, the reservoir will not be developed and produced, and therefore cannot be considered to contain reserves, even
142
-
.-sra
NATURAL DEPLETION MECHANISMS FOR OILRESERVOIRS
Original Pressure Single Infill
Single Well
Multiple Infills
Average Abandonment Pressure Figure 9.3-2 Relationship Between Well Spacing and Abandonment Pressure
Present
~
Value
p
~
.....
~~
~/
,
Cumulative Oil
,,
,,
.,.~,/I~b ," ,,
the decline of the oil rate will be sharper, as depicted in Figure 9.3-4. The decision whether to increase the capacity of the facility is based on an economic evaluation of the benefit of the additional oil and the cost of expansion.
a
,r ,,
,, ,
,,
'\
!
,, ,, ,, ,, ,, ,
is
~
E
o"
Constrained od\.\C\\O{\
,, ,,
F==:::::=::-'l'lQ!ol!!!!alC!F",lu!"id,;..Po-'Oil Production
.:»:
Number of Wells
Figure 9.3-3 Optimum Well Spacing if it has been adequately delineated through drilling. A facility sized large enough to handle the maximum initial production will continue to have high operating costs when oil volumes decline in the future, and will reach its economic limit earlier than a smaller, less expensive facility that limits initial production, but has lower operating costs. Sometimes facilities need to be installed in oil fields to handle increasing production volumes ofassociated gas and water. Installing large facilities that will not be utilized for many years may not be economic, and the use of constraining facilities may be necessary. When a naturally declining oil rate reaches a facility constraint,
Cumulative Oil
Figure 9.3-4
Effects of Facility Constraints on Economic Limit
Regulatory Constraints
In addition to the standard economic considerations of developing a reservoir (rate of return, payout, operating costs, and facility costs), there are also the regulatory constraints imposed by the local government agencies. The purpose of these regulations is to ensure the conservation and responsible exploitation of a depleting resource, to ensure that the equitable rights of
143
DETERMINATION OF OIL AND GASRESERVES
competing producers are met, and to protect the environment. Regulations with respect to well spacing, location of wells on a spacing unit, production rate, water-oil ratios, gas-oil ratios, and hydrogen sulphide emissions have been established to meet the objectives ofthese agencies. These regulations will, in some cases, impose constraints on development scenarios and thus affect the estimates of recoverable hydrocarbons. This topic is discussed in more detail in Chapter 23, The
Khan, A.R., and Caudle, B.H. 1968. "Scaled Model Studies of Thin Oil Columns Produced by Natural Water Drive." SPE 2304. Kuo, M.C.T. 1989. "Correlations Rapidly Analyze Water Coning." O&GJ, Oct. 1989, pp. 77-80. Kuo, M.C.T., and DesBrisay, C.L. 1983. "A Simplified Method for Water Coning Predictions." SPE 12067.
References
Marshal, D. 1957. "Mathematical Treatment of Water Invasion of Oil-Bearing Formations." Erd. Kohle, Vol. 10, Dec. 1957, p. 825.
Bournazel, C., and Jeanson, B. 1971. "Fast WaterConing Evaluation Method." SPE 3628.
Muskat, M. 1949. Physical Principles ofOil Production. McGraw-Hili, New York, NY.
Caudle, RH., and Silberberg, I.H. 1965. "Laboratory Models of Oil Reservoirs Produced By Natural Water Drive." SPEJ, Mar. 1965, pp. 25-36.
Muskat, M., and Wuckoff, R.D. 1935. "An Approximate Theory of Water Coning in Oil Production." Trans., AIME, Vol. 114, pp. 144161.
Regulatory Environment.
Chaney, P.E., Noble, M.D., Henson, W.L., and Rice, T.D. 1956. "How to Perforate Your Well to Prevent Water and Gas Coning." O&GJ, Vol. 55, May 1956, pp. 108-114. Chierici, G.L., Ciucci, G.M., and Pizzi, G. 1964. "A Systematic Study of Gas and Water Coning by Potentiometric Models." JPT, Aug. 1964, pp. 923-929. Dietz, D.N. 1953. "A Theoretical Approach to the Problem of Encroaching and By-Passing Edge Water." Proc., Konikl. Ned.-Akad, Wetenschap, Series B, Vol. 56, p. 83. Henley, D., Owens, W.W., and Craig, F.F. 1961. "A Scaled Model of Bottom Water Drives." JPT, Jan. 1961, pp. 90-98. Hurst, W. 1943. "Water Influx Into a Reservoir and Its Application to the Equation of Volumetric Balance." Trans., AIME, Vol. 151, p. 305. Hutchinson, T.S., and Kemp, C.E. 1956. "An Extended Analysis of Bottom Water Drive Reservoir Performance." Trans., AIME, Vol. 207, pp.256-261.
Pirson, SJ. 1950. Elements ofOil Reservoir Engineering. McGraw-Hill, New York, NY. Roberts, T.G., and Ellis, H.E. Jr. 1962. "Correlation of Gas-Oil Ratio History in a Solution-Gas-Drive Reservoir," JPT, Vol. 14, Jun. 1962, p. 595. Schilthuis, RJ. 1936. "Active Oil and Reservoir Energy." Trans., AIME, Vol. 118, p. 33. Sobocinski, D.P., and Cornelius, AJ. 1965. "A Correlation for Predicting Water Coning Time." JPT, May 1965, p. 594. Tracy, G.W. 1955."Simplified Form of the Material Balance Equation," Trans., AIME, Vol. 204, p. 243. Wahl, W.L., Mollins, L.D., and Elfrink, E.R 1958. "Estimation of Ultimate Recovery from SolutionGas Drive Reservoirs." JPT, Jun. 1958, p. 132. Welge, HJ. 1952. "A Simplified Method for Computing Oil Recovery by Gas or Water Drive." Trans., AIME, Vol. 95, p. 91.
144
_____________n
Chapter 10
DEPLETION MECHANISMS FOR NATURAL GAS RESERVOIRS
10.1
INTRODUCTIOJII
During the depletion of natural gas reservoirs, many factors affect the production performance. The basic characteristics and physical properties ofthe gas and its associated constituents or products, and its proximity and interrelationship to other fluids in the reservoir can either enhance or adversely affect the recovery from a pool. The most significant aspect, however, is the compressibility and, conversely, in the reservoir, the expandable nature ofpressurized gas. On average, a significantly higher percentage of the gas in a reservoir is recovered through natural depletion mechanisms than of the oil, which has lower compressibility. This chapter highlights some of the characteristics of the gas and the reservoir that influence recoveries and basic approaches in forecasting recoverable gas reserves.
10.2
CHARACTERISTICS OF NATURAL GAS
The gases that constitute natural gas belong mainly to the "paraffin series." The main constituent is methane. Impurities such as nitrogen, carbon dioxide, helium, and hydrogen sulphide may be present in natural gas. The Alberta Energy Resources Conservation Board classifies natural gas with less than one percent hydrogen sulphide as "sweet." When the hydrogen sulphide content is over one percent, the gas is classified as "sour." Natural gas found by itselfin a reservoir and completely in the gaseous state is classified as "nonassociated," (Figure 10.2-1). Gas found in an oil reservoir with no free gas present except that which is in solution is classified as "solution gas." Gas and oil may be found in a reservoir in many different combinations when the field is discovered, and the relationship of the gas and oil mayor may not change, depending on the reservoir and fluid characteristics and on drilling, completion and production practices. For example, gas may be found as free gas above the oil. This is called a "gas cap," and the gas is classified as "associated" gas. Under some
Associated Gas
Nonassociated Gas Source: Clark, 1960,
Figure 10.2-1 Classification of Gas Based on Source in Reservoir
reservoir conditions and producing practices, the dissolved gas may come out of solution in the reservoir and form a "secondary" gas cap or add to a natural gas cap. At low pressures in shallow fields, natural gas and crude oil appear as distinct substances in the reservoir (Figure 10.2-2, Reservoirs A and B). As the pressure at which petroleum is found rises with increased depth, gas dissolves in crude oil, and the high-boiling constituents dissolve in the gas phase. Some fields have both oil and gas in contact (Figure 10.2-2, Reservoir C). Deeper fields at pressures over about 27 600 kPa (4000 psi) and at temperatures ofmore than 95°C (200°F) contain singlephase fluids that are not immediately distinctive as oil or gas fields (Figure 10.2-2, Reservoir D). "Dry" gas reservoirs normally yield little or no surface liquid recovery with processing through normal lease separation equipment. A gas is "wet" if hydrocarbon liquids are extractable in surface separation equipment, and may be produced from a single-phase gas reservoir, a retrograde condensate gas reservoir, or an "associated gas" reservoir.
145
-/ DETERMINATION OFOILANDGAS RESERVES
Ground Level
.\ containing Dissolved G ........... 0\ Os
A
Water
_ _......
Water
B
Source: Katzet al., 1959.
It is also quite common to find a volatile oil rim. In this case, the gas cap would be exactly at the dew point.
DEFINITION OF RESERVOIR TYPES FROM PHASE DIAGRAMS
If the accumulation occurred as shown by point C, the reservoir would be in a single-phase (oil) liquid state, since the temperature is below the critical temperature. In this case, as the pressure declined, the bubble point would be reached (Point C I). Below this point, a free-gas phase would appear. This gas is classified as "solution gas."
Various types of reservoirs can be defined using pressure-temperature phase diagrams (Figure 10.3-1). The area enclosed by the bubble-point and dew-point lines is the region ofpressure-temperature combinations for which both gas and liquid phases exist. The curves within the two-phase region show the percentage ofthe total hydrocarbon volume that is liquid for any temperature and pressure. Initially, each hydrocarbon accumulation would have its own phase diagram, which would depend only upon the composition of the accumulation. A single-phase gas reservoir at discovery is shown by point A. Since the fluid in the reservoir during production remains at 150°C(300°F), it retains its gaseous state as the pressure declines along path A-AI' Furthermore, the composition of the produced gas does not change as the reservoir is depleted. However, cooling and pressure drop in the wellbore and surface facilities allow the condensing of gas along the line A-A2 • This accounts for the production of condensate liquid at the surface from a gas in the reservoir. Retrograde gas condensate reservoirs or dew-point reservoirs exist at pressures sufficient to be at or above the upper boundary of the two-phase envelope and at a
146
When a retrograde gas condensate reservoir has conditions on or very close to the dew-point line at the time of discovery, it means that the percentage of intermediates (C2 - C6) is high.
Occurrence of Oil and Gas
Figure 10.2-2
10.3
temperature between the critical and cricondentherm values, as shown by point B. Here the fluid is also in the one-phase gaseous state. As pressure declines because of production, the composition of the produced fluid will be the same as for reservoir A, and remain constant until the dew-point pressure is reached (Point B 1) • Below this pressure, liquid condenses out of the gas as fog or dew, leaving the gas phase with a lower liquid content. The condensed liquid adheres to the walls of the pore spaces of the rock, and is immobile. Thus the gas produced at the surface has a lower liquid content and the producing gas-condensate ratio increases. This process of retrograde condensation continues until a point ofmaximum liquid volume is reached (Point B2) . Vapourization of the retrograde liquid occurs from B2 to the abandonment pressure at point B) and can be noted by decreasing gas-condensate ratios on the surface.
Ifthe same hydrocarbon mixture occurred at point D, it would be a two-phase reservoir, consisting of a liquid or oil zone overlain by a gas zone or "gas cap." As the compositions of the gas and oil zones are entirely different from each other, they may be represented separately by individual phase diagrams. The oil zone will produce as a bubble-point oil reservoir and the gas cap will be at the dew point, and may be either retrograde as shown in Figure 10.3-2 (a) ornonretrograde as shown in Figure 10.3-2 (b). The initial in-place gas and condensate for gas condensate reservoirs, both retrograde and nonretrograde, may be calculated from the available production data by recombining the produced gas and condensate in the correct ratio to find the composition, average specific gravity (air = 1.000), pseudo-critical pressure, and pseudo-critical temperature of the total well fluid, which is presumably being produced initially from a single-phase reservoir.
DEPLETION MECHANISMS FOR NATURAL GASRESERVOIRS
Reservoir Temperature (oG) -18 4000
3500
10
38
66
94
122
150
Bubble Point or
DewPoint
Single Phese
or
Gas Reservoirs
Dissolved Gas Reservoirs
Retrograde Gas-Condensate Reservoirs
',A II
178 27600
24150
II
"II
·2/1
til 3000 'w
'iiI I
off/l
el 0..1
oS ~
'Cs'I
::>
l:l
~I
2500
~
IJ.J
a.
I I
~
.~
2000
I"
Q)
20700
I
I I "0 I 'S I u::: I .: I
::>
17250
l:l ~
a. ~
~
13 800
I
Q)
.~ Q)
&!I
8!
~ ~ ~
I
Q)
a::
-I
°1
£1
1500
10350
"'I 0-
1
I
A,
1000
6900
I
I I I
Source: AfterCraft, 1959.
100 150 200 250 Reservoir Temperature (OF)
300
350
3450
Figure 10.3-1 Pressure-Temperature Phase Diagram of a Reservoir Fluid
BP
011T
T
Temperature
Temperature
(a)
(b)
Source: Craft, 1959.
Figure 10.3-2
10.4
Phase Diagram of a Cap Gas and Oil Zone Fluid
GAS RECOVERY
Ideally, 100 percent gas recovery is the goal. For reservoirs producing by gas expansion and without water drive, there is no physical reason why the gas may
not be recovered down to near atmospheric pressure. However, the production rates decrease so rapidly when the pressure approaches atmospheric that some abandonment pressure is established for economic production. Most volumetric depletion reservoirs with reasonable permeabilities will produce 70 to 90 percent of the original gas in place. Sometimes the higher limit ofrecovery can be approached when operating costs are low and gas prices high. In other reservoirs, substantial losses will occur. But it is sometimes possible to minimize this loss through proper reservoir management and the application ofbasic principles ofreservoir engineering. The following are some of the reasons for low gas recovery: Drive Mechanism. In terms of drive mechanism, a frontal displacement-probably a gas-water contactalways results in a substantial residual gas saturation. This is often more than 40 percent in sandstones. In the case of near-total pressure maintenance by water 147
DETERMINATION OF OIL AND GASRESERVES
encroachment, more than 40 percent of the gas may be trapped behind the advancing gas-water contact.
Gas reserves in gas fields may be estimated by the volumetric and material balance methods.
Reservoir compaction drive in soft sediments has a similarly negative impact on gas recovery.
Volumetric Method
Over-Pressured Reservoirs. Over-pressured reservoirs, usually at considerable depth, can also have significant reductions in permeability to gas flow at abnormally high bottom-hole pressures during the gas exploitation process (Duggan, 1972). Phase Behaviour. If the reservoir temperature is less than the cricondentherm (maximum two-phase temperature), the potential exists for retrograde condensation of some of the heavier hydrocarbons as pressure declines and, therefore, a loss of valuable liquids.
Other Reasons. In addition, gas might be trapped due to the reservoir configuration, position and number of producing wells, production rates, water coning, migration offines, damage at the producing wellbore sandface, stratification, and loss of permeability due to facies changes. Low permeabilities often result in high abandonment pressures when reduced well spacing cannot be economically justified.
10.5
GAS RESERVES
"Gas reserves" refers to the fraction or portion of the original gas in place that is economically recoverable. Consequently, the recovery factor, RF, is defined as the ratio of gas reserves to initial gas in place and is usually expressed as a percentage: Gpo
RF= -
Gi
x 100
(I)
where Gp, = cumulative gas produced at abandonment conditions G, = initial gas in place Gas reserves are assigned to one of three groups:
I. Nonassociated gas reserves 2. Solution gas reserves 3. Associated gas cap gas reserves The determination of reserves of gas in these three groups is discussed in the following subsections.
10.5.1 Nonassociated Gas Reserves Determination Nonassociated gas reserves are those reserves that are not associated with recoverable oil reserves. Their production is limited only by market availability and contract terms.
148
The volumetric method is used for new gas fields before any significant production takes place. In reservoirs where no water influx is expected, recoverable raw gas, G, is calculated by the following: G=Ah.p(l-S.) - T~
r, r,
[Pi- - Po] Zi
z;
(2)
where G = original recoverable raw gas reserves (m") . A = drainage area (nr') h net pay thickness (m) porosity (fraction) .p Sw = connate water saturation (fraction) Tsc = base or standard temperature CK) 0 (273 + 0c) Pso = base or standard pressure (kPaa) T f = formation temperature CK) 0 (273 + 0c) Pi = initial reservoir pressure (kPaa) Zi = compressibility factor at Pi and T f P, = abandonment pressure (estimated) (kPaa) Z, = compressibility factor at P, and T f The base pressure used varies from 99.284 kPaa to 103.594 kPaa, but is usually 101.325 kPaa. The base temperature is normally 15°C (288°K). Abandonment pressure, P" can be estimated by the following rule of thumb: P, = 240 kPaa + 80 kPaaflOO m of depth The initial gas in place in the reservoir, minus the remaining gas at the selected abandonment pressure gives the recoverable raw gas as shown in Equation (2). In water-drive reservoirs, a residual gas saturation, Sgr' remains in the water-invaded zone. The recoverable gas, G, from the water-invaded portion of the reservoir is calculated by Equation (3).
T" [(I-S.)P, Sg,P o]
G=Ah.p-
r., r,
z,
--
z,
(3)
If water invasion of the reservoir amounts to less than 100 percent at abandonment, a higher effective residual gas saturation for the reservoir will result.
~!
DEPLETION MECHANISMS FOR NATURAL GASRESERVOIRS
Material Balance Method This method is applicable only to the reservoir as a whole, because of the migration of gas from one portion of the reservoir to another in both volumetric and water-drive reservoirs. For single-well reservoirs this method may be used directly, but in multiple-well pools the production information must be combined. The Law of Conservation of Mass may be applied to gas reservoirs to give the material balance as follows: mass of gas produced = initial mass of gas remaining mass of gas For the gas system under consideration, if the gas composition is constant, the number of moles of gas, both produced and remaining in the reservoir, is directly proportional to their masses. A material balance in terms of moles of gas may be written as follows: !1>=n;-nf
If there is a water drive, the final volume, Vf after producing a volume of gas, Gp , is: Vf = Vi - We + BwWp
(5)
where V f = final gas pore volume (does not include connate water) Vi = initial gas pore volume (does not include connate water) We= volume of water that has encroached into the reservoir at the final pressure
Pf Bw= the formation volume factor for water in reservoir volume per surface volume Wp = volume of water that has been produced from the reservoir =
ZnRT is applied in Equations
P;V, _ P,(V,-W,+BwWp ) Z,T
Z,T
Z,T
(7)
or G = P;V,T" _ V,T". P, p Z,P"T P"T Z,
(8)
For fixed values of Psc and Tsc, since Pi' Z, and Vi are also fixed for a given volumetric reservoir, Equation (8) may be written as follows:
P,
Gp =b-mZ
(9)
r
where
(4)
where subscripts p, i and f stand for produced, initial and final remaining at some later stage of production rather than at abandonment.
If the real gas law PV (4) and (5):
P;V, PrV,
-----
Z,T
(6)
where G, = volume of produced gas at standard pressure, P'o> and standard temperature,
Tsc If there is no aquifer present in the reservoir, there is no water influx and water production will be negligible. Then Equation (6) may be written as follows:
and
Equation (9) is the equation of a straight line, and indicates that for a volumetric gas reservoir the graph of the cumulative gas production, Gp, vs. the ratio P/Z is a straight line of negative slope "m." Figure 10.5-1 shows a plot of P/Z vs. cumulative gas production. The plot can be extrapolated to zero pressure to determine the initial gas in place or to any abandonment P/Z to find the recoverable gas. For the computation of initial in-place gas for constantvolume reservoirs, the following data is required: • Initial reservoir pressure • Cumulative gas volume • Stabilized shut-in reservoir pressure at the end of ' production • Gas deviation factors at these two reservoir pressures assuming the reservoir temperature remains constant This method is not applicable to water-drive gas reservoirs. With pressure reduction, when water enters the space occupied by gas, the pressures are maintained either almost completely or only in part depending on the nature of the water drive (Figure 10.5-2). In reservoirs where an aquifer provides a high degree of pressure support, the existence of a water drive is generally quite obvious. In reservoirs with only a partial pressure support, an active water drive may not be
149
"4 DETERMINATION OFOILAND GAS RESERVES
10.5.2 Solution Gas Reserves Determination Solution gas reserves are dissolved in the oil in a reservoir and can only be recovered if oil is produced. If solution gas cannot be conserved or sold, regulations may necessitate that the oil production be shut in. The rate of solution gas production depends on the rate of oil production and the producing gas-oil ratios (GORs).
N
ii:
I Cumulative Gas Production
Figure 10.5-1
Plot of P/Z vs. Cumulative Gas Production
apparent. A plot of P/Z vs. cumulative gas production in these reservoirs will indicate an overstated extrapolation of recoverable gas.
Complete Water Drive
If decline analysis is used to predict oil production, an extrapolation of the GOR trend can be conducted concurrently. More rigorous prediction methods can also be utilized as described in Section 9.2.1. As a rule of thumb, the ultimate solution gas recovery factor in solution gas drive reservoirs generally ranges from 50 to 65 percent. In oil reservoirs with an active water drive or waterfloods, the final recovery factor for the solution gas will be influenced by the degree ofpressure maintenance and sweep efficiencies, as well as residual oil and gas saturations.
10.5.3 Associated Gas Reserves Determination
t
Cumulative Gas Production
Figure 10.5-2
Effect of Water Drive on Pressure Decline
Models are available that use Equation (5), the basic material balance equation, for water drive reservoirs. An example is shown by Guerrero (1968). However, there are multiple unknowns in the material balance equation for water influx reservoirs, and calculations generally involve several assumptions on the reservoir description. Consequently, material balance predictions are often unreliable when a detailed understanding of the reservoir and supporting aquifer does not exist.
150
During the initial stages of oil production, GORs will generally remain at or above solution GOR until the critical gas saturation is reached. At this point the producing GOR will increase as described in Section 9.1.2.
The term "associated gas reserves" refers to a gas cap above oil reserves. Most, if not all, of the gas cap drive energy is required to maximize oil recovery. For this reason, associated gas reserves must ideally remain shut in until all the oil reserves have been produced. These gas reserves will be recovered during blow-down ofthe gas cap. Associated gas reserves are generally estimated using the volumetric method and an estimated abandonment pressure. As a gas cap adds inherent complexities to an oil reservoir, its presence may justify a more rigorous analysis or reservoir simulation to determine the appropriate depletion approach.
10.6
PIPELINE GAS RESERVES
The methods discussed in this chapter give reserves of raw gas. Before the gas is delivered to the point of sale, there are losses at the surface due to processing shrinkage and fuel consumption. These losses must be deducted from the raw gas reserves to calculate marketable pipeline gas.
--------------
.~;
41
DEPLETION MECHANISMS FOR NATURAL GAS RESERVOIRS
10.8
In sweet, dry gas fields, the surface loss is usually about 2 to 5 percent. For wet or sour gases, the surface loss can be estimated from the gas analysis, the recoveries of related products that are expected, and an allowance for plant fuel.
10.7
Rawlins and Schellhardt (1935) demonstrated that a gas well can be tested to predict its deliverability against a specific bottom-hole flowing pressure. An empirical relationship has been developed to relate the well gas flow rate at surface conditions with bottomhole flowing pressure and average reservoir shut-in pressure:
RESERVES OF RELATED PRODUCTS
Natural gas liquids and sulphur are recovered from the natural gas, and the reserves are estimated from the gas analysis and the gas reserves.
Q"
=
C (PR2·pl)"
(10)
where Q,,= flow rate at standard conditions of pressure and temperature C = a coefficient that describes the position of the stabilized deliverability line PR = average reservoir shut-in pressure Pp = reservoir flowing pressure n = an exponent equal to the reciprocal of the slope of the stabilized deliverability line
10.7.1 Natural.Gas Liquids For the development of reserve estimates, natural gas liquids are defined as those hydrocarbon liquids that, in the reservoir, are either gaseous or in solution with crude oil and that are recoverable as liquids by condensation or absorption in field separators, scrubbers, gasoline plants, or cycling plants. Natural gasoline, condensate, and liquefied petroleum gases are in this category.
Limits of n vary from 0.5 for fully turbulent to 1.0 for completely laminar flow in the formation, reflecting the degree ofturbulence.
Natural gas liquids are in a sense an intermediate product-lighter than what is usually considered crude oil and heavier than what is usually considered natural gas.
The P/Z vs. cumulative gas production relates the static reservoir pressure to cumulative gas. The results of isochronal (back pressure) testing relates static reservoir pressure, well flow rate, and sandface flowing pressure (Figure 10.8·1).
Natural gas liquid recoveries can be estimated as shown in Table 10.7·1.
10.7.2 Sulphur Sulphur is recovered as a by-product if hydrogen sulphide is present as an impurity in the natural gas. Sulphur recovery can be estimated as shown in Table 10.7·1.
Table 10.7·1
GAS DELIVERABILITY FORECASTING
Well performance estimates are made during the development stage of the gas reservoir and also during the depletion of the gas field. Basically, this involves establishing well production rates vs. reservoir pressure (gas well deliverability) that exist during the life of the gas reservoir (Figure 10.8·2).
Recoveries of Related Products
Related Product
SI Units
Imperial Units
Recovery Fraction Range
For Shallowcut use
For Deepcut use
Propane
m'/IO'm' (raw gas)
=
bbl/lO'cf vol. % x 36.9 x recovery (fraction) (raw gas)
=
vol. % x 6.54 x recovery(fraction)
o to 0.90
0.50
0.90
Butane
m3/10'm3 (raw gas)
=
bbl/.JO'cf vol. % x 43.0 x recovery (fraction) (raw gas)
=
vol. % x 7.62 x recovery (fraction)
o to 0.95
0.75
0.95
=
up to 1.00 vol. % x 10.15 x recovery(fraction)
0.95
1.00
Pentanes Plus m3/lO'm3 = vol. % x 57.3
bbl/lO'cf x recovery (fraction) (raw gas)
(raw gas) Sulphur
m3/IO'm3 (raw gas)
=
bbl/lO'cf = vol. % x 0.377 vol. % x 13.6 x recovery (fraction) x recovery (fraction) (raw gas)
0.95 to 1.00
Source: After Gas Processors Suppliers Association, 1981,
151
?
-DETERMINATION OFOILANDGASRESERVES
P~----Shut-In Reservoir Pressure Stabilized Deliverabilily Curve
i!!
" Ul Ul
i!!
n,
10
0>
e
.~
u:: ~
.~
" Ul
a: " 10.1 +----r:-----.:---,-L----1 2 10 10 10' 10' Gas Flow Rate
Figure 10.8-1
o
Gas Flow Rate
AOF
Back Pressure Plot
Similarly a wellhead gas deliverability plot in wellhead flowing pressure vs. gas flow rate can be generated from the wellhead back pressure plot. Further discussion on back pressure testing is beyond the scope of the monograph. For further details, the reader is referred to Theory and Practice ofTesting of Gas Wells (Energy Resources Conservation Board, 1975), Back Pressure Test for Natural Gas Wells (Railroad Commission of Texas, 1972) and "Methods for Predicting Gas Well Performance" (Russell et aI., 1966).
10.9
0+------------+-
WELL SPACING
Optimum well spacing for the exploitation of gas reservoirs may be substantially different than for oil reservoirs. Where spacing regulations govern, spacing would normally be wider for a gas reservoir than for an oil reservoir. These regulations recognize the increased mobility of gas as compared to oil, and the corresponding greater migration capability ofgas during producing operations; thus the spacing assigned to a gas well is considerably greater-typically, 259 hectares (640 acres) per well. However, a denser well spacing may exist in areas with shallow,low-permeability reservoirs.
10.10 CYCLING OF GAS CONDENSATE RESERVOIRS WITH DRY GAS Incentive exists in cycling of gas condensate reservoirs with "dry" gas in those cases where natural depletion of the reservoir will result in substantial loss of liquid hydrocarbons in the reservoir. This occurs in
Figure 10.8-2
Gas Deliverability Plot
volumetric reservoirs where retrograde condensation behaviours exist (liquids forming as the pressure declines), and in water-drive gas fields where "wet" gas is trapped. It has been noted that liquid hydrocarbons formed during pressure depletion of a reservoir are not normally revapourized at lower reservoir pressures and, therefore, are trapped as a residual liquid saturation. Under these circumstances, the gas in the reservoir may be "cycled" to reduce the loss of liquids. In this operation, gas is produced from the reservoir, the liquid hydrocarbons are extracted, and the dry gas is reinjected. This reduces the rate of pressure reduction in the reservoir, which is responsible for the retrograde condensation. The dry gas re-injected may be only part of the gas produced, or it may be all of the gas produced, or it may even be gas in excess of that produced, so that the reservoir voidage is fully replaced. There is evidence (Smith and Yarborough, 1968) that at least part of any liquid saturation that formed prior to the implementation of dry gas cycling operations, will be revapourized into the dry gas. To achieve maximum benefit from dry gas cycling, cycling should be initiated before the dew point of the reservoir hydrocarbon fluid is reached. In reservoirs where rock characteristics are favourable, cycling with dry gas should provide recovery of part of the liquids which otherwise would be lost. Not all cycling projects are successful. Sprinkle et al. (1971) have reported the adverse influence of stratifi-
152
s
DEPLETION MECHANISMS FOR NATURALGAS RESERVOIRS
cation on gas cycling operations. The presence of a high permeability layer in the reservoir was believed to be the cause for poor liquid hydrocarbon recoveries and resulted in the ultimate abandonment ofthe gas cycling project in a Texas Gulf Coast Frio Sand reservoir. Income from dry gas cycling projects will initially be all or mainly from liquid hydrocarbon sales and, later, during "blow-down," from the sale of both gas and liquids, but the rate ofliquid recovery will be declining.
10.11 SECONDARY RECOVERY OF GAS Secondary recovery of gas is uncommon because primary recovery usually yields a high percentage of the gas originally in place (70 to 90 percent). New operating practices, however, have sometimes made commercial deposits out of some that were considered to be uneconomic. Boyd et al. (1982) described secondary gas recovery from the watered-out Frio gas reservoir in the Double Bayou Field, Chambers County, Texas.
10.12 ENHANCED GAS RECOVERY Enhanced gas recovery has been traditionally used to describe methods of unconventional gas recovery from tight gas sands, Devonian shales, coal-bed methane, and methane from geopressured aquifers. However, many difficult problems such as technology, risk and economics remain barriers to progress in this direction. In general these reserves are not commercially viable without subsidy. Higher recovery from conventional gas reservoirs is a more likely place to look for additional gas and gas condensate production. The largest obvious source of gas from discovered reservoirs would be those reservoirs that have had strong water drives. It is worth mentioning that approximately two-thirds of the gas reservoirs of the world have an original gas-water contact, and approximately 50 percent of these reservoirs have at least a partial displacement with water.
References Boyd, W.E., Jr., Christian, L.D., and Danielsen, c.L. 1982. "Secondary Gas Recovery from a WateredOut Reservoir." Paper presented at the fall SPE meeting, New Orleans, LA., Sep. 1982, SPE No. 11158. Clark, N.J. 1960. "Elements of Petroleum Reservoirs." SPE of AIME, Dallas, TX. Craft, B.C., and Hawkins, M.F. 1959. Applied Reservoir Engineering. Prentice-Hall, Inc., Englewood Cliffs, N.J. Duggan, J.O. 1972. "The Anderson "L" - An Abnormally Pressured Gas Reservoir in South Texas." JPT, Feb. 1972. Energy Resources Conservation. Board. 1975. Theory and Practice ofthe Testing ofGas Wells. 3rd ed., Calgary, AB, Canada, Second Printing, 1978. Gas Processors Suppliers Association. 1981. Engineering Data Book (9th ed., 5th rev.). Guerrero, E.T. 1968. Practical Reservoir Engineering, The Petroleum Publishing Co., Tulsa, OK. Katz, D.L., Cornell, D., Kobayashi, R., Poettman, F.H., Vary, J.A., Elenbass, J.R., and Weinaug, C.F. 1959. Handbook ofNatural Gas Engineering. McGraw-Hill Book Co., New York. Railroad Commission of Texas. 1972. Back Pressure Testfor Natural Gas Wells. Oil and Gas Engineering Department, State of Texas. Rawlins, E.L., and Schellhardt, M.A. 1935. Back Pressure Data on Natural Gas Wells and Their Application to Production Practices. US Bureau of Mines, Monograph 7. Russell, D.G., Goodrich, J.H., Perry, G.E., and Bruskotter, J.F. 1966. '.'Methods for Predicting Gas Well Performance." JPT, Jan. 1966, pp. 99-108. Smith, L.R., and Yarborough, L. 1968. "Equilibrium Revaporization of Retrograde Condensate by Dry Gas Injection." SPEJ, Mar. 1968, pp. 87-94. Sprinkle, T.L., Merrick, R.J., and Caudle, RH. 1971. "Adverse Influence of Stratification on a Gas Cycling Project." JPT, Feb. 1971, pp. 191-194.
153
D
Chapter 11
ENHANCED RECOVERY BY WATERFLOODING
11.1
INTRODUCTION
Waterflooding is the process of injecting water into a formation for the purpose of displacing oil to producing wells. The displacement of oil by water is governed by wettability, pore size distribution and geometry, rock heterogeneities, and fluid properties. Waterflooding is a proven technology to improve recovery, but the degree of improvement and economic viability is dependent upon the following: • The type of flood scheme implemented
11.2.1
Mobility Ratio
D' Arcy developed an empirical relationship for the velocity of a fluid through a porous medium as a function ofpressure differential, viscosity, and a proportionality constant (permeability). The mobility ofa fluid is the effective permeability of the rock to that fluid divided by the viscosity of the fluid. For a frontal displacement scheme, the mobility ratio, M, is the ratio of the mobility of the displacing phase behind the flood front to the displaced phase ahead of the flood front.
• Properties of the reservoir rock (I)
• Properties of the oil • Well spacing • Economic factors (i.e., cost of the scheme, oil price, royalties, regulatory constraints) Waterflooding is classified as "secondary" recovery because it supplements recovery of oil by natural or "primary" depletion. In certain reservoirs, mobility ratios are improved by the addition of polymers, and interfacial tension is reduced by the addition ofsurfactants to the injected water. These processes are "tertiary" recovery schemes and are referred to as "polymer" and "micellar" flooding, respectively. Based on a statistical review of waterfloods in western Canada, total recovery factors generally vary from 16 to 45 percent with an average of 30 percent of original project oil in place. These values are typically at least double the primary recovery factor values. This chapter reviews the waterflooding process, the industry methods used to estimate reserves and production forecasts and the factors that affect the results, the accuracy of these methods, when and how to apply the them, and typical statistical data.
11.2
DISPLACEMENT PROCESS
The displacement process is governed by several fundamental principles that include mobility ratio, interfacial tension, and fractional flow.
where
I<.w =
relative permeability to water k,o = relative permeability to oil Ilw = water viscosity (cp) Ilo = oil viscosity (cp)
For a waterflood scheme, water mobility is determined at the average water saturation at water break-through. Oil mobility is determined at the initial connate water saturation. Mobility ratios for water displacing oil generally vary from 0.1 to 10. Increased mobility ratios have a detrimental effect on displacement, areal sweep and vertical sweep efficiencies, as discussed in the following subsections.
11.2.2
Interfacial Tension
Interfacial tension is a thermodynamic property of an interface between two phases. Typical values of interfacial tension between oil and water at reservoir conditions range from 10 to 30 dynes/em. Interfacial tension generally increases with increasing molecular weight ofthe reservoir fluid and decreases with increasing reservoir temperature. In water-wet rocks, interfacial tension tends to create bubbles of oil that block pore throats. In oil-wet rocks, interfacial tension tends to bind the oil to the rock surface. Interfacial tension is one of the major reasons why oil becomes increasingly more difficult to recover as water saturation increases.
154
---
1
ENHANCED RECOVERY BYWATERFLOODING
Over the range of interfacial tensions encountered in waterflooding, residual oil saturations are relatively constant. Residual oil saturations decline when interfacial tension is reduced to less than one dyne/em and approach zero when interfacial tension is approximately 0.001 dyne/em.
using relative permeabilities to oil and water determined in laboratory tests. Frontal advance theory and the application of fractional flow curves are presented in considerable depth by Craig (197Ia) and Willhite (1986). Typical fractional flow curves are illustrated in Figures 11.2-1 and 11.2-2.
11.2.3
These fractional flow curves illustrate that the displacement of oil from a water-wet rock is more efficient than from an oil-wet rock. Water injection and water production volumes will be higher for an oil-wet reservoir than for a water-wet reservoir.
Fractional Flow
Fractional flow is the fraction of the total fluid flow that is due to the flow of the displacing phase, and is a function of the saturation of the displacing phase. The simplified form of the fractional flow equation, excluding gravity and capillary forces, is as follows: 1 (2)
where fw
=
fractional flow
Fractional flow is a function of water saturation since relative permeabilities to oil and water are functions of water saturation. Fractional flow curves are constructed
It is noted that the fractional flow following breakthrough represents the producing water cut at the sandface. In single layer displacement, remaining oil saturation to waterflooding should be determined from fractional flow curves at the estimated economic water cut limit. In multi-layerdisplacement, it is common practice to assume that the residual oil saturation to waterflooding is equal to the endpoint saturation from relative permeability data. This is a consequence of producing well oil cuts being maintained at economic rates by layers that have not broken through with water.
0.8
-• -'" 3: -
c: 0.7 ttl
0.6
0
~
0
0.5
u:
a; 0.4 c:
0 :;:::
o
!!1
0.3
u.
0.2 0.1
o
L-~===----'-_-'---J_-'-_---'
20
30 40 50 60 70 80 Water Saturation (% pore vol.) Source: Craig, 1971 a.
Figure 11.2-1
Effect of Oil Viscosity on Fractional Flow Curve, Strongly Water-Wet Rock
0~::::....J-......L_--'---'--'--
10
20 30 40 50 60 70 Water Saturation (% pore vol.) Source: Craig, 1971 a.
Figure 11.2-2
Effect of Oil Viscosity on Fractional Flow Curve, Strongly Oil-Wet Rock
155 n
DETERMINATION OFOILANDGASRESERVES
In practice, average remaining oil saturations will be slightly higher than endpoint residual oil saturation values due to economic limit constraints. In dipping reservoirs, fractional flow data are adjusted for gravity and capillary effects. For oil being displaced updip, the performance ofa waterflood improves as dip increases. Capillary pressure effects are assumed to be negligible for most reservoir flow systems.
11.3
TYPES OF WATERFLOODS
The two general types of waterflood schemes are classified by the primary direction of the displacement process, i.e., vertical or horizontal. Vertical Waterflood Schemes. Water is injected at wells completed at the bottom of the formation, and oil is produced at wells completed at the top of the formation (Figure 11.3-1). The higher density of water as compared to oil results in water gravitating to the bottom of the formation and displacing oil in an upward direction. Oil production
t
t
o interval Completion
--> --
I
Principal direction of oil displacement Oil
-----------
- ' -Water injection
Water
• Pembina Cardium • Wainwright - Sparky Swan Hills - Beaverhill Lake Steelman - Midale , 0 A----
I II I I
A Injector
0
0---- A
• 0 A---
Cross Section for Vertical Waterflood
0
I
Producer Principal direction of oildisplacement - horizontal
A
Figure 11.3-2 Plan View for Horizontal Waterflood Horizontal flood schemes are typically classified by the type of injection pattern. The most common, as illustrated in Figure 11.3-3, include the following: • Five-spot • Inverted nine-spot Line drive • Peripheral A combination of the vertical and horizontal processes is used in dipping reservoirs. Other types ofpatterns are discussed and illustrated by Craig (197Ib).
11.4 Figure 11.3-1
A
ANALYSIS METHODS AND WHEN TO APPLY THEM
There are five general types of reserve and production forecast methods for waterfloods in common use: 1. Volumetric analysis
This type of scheme is best suited to relatively thick formations and is most commonly applied to reef reservoirs such as the following in Alberta:
2. Decline performance analysis
• Rainbow, Virgo, Zama, Shekelie-Keg River
4. Analytical performance prediction
• Pembina, West Pembina-Nisku
5. Numerical simulation The volumetric method is used only to calculate reserves, whereas the other methods may be used to calculate reserves and production forecasts. Wherever possible, reserves should be calculated using more than one method in order to substantiate the results and increase confidence. The following subsections discuss the applicability of the methods at various stages of depletion.
Horizontal Waterflood Schemes. Water is injected in a pattern of wells, and oil is produced from wells completed between injectors (Figure 11.3-2). Pressure gradients caused by injection and production result in displacement of oil in a horizontal direction. This type of scheme is best suited to relatively thin or layered formations and is commonly applied to blanket or channel type sands as well as carbonate reservoirs such as the following reservoirs in western Canada:
3. Comparison to analogous pools
156
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _d
ENHANCED RECOVERY BYWATERFLOODING
A
<;>
I
A
Q
Q
I
-0- -
I
-?- -
I
-0-- -
{>
I
I
\>A\>A\> I
?- -
-0- -
I
I
I
I
I
-9-- -0-- - \>
6 A 6 A 6 Inverted Nine-Spot A Injection well
'f -0- -0- - 0 - - 9 I I I
I I I
I I I
I I I
o
Production well Pattern boundary
,A"
I I I
/ / /
I I I
I I I
I I I
I I I
I I I
I I I
I I I
-, <, 0
-,
/
~
/
-,
'f -6- -6- -6--Q I I I
<,
// / 0
r--b--b--k--* I I I
"
I I I
~-lr--lr--I:r-" Direct Line Drive Source: After Craig, 1971 b.
0
"
/
"0
" " -,
""
~
/ /
0 /
-,
/ /
""If
/ /
Peripheral
Figure 11.3-3 Flood Patterns for Horizontal Flood Schemes
11.4.1
Pool Discovery
At pool discovery, there is normally insufficient reservoir data to accurately calculate waterflood reserves by any method. If the reservoir is seismically defined, waterflood reserves may be calculated using volumetrics or analogies and are normally categorized as "possible" because of the considerable uncertainties in reservoir definition.
11.4.2
Delineated Pool: Immature Depletion
Once a pool has been delineated and on primary production for a reasonable period of time, waterflood reserves can be calculated more accurately since reservoir size and configuration will have been established, reservoir properties can be measured at various points across the pool, oil properties will have been established, and the primary depletion mechanism can be established. At this stage, volumetric, analogous comparison, performance prediction, and numerical simulation methods may be utilized. Properly assessed volumetric and analogy methods are reasonably accurate at this stage
and are normally used to assess waterflood feasibility. Performance prediction methods alone are only approximate, but are reasonably accurate if adjustments are made to fit volumetric reserves and analogies. Numerical simulation is commonly performed if waterflood feasibility has been established by analytical techniques. This technique is generally accurate for ultimate recovery predictions, "provided" reservoir properties are accurately defined and numerical effects are properly handled. Frequently, however, reservoir rock properties, layering and heterogeneities are not accurately known, and unreliable break-through predictions result. Analogies in these cases sometimes yield more reliable results if the analogous pools have similar heterogeneities and rock properties. If economically feasible, waterflood reserves at this time are frequently classified as "probable." Where strong analogies can be made to similar successful flood schemes, a portion of the reserves may also be classified as "proved." The degree to which proved reserves are assigned depends upon the type of reservoir, the reliability ofthe data, the commitment of
157
DETERMINATION OFOILAND GASRESERVES
the operator to implement a scheme, and the strength of the analogies.
11.4.3
Post-Injection Startup
After startup of injection, reserves are generally calculated in the same manner as that described in Section 11.4.2. Slightly higher confidence may be placed on the calculated results as water injectivity and potential premature break-through problems can be ascertained.
11.4.4
Post-Waterflood Response
After waterflood response has been exhibited (i.e., oil production increases and gas-oil ratio (GaR) decreases), more of the possible and probable waterflood reserves may be reclassified as "proved" or "probable." There is basically no change in the way volumetric, analogy and performance prediction methods are utilized at this stage of depletion; the only difference is in the confidence level ofthe results. Numerical simulation results become more accurate, however, as reservoir and rock properties are tuned to match actual response.
11.4.5
volumetrically ifhistorical oil-water contact movements are measured. Changes in contact levels compared to mapped pore volumes yield in situ determination ofdisplacement and sweep efficiencies which may be used to assess remaining reserves. Hydrocarbon pore volume or original oil in place vs. depth relationships are required to evaluate in situ recovery efficiencies.
11.5
VOLUMETRIC ANALYSIS
11.5.1
Overview of Method
The volumetric equation for the calculation of waterflood reserves is a relatively simple one (Slider, 1983a): N p=f ''"lE V [Sop 'tswB op
Mature vertical waterflood schemes may not have established oil production decline or water cut trends as a result of regulatory production rate limitations imposed on oil wells and manual restrictions to prevent water coning. Recoveries can be accurately predicted
(3)
or
where Npf
total waterflood reserves from commencement of the flood to abandonment (stm") average porosity within the gross swept area of the flood scheme E, = total sweep efficiency = EH x Ev x Ec EH = horizontal sweep efficiency (areal) Ev vertical sweep efficiency Ec = conformance efficiency (continuity) V sw = gross swept rock volume of the flood scheme (m") Sop = oil saturation within the gross swept volume at the start of the flood (fraction) Sor = residual oil saturation (fraction) Bop = oil FVF @ start of flood (m3/m3 ) B or = oil FVF @ abandonment of flood (m3/m3)
Mature Waterflood
Mature horizontal waterflood schemes exhibit trends of increasing water cut and declining oil production. Once the trends have been established, decline performance analysis may be used to calculate reserves and oil production forecasts. As discussed in Chapter 18, normally over 50 percent recoverable reserve depletion is required before decline analysis is performed. Properly assessed decline analysis is the most accurate conventional method to determine proved and probable producing reserves. Numerical simulation techniques can be more accurate, but the expense of performing the simulation may not be warranted unless operating and optimization strategies are being examined. Waterflood recoveries obtained from decline analysis are frequently rationalized volumetrically. This procedure will indicate whether all areas of the reservoir are being efficiently flooded. Additional nonproducing proved or probable reserves may be assigned to areas of the reservoir that require infill or delineation drilling, additional injection well conversions, or recompletion workovers to improve recovery.
~]
B
The equation is straightforward, but the derivation of each parameter of the equation may not be. Waterflood reserves are frequently confused with total reserves. Total reserves are equal to primary plus waterflood reserves. Similarly, total recovery factor is equal to the primary plus waterflood recovery factors. Typically, the total recovery factor for waterflood schemes is at least double that of primary recovery.
11.5.2
Parameters and Factors Affecting Analysis
The individual parameters that make up the volumetric waterflood equation are discussed in this section. More complete discussions are presented by Craig (1971), Willhite (1986) and Slider (1983).
158
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _51
ENHANCED RECOVERY BYWATERFLOODING
Horizontal Waterflood Schemes
100
1. Porosity Average porosity, cp, within the gross swept reservoir should be used in the calculation. It should be noted that this may not equal average pool porosity. 2. Total Sweep Efficiency Total sweep efficiency, E" has three components: horizontal efficiency, EH , vertical efficiency, Ev, and conformance efficiency, Ec . Many evaluators rearrange Equation (3) and incorporate SolBop- So/Borin a fourth component term, displacement efficiency, ED:
0..80 Q)
~
(f)
'" 70 ~
-c
""
60
50
(4)
~:::::JLU-lWlL_---L---l---LilWJ 0.1
1.0
10
Reciprocal of Mobility Ratio Source: After Craig, 1971d.
Some horizontal waterflood schemes exhibit piston-like oil displacement. The oil wells produce water-free until flood front arrival, and then water out within a few months. This behaviour can result from unstratified deposition, water-wet characteristics, or the overdisplacement of water injection volumes relative to pattern producing rates. When this behaviour occurs in a number of wells, the shape and position of waterflood fronts can be mapped, enabling in situ measurement of sweep efficiency by the comparison of swept pore volumes with either injected water or produced oil volumes. The accuracy of this method is largely a function ofthe accuracy of the mapped shape of the flood front.
3. Horizontal Sweep Efficiency Horizontal (or areal) sweep efficiency, EH , may be defined as the areal fraction ofa waterflood pattern contacted by injected water. This fraction is affected by pressure gradients, permeability trends, mobility ratios and injected volumes. Values of EH at water breakthrough for various waterflood pattern configurations have been determined through laboratory models in numerous studies (Craig, 197Ic). With continued water injection after break-through, EH increases as a function ofthroughput volumes until it reaches 100 percent. For volumetric reserve calculations, horizontal sweep efficiencies are determined for conditions at economic water cut limits. A number of design correlation charts have been developed to determine EH ; these are summarized by Craig (1971d). Figure 11.5-1 illustrates the correlation for a five-spot flood pattern. As can be seen from this plot, horizontal sweep efficiencies are generally over 90 percent since economic water cut, fw ' limits are typically greater than 95 percent.
Figure 11.5-1
Effect of Mobility Ratio on Oil Production for the Five-Spot Pattern
This plot also illustrates that EH decreases as mobility ratio increases. Thus, high viscosity oil (commonly low API gravity) reservoirs will have a lower EH and a lower recovery factor than similar low viscosity (high API gravity) oil reservoirs. Permeability trends must be addressed when horizontal sweep efficiency is being determined. Unfortunately, these are frequently not identified until after implementation of a waterflood scheme when wells on trend with water injectors prematurely water out. These problems are usually rectified by converting the scheme to a line drive waterflood with alternating rows of injectors and producers oriented along the permeability trend. For example, most Cretaceous Cardium reservoirs in west central Alberta have southwest to northeast permeability trends resulting from tectonic stress during the building of the Rocky Mountains. Horizontal sweep will also be affected by nonuniform pressure sinks at production wells. EH is normally not affected by gas saturations prior to waterflooding. However, if gas saturations are too high prior to waterflooding, cusping of the waterflood front at the producing well prior to fill-up may occur and adversely affect horizontal sweep efficiency. Of greater operational significance when high free gas saturation exists is the high reservoir voidage created by high GORs. To maintain cash flow, it is common practice to continue oil production during waterflood fill-up. High GORs result in high voidage replacement requirements and defer re-pressuring ifinjectivity is low or injector-to-producer ratios are low. 159
DETERMINATION OFOIL AND GAS RESERVES
4. Vertical Sweep Efficiency Vertical sweep efficiency, E y , accounts for incomplete sweep of reservoir layers at abandonment of the waterflood scheme. Incomplete vertical sweep is caused by the stratified nature ofmost reservoirs. Strata are flushed with water in descending permeability sequence. At economic water cut limits at the producing well, not all strata may be flushed with water, and the vertical sweep efficiency will then be less than 100 percent. Vertical sweep efficiencies are commonly calculated from methods that order flow capacity thicknesses and permeabilities from core analyses. The two most common techniques are the Stiles Method, which primarily concerns capacity thickness ordering (Slider, 1983b), and the Dykstra Parsons Method, which relates statistical variations in permeability with floodout behaviour of flood pot tests made on California core samples (Craig,197Ie). In both methods, Ev is a function of mobility ratio and permeability contrast. Ev decreases as mobility ratio and permeability contrast increase. Thus reservoirs with thin high permeability streaks have low vertical sweep efficiency. Care must be taken to ensure that the stratified reservoir assumption is valid in both methods. Some reservoirs that undergo post-depositional porosity alteration have high permeability contrasts on core, but these contrasts may be so random in nature that the reservoir will appear homogeneous, and piston-like displacement may occur with virtually 100 percent vertical sweep efficiency. Other factors that affect vertical sweep efficiency include gravity and cross-flow between layers. Due to gravity forces, water will tend to move at the bottom of the reservoir and, in uniform permeability distributions in horizontal reservoirs, this movement tends to decrease vertical sweep efficiency. Ifreservoir permeability decreases with depth, however, gravity forces will improve vertical sweep. In dipping and vertical reservoirs, gravity forces can be used to advantage by injecting downdip and displacing oil updip. Cross-flow between layers tends to improve Ev at favourable mobility ratios (low) and diminish it at unfavourable mobility ratios (high). 5. Conformance Efficiency Conformance efficiency, Ec , or continuity, is a term used to account for discontinuous reservoir pore volume. In the past, engineers widely assumed that all pore spaces in a reservoir are interconnected with each other, 160
but infill drilling results throughout North America indicate that reservoirs are less continuous than had been assumed. Generally Ec is difficult to quantify and is usually back-calculated in mature producing pools where reserves from decline analysis do not rationalize volumetrically using only vertical and horizontal sweep efficiencies. Without considering continuity, infill drilling programs technically do not usually increase ultimate recoverable reserves; they only accelerate recovery. In reservoirs with poor continuity, infill drilling will improve continuity and, therefore, reserves by accessing additional pore volume. A more complete discussion of infill drilling and continuity is presented by Gould and Sarem (1989). 6. Gross Swept Volume Gross swept volume, Vsw, refers to the reservoir rock volume that is subject to waterflood sweep. In a horizontal sense this includes the area within waterflood patterns and a portion ofthe area outside the waterflood patterns. A common error in waterflood analysis is to utilize entire pool volumes instead of gross swept volumes. A procedure for determining gross swept areas discussed by Slider (1983c) is dependent on the gas saturation existing at the start of a flood scheme. The higher the gas saturation at the start of the flood, the lower the swept fraction of oil outside the enclosed flood pattern. When reservoir permeability trends exist, they should also be considered when 'estimating gross swept areas. The vertical component of gross swept volume is frequently overlooked in volumetric waterflood analysis. Gross swept volumes should reflect layers which are receiving injected water volumes. In thick stratified reservoirs some layers may be ofpoor quality and may not be completed or may not be receiving injected water volumes due to formation damage.
7. Oil SaturatIon at Start of Flood Oil saturation, Sop, at the start of a flood for a solution gas drive reservoir may be determined using the following equation (Slider, 1983e): SOP =
(N - N pp ) Bop (1 - Sw) NB o;
where N = oil in place (stm') N pp = primary oil production (stm') Bop = oil FVF after primary depletion (m 3/m3 ) connate water saturation (fraction) Sw a, = initial oil FVF (m3/m3)
(5)
ENHANCED RECOVERY BYWATERFLOODING
Initial oil in place is calculated by either material balance or volumetric methods. Connate water saturation is measured by log analysis or capillary pressure tests. This Sop calculation assumes the saturation is uniform at the star! of the flood. 8. Residual Oil Saturation Residual oil saturation refers to the microscopic oil saturation left in reservoir rock. Because oil and water are immiscible, surface tension of fluids with reservoir rock results in incomplete displacement of oil by water. The efficiency of this displacement is a function of the reservoir wettabiIity and pore throat size and configuration. Residual oil saturations are most commonly determined by flooding reservoir core samples with multiple pore volumes ofwater in either steady-state or unsteady-state tests. Since Sor is dependent upon wettability, care must be taken to ensure that the rock samples do not have altered wettability properties as a result of core handling. The effects of core handling on wettability are discussed at length by Anderson (1986). The accuracy and reliability of Sor measurements generally decrease from native state to restored state to cleaned cores. The most significant conclusions from Anderson's literature survey are summarized as follows: • Removal of a core from the reservoir may increase oil wettability due to the decreased solubility of wettability-alteringcompounds as a result of temperature and pressure reduction. • Core flood tests conducted qt ambient vs. reservoir temperature and pressure may exhibit oil-wet characteristics resulting in a hig9 estimate of residual oil saturation. • Cleaning and drying of core samples prior to use in core flood tests tend to induce water wettability and result in a low estimate of residual oil saturation. Ideally, multiple core samples should be tested and averaged to determine Sor because most reservoirs are heterogeneous, and one sample may not be representative of the average reservoir. Due to cost and core availability considerations, this is not always feasible. In Alberta, the Energy Resources Conservation Board publishes a guide of nonconfidential core flood tests (Energy Resources Conservation Board, 1993). Values of Sor may thus also be estimated by analogy to other pools of similar geologic horizon in the same geographic area. Another means of estimating Sor whhe core flushing tests are not available is by examining average Sor
values from conventional core analyses. These values should be adjusted to reservoir conditions using the oil formation volume factor. In situ residual oil saturations are sometimes taken in waterflooded portions of reservoirs using log or sponge coring techniques. This could only be performed in a mature waterflood or pilot project. Residual oil saturation may also be affected by trapped gas saturations, Sgt, when initial gas saturations are present in the reservoir prior to waterflooding. Experimental studies discussed by Craig (1971 f) indicate a reduction in Sor with Sg! in water-wet rocks but not in oil-wet rocks. These correlations assume that no compression or resolution of gas occurs. In most waterflood schemes,gas saturation is reduced by re-pressuring which reduces the impact on residual oil saturation. Dardaganian (1985) discussed the effect of free gas saturation on waterflooding and a method for determining the optimum pressure at which to initiate a waterflood. Vertical Waterflood Schemes
In vertical waterflood schemes, gravity results in oil displacement across stratified layers; hence, Ev and EH are usually taken to be 100 percent. Gross swept volumes, however, should be adjusted downwards to reflect the following: Sandwich Loss. This is the volume of oil remaining at the top of the reservoir after waterflooding. As a result of water coning, not all of the reservoir can be swept with water before producing wells reach economic water cut limits. Also, ifthe reservoir is updip ofproducing wells, attic oil losses may result. Coning correlations have been developed (Kuo, 1989) to predict sandwich loss; however, they are highly dependent on mobility. Typically, sandwich losses can vary from 2 to 15 feet for mobility ratios of I to 10 respectively in unfractured reservoirs. Unswept Volumes (along the periphery of the pool). Under perfect gravity segregation in a homogeneous pool, water will areally displace the entire reservoir. However, discontinuities, permeability channels, and restrictions may limit volumes swept by the flood. These effects may be incorporated by reductions in either the swept volume or horizontal and conformance factors. Vertical waterflood schemes have been implemented in pinnacle reefs in northern Alberta where the pools have been essentially depleted under a primary recovery mechanism. The success of these schemes may be questionable as primary depletion has established 161
DETERMINATION OF OIL AND GASRESERVES
high gas saturations. These likely form gas caps that improve primary recovery to levels approaching typical secondary recovery values.
This greatly increases confidence in the calculations. Waterfloods of this type include vertical and horizontal bank displacement schemes.
In situ sweep efficiency in vertical waterflood schemes with a high degree of gravity segregation can be measured by comparing oil recovery to water-flushed hydrocarbon pore volume. The flushed pore volumes are determined using oil-water contact measurements from log analysis and pore volume vs. depth correlations. The accuracy of the sweep efficiency evaluation in this method is a function of the accuracy in the oilwater contact measurement and pore volume vs. depth plots.
11.6
DECLINE PERFORMANCE ANALYSIS
11.6.1
Overview of Method
Decline analysis is used to evaluate mature waterflood schemes. Historic decline trends are used to extrapolate future trends; the two generalized methods are oil rate and oil cut declines. Decline trends may be either expo. nential or hyperbolic. The mathematics of decline analysis are discussed in Chapter 18.
Factors affecting the accuracy of oil-water contact measurement include the following:
11.6.2
Completion Status. Is the log run in a cased or open hole? Cased hole interpretations are more subtle and difficult to interpret. Producing Status. Is the log run in an observation or producing well? Oil-water contacts measured in shutin producing wells may be higher than the pool average if the water cone is not allowed sufficient time to settle. • Variability of Measurements. Is measurement made at one or a number of wells, and does the value vary significantly? The more variation, the more interpretation required to derive a pool average value. The sensitivity ofthe calculated sweep efficiency to oilwater contact variations should also be checked to gain confidence in the answer, e.g., does a 0.91 m (3 foot) change in interpreted contact change calculated sweep efficiency 5 or 50 percent? Once sweep efficiency has been derived, the remaining uncertainties pertain to the determination of remaining gross swept pore volume, which is a function of the geological mapping and petrophysical properties of the reservoir.
11.5.3
Reliability of Results
The accuracy of volumetric reserve calculations is a function ofthe accuracy ofthe parameters in the analysis. An engineer using volumetric analysis must assess the uncertainties in the parameters when assigning proved and probable reserve values. There is usually considerable uncertainty in assessing volumetric parameters in proposed flood schemes prior to implementation. These uncertainties diminish as additional data is gathered. In certain types of mature waterfloods, total sweep efficiencies can be determined from performance data.
162
Factors Affecting Analysis
Oil decline trends in waterfloods are a consequence of increasing water cuts and constant or declining wellbore flow capacity. The shape of the decline trend is a function of relative permeability, mobility and reservoir permeability variation. Ideally, under stable conditions, extrapolation of oil cut and oil rate declines should yield the same reserve value. Economic oil cut and oil rate limits should be determined for design fluid lifting capacity and used as endpoints in the decline analyses. Pools with a high degree of stratification, permeability variance, or dual porosity behaviour will tend to decline in a hyperbolic or harmonic fashion. Most reservoirs, however, exhibit exponential decline behaviour. For horizontal floods, the following points should be considered when analyzing production decline trends: I. Total fluid production should be plotted along with oil cut and oil rate data. Increasing fluid rates may be achieved by increased drawdown at producing wells, increased reservoir pressure through overinjection, or well stimulation. While total fluid rates are increasing, oil rate decline trends are dampened. The use of oil cut trends is preferred in these cases as oil rate trends will yield optimistic results. Conversely, if total fluid rates are declining, the use of the oil rate decline trend will yi~ld conservative results unless the fluid rate dechne cannot be arrested. 2. When wells are grouped for decline analysis, care should be taken to ensure that wells within the group have experienced water break-through. A preferred method is to group wells with similar water cut.s. This ensures that there will be no sudden change III decline behaviour as a result of water break·through.
ENHANCED RECOVERY BYWATERFLOODING
3. Generally, oil cut trends should not be used ifwater cuts are still less than 50 percent.
11.7
COMPARISON TO ANALOGOUS POOLS
4. Infill wells should be grouped separately for decline analysis. Decline trends of the infills and original wells, pre- and post-infill drilling, can be used to assess incremental recovery associated with the infill program.
11.7.1
Overview of Method
5. Yearly voidage replacement ratios should be checked when decline trends are being analyzed. Underinjection will cause gradual pressure loss, accelerated oil rate declines and dampened oil cut declines. The reverse is true for overinjection. This is most sensitive in low GOR and low API gravity oil reservoirs. 6. Decline analysis can be used as a diagnostic tool. Declining fluid production rates when voidage is being maintained may be due to formation damage or pumping equipment failure. 7. Producing conditions should be verified when declines are being analyzed to ensure that declines are real and have not been imposed by operating constraints. These comments also apply to vertical waterflood schemes, which generally exhibit a more sudden waterout behaviour. Flood-out is controlled by coning rather than by stratification characteristics. Thick vertical flood schemes will exhibit a relatively extensive water-free production period followed by a steep decline trend. When reserves are determined by decline analysis in vertical flood schemes, completion intervals should be checked to ensure that wells are completed at the top of the productive zone. If not, additional reserve assignments are warranted.
11.6.3
R.eliability of Results
Decline analysis is one of the more reliable methods of estimating reserves. The reliability of the method in· creases with the maturity ofthe pool and the smoothness of the data. At the start of production decline, when trends are not clearly established, interpretation of declines may vary from engineer to engineer. Most engineers adhere to exponential trends until hyperbolic trends can be confidently quantified. Engineers with experience in analyzing decline trends of pools similar in nature to the subject pool may have more confidence in assessing a certain type of decline and thus may use a different trend than an engineer with less experience.
Predicting waterflood performance by the analogy method refers to the comparison of a previous mature waterflood project to a proposed or current project in order to predict results ofthe proposed or current project. The method is usually reliable and is best applied in conjunction with the volumetric method. The method can be used to determine recoverable reserves as well as production and injection forecasts. A rigorous analogy involves comparison of all volumetric recovery parameters.
11.7.2
Procedure and Factors Affecting Analysis
The first step in a rigorous analogy analysis is to rationalize the volumetric parameters in the analogy pool. The recoverable reserves in the analogous pool should be well-established from decline analysis. The oil in place should also be well-established from mapping and volumetric calculations. The volumetric parameters should be determined as described in Section 11.5. The only unknown variable that is not definable empirically is the conformance efficiency Ee. Once all the other volumetric parameters have been derived, this value can be back-calculated to match recoverable reserves. In addition to continuity, this factor will include any error or anomalies associated with the determination of the other parameters in the volumetric equation. The next step of the analogy procedure is to calculate reserves of the proposed or current waterflood scheme using the volumetric method. If differences in mobility ratios, oil saturations and permeability variations exist between the analogy and the proposed waterflood project, these differences should be incorporated in the analysis. The conformance efficiency of the analogy project should be applied to the proposed project. The underlying assumption in the analogy method is that the continuity and anomalies associated with the analogous pool will also apply to the proposed or current waterflood scheme. When the analogy project is similar to the proposed project in terms of geological deposition and oil gravity, the analogy method is usually simplified by comparing the recovery factor ofthe analogy project to the proposed project. This is also often performed when the specific volumetric parameters are not well-defined
163
DETERMINATION OF OIL AND GASRESERVES
in the analogy or proposed waterfloodproject due to an absence of reliable laboratory core tests. Analogies can also be used to predict production and injection performance. Rigorous application of ratedependent analogies is describedby Slider(1983d). This procedure is useful in estimating waterflood response time, magnitude of oil productivity improvement and flood-out behaviour after response. The procedure involves plotting oil rate divided by effective injection rate vs. cumulative effective injection divided by ultimate flood recovery. This plot provides a normalized relationship that can be applied to a flood scheme of any size.
11.7.3
Reliability of Results
When analogies are applied in an analysis, it is good engineering practice to provide a comparison of reservoir properties so that the reader can judge the strength of the analogies being made. The strength of the analogy is critical to the assessment of proved or probable reserves. The closer the analogy project to the proposed project in terms of geographic area, geologic horizon, oil viscosity, waterfloodpattern and orientation,permeability variation, residual oil saturation and degree of depletion prior to waterflooding, the strongerthe analogyand the more reliable the results. An analogyshouldbe chosen that is typical of performance and not one that is clearly the best or worst performance. Analogiesare best utilized prior to or immediatelyafter implementationof a waterflood project.
11.8
ANALYTICAL PERFORMANCE PREDICTION
11.8.1
Overview of Methods
The analytical methods summarized in Table 11.8-1 yield production and injection forecasts for horizontal waterflood schemes. The Higgins-Leighton (1962) Methodhas fewerlimitingassumptionsthan othertechniques and is adaptable to varioustypes of patterns.The method models a flood pattern as a series of parallel stream flow tubes and is available in computerprogram format. For composite injection and producing rate, WaR and recovery vs. time, Craig (1971 g) recommended the use of the Craig-Geffen-Morse (Table 11.8-1) Method coupled with the Caudle and Witte (1959) correlation for injection rates. This procedure splits a waterflood forecast into four stages:
164
Stage I Periodprior to interference of oil banksaround injectors Stage 2 Period from interferenceto fill-up of gas pore space Stage 3 Period from fill-up to water break-through Stage 4 Period from water break-throughto flood-out During Stage I, water injection rates are calculated using radial flow equations. Water injection rates during Stage 2 are calculated using the Caudle and Witte conductance ratio. Oil production in Stages I and 2 is assumed to be negligible or zero. If oil production is significant, then adjustments are made to fill-up times and volumes. Oil production during Stage 3 is equal to water injection rates. Afterwaterbreak-through in Stage 4, the following are calculated: • Horizontalsweepefficienciesas a functionof breakthrough areal sweep and injected water volumes • Water-oil ratios from frontal advance theory • Injectivities from Caudle and Witte Oil producing rates from producing WaRs and injection rates The method can handle multi-layer effects by normalizing injection, production, and pore volume data for each layer and multiplying the results by single layer calculations.
11.8.2
Reliability of Results
All waterflood predictive methods have underlying simplifyingassumptions. The accuracy ofthe methods therefore relies on the validity of these assumptions in additionto the accuracyof the reservoirdescription. The Craig-Geffen-Morse Method is one of the more rigorous methods; however, the following limiting assumptions apply: • Waterflood response is injectivity-driven. • Gravity effects are negligible. • There are no cross-flow effects. • There is no lateral variation in reservoir properties. • Reservoir continuity is 100 percent. • Oil production is negligible prior to fill-up. • Capillary effects are negligible. • There is no bottom water or gas cap. The engineer must judge the significance and validity of these assumptionsto the reservoir being analyzed. It is recommended that the production profile resulting fromthepredictive methodbe adjusted to matchreserves calculated by volumetric, decline or analogy methods.
_________.a
ENHANCED RECOVERY BYWATERFLOODING
Table 11.8-1
Classification of 33 Waterflood Prediction Methods
Basic Method
Modification
A. Methods primarily concerned with permeability heterogeneity-injectivity 1. Dykstra-Parsons (1950) (a) Johnson (1956) 2. Stiles (1949)
(b) Felsenthal-Cobb-Heuer (1962)' (a) Schmalz-Rahme (1950)' (b) (c) (d) (a) (b)
Arps ("Modified Stiles") (1956) Ache (1957) Slider(1961) Muskat (1950) Prats et al. (1959)'
C. Methods primarily concerned with the displacement process 1. Buckley-Leverett (1942) (a) (b) (c) (d) (e) (I) 2. Craig-Geffen-Morse (1954) (a) 3. Higgins-Leighton (1960-1964)
Terwilliger et al. (1951) Felsenthal-Yuster (1951) Welge (1952) Craig-Geffen-Morse (1954)3 Roberts (1959) Higgins-Leighton (1960-1964)' Hendrickson (1961)
3. Yuster-Suder-Calhoun (1949) 4. Prats-Matthews-Jewett-Baker (1959) B Methods primarily concerned with areal sweep efficiency 1. Muskat (1946) 2. Hurst (1953) 3. Atlantic-Richfield (1952-1959) 4. Aronofsky (1952-1956) 5. Deppe-Hauber (1961-1964)
D. Miscellaneous theoretical methods 1. Douglas-Blair-Wagner (1958) 2. Hiatt (1958) 3. Douglas-Peaceman-Rachford (1959) 4. Naar-Henderson (1961) 5. Warren-Cosgrove (1964) 6. Morel-Seytoux (1964) E. Empirical methods I. Guthrie-Greenberger (1955) 2. Schauer (1957) 3. Guerrero- Earlougher (1961) Source: After Schoeppel, 1968. Note: Complete citations for all of the references listed in this table are given at the end of the chapter. I Also applies to Stiles method. 'Also applies to Yuster-Suder-Calhoun and Schauer methods. 3Also concerned with areal sweep problem. Also recognized as basic method.
165
q
I
DETERMINATION OF OIL AND GASRESERVES
Predictive methods are normally applied at the waterflood design stage to assist in scoping economics and facility design rates. The methods can be used, however, at any stage of waterflood depletion and history-matched to actual performance by tuning reservoir rock properties. The reliability of the procedure increases if this is performed.
Sensitivity studies on gridblock sizing should be performed to ensure that the selected sizing is sufficiently accurate. Increased grid definition should be used in highly heterogeneous areas and around wellbores.
11.9
NUMERICAL SIMULATION
11.9.1
Overview of Method
If reservoir projects are fairly consistent across a proposed waterflood area, partial pattern waterflood simulations are frequently performed and the results scaled up to reservoir dimensions. The examination of a partial pattern can result in better grid definition for more accurate results.
The most advanced method for determining waterflood reserves and performance predictions is numerical simulation, which can be described as the use of digital computers to numerically solve mathematical models representing physical reservoir systems. Simulation techniques are discussed in Chapter 17. Aspects of'simulation that are relevant to waterflooding are presented in this section.
11.9.2
Parameters and Factors Affecting Analysis
The following factors may affect numerical simulation results: Model Phases. Waterflood simulations are performed using Beta or black oil models. When the reservoir is above the bubble point, only two phases (oil and water) are required. Ifthe reservoir is below the bubble point, then three phases are required (oil, water and gas). The relative permeability and physical properties of the phases are required in the simulation. The physical properties are usually easily measured and accurate; however, the accuracy of relative permeability data is less reliable and can significantly influence the simulation results. Model Dimensions. Waterflood simulations are typically two- or three-dimensional. Three-dimensional studies are required where there is distinct layering or important gravitational influences. Two-dimensional cross-sectional simulation studies are frequently used to quantify the effects of gravity segregation. Results may then be incorporated in 2-D areal studies on horizontal waterfloods through the use of pseudo-functions. Grid Block Sizing. Numerical simulation involves a trade-off between calculation time and accuracy. The more grid blocks used to define a reservoir, the more accurate the calculated results. However, calculation time and cost also increase, and in many cases, prohibitively.
I
Large areal waterflood simulations should employ several grid blocks between wells so that pattern modifications and infill drilling may be studied.
Grid Block Orientation. Grid blocks should be oriented along permeability trends and geological layers. The orientation and size of the grid may affect the manner in which water break-through occurs. This problem is most pronounced where there are high contrasts in the water-oil mobility ratios. More advanced simulators use variational or nine-spot finite difference approximations to eliminate this effect. Timestep Sizing. Also related to grid block sizing is timestep sizing. Saturation fronts cannot pass through a grid block in one timestep. Thus, the smaller the gridblock, the shorter the timestep must be to ensure this condition is not violated. A smaller timestep means more timesteps, and hence a higher number of calculations to be performed ina simulation run. Most modem simulators utilize automatic timestep selection to optimize running times. History Match. Once a reservoir description is set up in the model, a simulation history match is run by entering actual oil production, water injection, or pressure constraint data. Calculated results such as reservoir pressure, water-oil ratios, gas-oil ratios, and oil rates are compared to actual results to judge the accuracy of the model. Model parameters are then revised and the model rerun to get a better match. This process is a trial-anderror procedure and relies on the judgement of the simulation engineer to revise the properties in an appropriate manner. The normal practice is to revise poorly defined properties first.
11.9.3
Reliability of Results
The numerical simulation technique is the most rigorous method of determining reservoir flow behaviour and compensates for most of the shortfalls experienced by analytical methods. Assuming that the model is set up appropriately to compensate for
166
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _A
ENHANCED RECOVERY BYWATERFLOODING
the numerical factors as described, the limitation ofthe simulation results rests solely on the accuracy of the reservoir description.
11.10
WATERFLOODING VARIATIONS
11.10.1
Naturally Fractured Reservoirs
When reservoir properties are established to a high degree of confidence by good well control, production results, pressure tests, and core studies, and the simulation history match requires very little alteration to the reservoir description, then a high degree of confidence may be placed on the results. As the number of alterations performed to achieve a match increases and the duration of the history-match period decreases, less confidence can be placed on the results. Confidence is a function of the reasonableness of the alterations made during the history-match process.
Displacement of oil by water in naturally fractured reservoirs is related to capillary and gravity forces acting on individual matrix blocks. "Imbibition" is the mechanism by which the nonwetting phase is displaced by the wetting phase in porous media due to the effects of capillary pressure. Oil recovery by imbibition is an important process in waterflooding of fractured reservoirs. As injected water advances along the fractures and is imbibed into the matrix, an equivalent volume of oil is released to the fractures. A discussion ofthe theory of waterflooding in fractured reservoirs has been presented by de Swaan (1978).
Example
The imbibition process is the dominant displacement mechanism when matrix blocks are small and capillary pressures are high. Gravitational pressure governs displacement when matrix blocks are tall and capillary pressures are low. For oil-wet rocks, external forces (gravity and applied pressure differentials) must overcome capillary pressures to recover oil from the matrix blocks. It follows then that matrix blocks must be of a certain height in order for waterflooding to be successful in oil-wet reservoirs. Knowledge of reservoir rock wettability is important in the evaluation of reserves in fractured reservoirs. For rocks that tend to be oil-wet, such as some dolomites, spontaneous imbibition of water will not occur. Fortunately, most dolomitic rocks have somewhat low capillary pressure due to large pore sizes, and gravity dominates the displacement process. Bridging between matrix blocks may result in a more continuous capillary network that will improve oil recovery.
This example demonstrates the kind of problem that simulation engineers may encounter in a history match in an immature waterflood. Water break-through has not occurred at a producing oil well when the simulator predicts that the well should be producing significant water. The engineer must decide which of the following applies: o
The relative oil-water permeability data is wrong.
o
There is directionalpermeability diverting water away from the producing well.
o
The pore volume between the injector and producer is too low.
o
There is a flow barrier between the producer and injector.
o
The reservoir model has been layered when no layering is in fact occurring.
The solution may be anyone of these. An incorrect alteration may still achieve a history match, but will result in incorrect forecasts. Simulations performed in proposed waterflood schemes will only have the primary depletion with which to history-match. While these should give a good determination of oil in place and oil and gas flow behaviour, they do not address directional permeability, reservoir layering, flow restrictions or actual relative oil-water permeability characteristics. In vertical schemes they will not address water coning characteristics, which are critical to flood performance. The assignment ofproved and probable reserves from a simulation study must address these limitations.
Oil recovery by imbibition is described by a timedependent transfer function that is determined in the laboratory and can be modelled mathematically as presented by Aronofsky et al. (1958). Laboratory tests on small reservoir samples are scaled according to rules presented by Mattax and Kyte (1962) to determine recovery from reservoir matrix blocks contacted by injected water. Displacement efficiencies within the fracture system, expected to approach 100 percent, are governed by gravity forces and applied pressure differentials since capillary pressures are negligible. Often, the original oil in place within the fracture system represents a small fraction of the total system. The areal and vertical sweep efficiencies in fractured reservoirs are expected to be similar to those in
167
DETERMINATION OFOILANDGASRESERVES
unfractured reservoirs provided that injection patterns in horizontal waterfloods properly account for permeability anisotropy. The water injection wells should line up parallel to the permeability trend. The influence of vertical fractures on areal sweep efficiency has been presented by Crawford and Collins (1954) and Dyes et al. (1953). It is noted that their studies were conducted for hydraulically fractured wells, but the results can be applied to naturally fractured reservoirs. In vertical waterflood schemes, the oil-water contact should be monitored to determine the in situ sweep efficiency. The measured oil-water contact represents the fluid contact within the fracture system and corresponds to the free water level since capillary pressures in the fractures are generally negligible. The recovery efficiency within the water-flushed portion ofthe reservoir will increase with time and can be history-matched using the model presented by Aronofsky et al (1958). The final recovery efficiency is obtained from the history-matched model. Waterflood reserves are determined by applying the final recovery efficiency to the estimated gross swept volume to account for sandwich losses at the top of the reservoir. The accuracy ofreserve estimates in naturally fractured reservoirs is dependent on accurate interpretation offracture characteristics such as frequency, width, orientation and distribution, reservoir structure, laboratory tests and reservoir monitoring. Proved and probable reserve estimates must consider the uncertainties and reasonable range of values associated with these parameters.
11.10.2 Polymer Flooding In reservoirs with unfavourable mobility ratios, watersoluble polymers may be added to improve displacement and sweep efficiencies. The mobility of the displacing phase is reduced due to an increase in fluid viscosity and an alteration in relative permeability related to polymer retention and modification of pore sizes. Reduction of water mobility has the disadvantage of reducing injectivity, so this limits the economic application of polymer flooding to high permeability reservoirs. Estimation ofreserves for polymer floods is similar to that for a waterflood but with a more viscous displacing phase. In the assignment ofproved and probable reserves, the polymer slug sizing and potential losses due to dissipation and polymer degradation must be considered.
11.10.3 Micellar Flooding Reduction of interfacial tensions is a major objective in micellar flooding processes. Surfactants (or surface-
168
active agents) are added along with polymers to injected water to reduce interfacial tension, and thereby reduce residual oil saturation and improve oil recovery. The reduction of residual oil saturation due to the addition of surfactants is determined in the laboratory. The technical success of a surfactant flood is dependent upon how much of the surfactant is lost due to adsorption, precipitation, the irnmobility ofthe surfactant-rich phase, and dissipation. The estimation of reserves for micellar floods is similar to that for a waterflood, but with a more viscous displacing phase. In the assignment of proved and probable reserves, the micellar-polymer slug sizing and potential surfactant losses must be considered.
11.11
STATISTICAL WATERFLOOD ANALYSIS SURVEY
11.11.1
Overview of Database
In order to illustrate typical usage of various reserve analysis methodologies and resulting recovery factors, reserve data on approximately 200 waterflood units in the western Canadian sedimentary basin were compiled from the database of an independent petroleum consulting firm. Recovery factor statistics are presented in Table 11.11-1, and methodology statistics are summarized in Table 11.11-2. The recovery factors presented are proved plus probable values derived by dividing ultimate recoverable reserves by original oil in place within the unit boundaries. Ultimate recoverable reserves are values calculated by the evaluators using the various assigrunent methodologies. Original oil-in-place values are based on estimates prepared by the operator, a governmental regulatory body, or an independent consultant. Depletion refers to cumulative oil production at the time of the evaluations divided by original oil in place. The ranges presented represent 7.5 percent of the sample points. Most of the sample points represent horizontal flood schemes. The data are from a sampling of waterfloods in western Canada, and the table is intended to show typical values and ranges of results.
11.11.2 Discussion of Results The following general observations may be made from a review of the data: I. The typical total recovery factor is 30 percent; range is 16 to 45 percent. 2. The average API gravity of samples is 33°; range is 23°t041°.
ENHANCED RECOVERY BY WATERFLOODING
Table 11.11-1 Summary of Recovery Factors: A Sampling of Western Canadian Waterfloods
Geologic Horizon
Litbology
No. of Data Points
Upper Cretaceous Lower Cretaceous Mannville Jurassic Triassic Triassic Permian Mississippian Mississippian Devonian Devonian
SS SS SS SS SS Carb SS SS Carb SS Carb
42 16 27 36 13 7 3 5 32 3 22
38 33 26 23 40 37 40 21 34 42 39
37-41 20-41 19-36 21-26 39-42 36-42 40-40 14-22 30-40 41-43 37-42
23 29 27 35 40 35 28 17 33 38 32
15-35 18-40 20-35 16-54 25-51 25-43 6-34 8-22 21-46 25-51 19-53
65 62 54 73 76 67 73 36 74 55 65
206
33
23-41
30
16-45
67
Average/Total
Table 11.11-2
Oil Gravity °API Range (75%) Average
Reserve Analysis Technique Distribution Depletion (%) No. of %of Data Points Samples Average Range
Volumetrics Decline Vol. & Decline Analogies Sim. Studies
51 110 27 12 6
Total
206
25 53 13 6 3
46 72 68 52 35
25-75 52-87 43-87 30-74 26-44
3. Assuming a recovery factor of 10 to 15 percent, the ultimate recovery under waterflood is typically at least double that of primary recovery.
Total Proved Plus Probable Recovery Factor Range Average (75%)
Aronofsky, J.S., Masse, L., and Natanson, S.O. 1958. "A Model for the Mechanism of Oil Recovery from the Porous Matrix Due to Water Invasion in Fractured Reservoirs." Trans., AIME, Vol. 213, pp.17-19. Caudle, B.H., and Witte, M.D. 1959. "Production Potential Charges During Sweepout in a FiveSpot Pattern." Trans., AIME, Vol. 216, pp. 446-448. Craig, F.F. 1971a. "The Reservoir Engineering Aspects of Waterflooding." SPE Monograph No. 3, pp. 29-44. ------.197Ib.pp.48-49. ------. 1971c. pp. 50-52.
4. Very few waterfloods exist for pools under200API gravity.
- - - -.. 1971d. pp. 108-111.
Decline analysis is generally not used until depletion is over 50 percent. The reason is the lack of definitive decline trends in immature stages ofwaterflood recovery. The data also suggest that waterflood declines start at approximately 50 percent depletion.
- - - -.. 1971f. pp. 41-43.
References Anderson, W.O. 1986. "Wettability Literature Survey." JPT, Oct. 1986, p. 1125.
Depletion (%)
- - - -.. 1971e. p. 64. - -.. 1971g. p. 93. Crawford, P.B., and Collins, R.E. 1954. "Estimated Effect of Vertical Fractures on Secondary Recovery." Trans., AIME, Vol. 201, pp. 192-196. Dardaganian, S.O. 1985. "The Application of the Buckley-Leverett Frontal Advance Theory to Petroleum Recovery." Trans., AIME, Vol. 213, pp. 365-368.
169
7
_.s.. .~
I I ,
DETERMINATION OF Oil AND GASRESERVES
de Swaan, A. 1978. "Theory of Waterflooding in Fractural Reservoirs." SPE Journal, Apr. 1978, pp.117-122. Dyes, A.B., Kemp, C.E., and Caudle, B.H. 1953. "Effect of Fractures on Sweep-out Pattern." Trans., AIME, Vol. 213, pp. 245-249. Energy Resources Conservation Board. 1993. PVT and Core Studies Index. Guide G- 14, Calgary, AB. Gould T.L, and Sarem, A.M.S. 1989. "Infill Drilling for Incremental Recovery." JPT, Mar. 1989, p. 229. Higgins, R.V., and Leighton, AJ. 1962. "A Computer Method to Calculate Two-Phase Flow in Any Irregularly Bounded Porous Medium." JPT, Jun. 1962,pp.679-683.
Kuo, M.C.T. 1989. "Correlations Rapidly Analyze Water Coning." OGJ, Oct. 1989, pp. 77-80. Mattax, C.C., and Kyte, lR. 1962. "Imbibition Oil Recovery from Fractured Water-Drive Reservoir." Trans., AIME, Vol. 201, pp. 192-196. Schoeppel, R.J. 1968. "Waterflood Prediction Methods - 7, Comparative Evaluation." O&GJ, Jul. 1968, p. 73. Slider, H.C. 1983a. Worldwide PracticalPetroleum ReservoirEngineeringMethods. Petroleum Publishing Company, Tulsa, OK, p. 551. - -. .l983b. p. 569. - - . 1983c.p. 557. - - . 1983d. p. 600. - - . 1983e. p. 554. Willhite, G.P. 1986. Waterflooding. SPE Textbook Series, Vol. 3, pp. 53-110.
170
-
1
Chapter 12
ENHANCED RECOVERY BY HYDROCARBON MISCIBLE FLOODING
12.1
INTRODUCTION
After discovery, most oil reservoirs produce under the natural energy of the reservoir. The primary drive mechanism for these reservoirs varies significantly. For pools with an initial reservoir pressure above the bubblepoint pressure, the energy is initially obtained by fluid expansion and rock compaction. Later in the life of the reservoir, when the pressure falls below the bubble-point pressure, additional energy will result from gas liberation and expansion. Usually these pools have recovery factors ofless than 20 percent. For pools with a bottom or edge water drive, oil is displaced by water, and the rate of decline of the reservoir pressure is reduced by the encroaching aquifer. The recovery factor for this type ofreservoir can be as high as 60 percent (e.g., Fenn Big Valley D2A Pool, Alberta, Canada). Reservoirs with a gas cap produce oil because of gas cap expansion. Other reservoirs may have both a gas cap and an aquifer. The recovery factor for these reservoirs can be as high as 80 percent (e.g., Westerose D-3 Pool, Alberta). The high recovery factor is due to rich gas sitting at the bottom of a thick gas pay zone because of gravity. This rich gas effectively acts as a solvent, and the result is a vertical miscible displacement of oil. In most reservoirs, oil recovery may be improved by the implementation of an enhanced oil recovery (EOR) scheme. EOR schemes can be classified as secondary and tertiary floods. Water injection for pressure maintenance, pattern waterflooding and immiscible gas injection are secondary EOR schemes. Hydrocarbon miscible floods and carbon dioxide miscible floods are tertiary EOR schemes. The terms "secondary" and "tertiary" indicate the EOR technology only, and not the state of the pool being flooded. Therefore, if a project is being miscible flooded before any waterflood, the miscible project is deemed a tertiary EOR project. In a waterflood or an immiscible gas flood, the displacing fluid is not soluble in the displaced oil. The displacement results in a residual oil saturation due to the interfacial forces between the displacing fluid and
the displaced oil. In contrast, miscible fluids are soluble in oil, so there will be no interfacial force between oil and solvent and the theoretical residual oil saturation will be zero. This chapter is limited to miscible flooding with hydrocarbon solvents. Miscible flooding is a proven technology that increases reserves. However, the improvement of the reserves estimation and the economic viability are affected by the following: • • • • • • • • • • •
Reservoir geology Rock properties Reservoir fluid properties Solvent composition and slug size Chase gas composition and slug size Implementation cost Well spacing and well patterns Flood types Oil, gas and condensate prices Royalty regime Stage of implementation
This chapter reviews recognized hydrocarbon miscible flood processes, the methods for the estimation of reserves, the accuracy of these methods, and the factors affecting this estimation as reported in the literature.
12.2
TYPES OF HYDROCARBON MISCIBLE FLOODS
Hydrocarbon miscible floods are the most common tertiary EOR schemes in western Canada. They can be subdivided into vertical or horizontal miscible floods.
12.2.1
Vertical Miscible Floods
Vertical miscible floods are usually implemented in pinnacle reefs or reservoirs with a high relief angle. In Alberta, the majority of these are in Rainbow Lake, Brazeau River, Pembina/West Pembina, and Wizard Lake. The solvent is injected as a blanket at the top of the reservoir to take advantage of a gravity-stabilized displacement. Subsequent chase gas injection drives solvent downward.
171
~ ....
..,
DETERMINATION OFOILANDGASRESERVES
The highest wells in the structure are usually chosen as the injectors to maximize oil displacement, and the producing wells are completed at the lowest porous interval above the oil-water contact. Production rates are controlled to restrict solvent and water production. Horizontal wells are becoming popular in vertical miscible floods. These wells are usually drilled as producers near the water-oil contact to reduce water and gas coning problems and thus increase the production rate and reduce the sandwich loss. Innovative completion techniques such as perforation below or at the oil-water contact have also resulted in reduced sandwich losses. The expected incremental recovery compared to upward waterflood is in the range of IS to 40 percent. The high incremental recovery factor in vertical displacement is due to high volumetric sweep efficiency as a result of a gravity stable displacement. The vertical miscible displacement is ideal in homogeneous reservoirs. In heterogeneous reservoirs with horizontal shale barriers, a substantial reduction in incremental oil recovery improvement can be expected as a result of poor vertical sweep efficiency (e.g., Golden Spike D-3 Pool, Alberta).
12.3
METHODS OF ACHIEVING MISCIBILITY
12.3.1
First-Contact Miscible Process
The simplest and most direct method of achieVing miscibility is to inject a solvent that is completely soluble in the oil in all proportions. Such solvents are called "first contact miscible" (FCM) and are the most expensive. As the ternary diagram shown in Figure 12.3-1 indicates, combining the oil and solvent in any proportion results in a single phase, i.e., no two-phase region is developed. Cost savings can be balanced against process risk by injecting less expensive "multi-contact miscible" (MCM) solvents which are subdivided into condensing and vapourizing processes. Intermediate
•
Solvent \
\
\ \ \ \ \ \
12.2.2
In horizontal miscible floods, solvent and water are injected alternately to mobilize residual oil and push it to the producers. After the injection of solvent and water, chase gas (which is miscible with solvent) and water are injected to extend the solvent bank size and complete the displacement process. After injection of 25 to 40 percent hydrocarbon pore volume (HPV) of solvent and chase gas, the process reverts to horizontal waterflood to depletion. In other words, in the early stage of the miscible project, oil is replaced by miscible solvent and moved toward producing wells. Later, residual solvent is mobilized by chase gas and moved to where it can contact more residual oil. Through this process, an expensive commodity, residual oil, is replaced with a cheaper commodity, chase gas. The majority of these floods are implemented in Swan Hills (Swan Hills A and B, South Swan Hills, Virginia Hills, Judy Creek A and B), Kaybob (Kaybob BHLA and Kaybob South Triassic), Goose River and FennBig Valley areas of Alberta. The expected incremental recovery factor of 5 to IS percent results from gravity override, viscous fingering, and the inability to control injection profiles.
172
\
Horizontal Miscible Floods
\
..
Reservoir Oil
Light
Heavy
Figure 12.3-1 Pseudo-Ternary Diagram Indicating First-Contact Miscibility
12.3.2
Multiple-Contact Miscible Process
In a condensing process, the intermediate hydrocarbons from the injected solvent condense into the reservoir oil to create a mixing zone. Initially, a given volume of solvent contacts the reservoir oil, resulting in a mixture, M I, which separates into an equilibrium gas, 01, and liquid, Ll (Figure 12.3-2). Further injection of solvent pushes the more mobile equilibrium gas ahead of the liquid, and the solvent contacts liquid, L I, resulting in a mixture, M2. The mixture again separates into equilibrium gas and liquid phases (G2 and L2, respectively). This process is repeated and, after a series of chain flashes, results in the formation of a two-phase envelope on the ternary diagram. The composition of the
I
I·
ENHANCED RECOVERY BY HYDROCARBON MISCIBLE FLOODING
equilibrium liquid travels up the bubble-point curve, becoming richer in the components of intermediate molecular weight as they condense out of the solvent and into the oil. However, as the equilibrium liquid becomes richer, the amount ofthe intermediate components lost from the solvent into the oil at each contact becomes less, and the vapour flashed at each contact and pushed ahead into the reservoir also becomes richer. Intermediate
the reservoir pushes the equilibrium gas, G I, further into the reservoir. This gas contacts fresh reservoir oil resulting in a mixture, M2, which separates into an equilibrium gas, G2, and a liquid, L2. Further injection causes gas, G2, to flow ahead and contact fresh reservoir. In this process, the composition of the gas at the displacing front is getting richer and progressively moving along the dew point until it reaches the composition that is directly miscible with the reservoir oil. Intermediate
Reservoir Oil
G~2g~~ M2
Reservoir Oil
L1
Light
Heavy
Solvent
Light
Figure 12.3-2 Development of Multiple-Contact Miscibility Condensing Process
This in situ multiple contact generation of miscibility establishes a "transition zone of contiguously miscible fluid compositions from the reservoir oil composition through compositions Ll, L2, L3, ... Ln on the bubblepoint curve to the injected gas composition." That is, the solvent is in first-contact miscible with the equilibrium liquid LM-I, which is in first-contact miscible with equilibrium liquid LM-2, and so on. This process dominates the leading edge of the multiple transition zone.
12.3.3
Vapourizing Multiple-Contact Miscibility
In a vapourizing process, the intermediate weight hydrocarbons from the reservoir oil vapourize into the injected solvent to create a mixing zone. In this process, miscibility can be achieved with natural gas, flue gas, carbon dioxide or nitrogen, provided that the reservoir pressure is above the minimum miscibility pressure. The development of miscibility in a vapourizing solvent drive can be explained with the help of the ternary diagram in Figure 12.3-3. Initially, a given volume of solvent contacts the reservoir oil, resulting in a mixture, MI, which separates into an equilibrium gas, Gl, and liquid, LI. Subsequent injection of solvent into
Heavy
Figure 12.3-3 Development of Multiple-Contact Miscibility Vapourizing Process
If the reservoir pressure is close to the bubble-point pressure, small pockets of gas may be formed at the structurally high area of the reservoir. These pockets of gas may dilute the equilibrium gas to the extent that the miscibility is lost.
12.4
EXPERIMENTAL METHODS TO DETERMINE MISCIBILITY
Four methods have been widely used by the industry to determine miscibility and design the composition of solvent and chase gas: 1. 2. 3. 4.
The pressure composition diagram (P-X) The multi-contact ternary diagram The slim tube test The rising bubble apparatus (REA)
12.4.1
poX Diagram
A typical poX diagram is perfomed as a screening test by combining reservoir fluid with increasing mole fractions of injection solvent, and measuring the saturation pressure of each mixture. The cricondenbar, critical point, and solubility limit can be determined
173
DETERMINATION OFOILANDGASRESERVES
at the operating temperature (Figure 10.2-1). The hydrocarbon mixture is deemed acceptable for injection ifthe cricondenbar lies below the reservoir operating pressure. This defines an FCM solvent.
12.4.2
Multi-Contact Ternary Diagram
The test is performed at reservoir pressure and temperature by combining the reservoir fluid with solvent. The compositions of the resultant equilibrium vapour and liquid are determined and become the first points on the phase envelope. The next step depends upon which multiple-contact miscibility (MCM) process is being simulated. For a condensing MCM process, the equilibrium gas is discarded and more solvent is added to the equilibrium liquid. For a vapourizing MCM process, the equilibrium liquid is discarded and more oil is added to the equilibrium gas. The procedure is repeated several times; tie-lines are defined after each step, and the appropriate phase envelope is generated. The hydrocarbon mixture is deemed immiscible if the solvent lies on the extension of a tie-line.
12.4.3
Slim Tube Test
The slim tube test apparatus consists of a long (usually more than 20 m) coiled stainless steel tube packed with glass beads or crushed silica. The porous medium is initially saturated with reservoir oil at the desired test temperature and pressure. Solvent is injected at one end, and miscibility is determined through visual observation ofthe transition zone, the recovery factor ofthe oil and the break-through performance ofkey solvent components (e.g., CI, C2, C3). Unlike the ternary and PoX diagrams, which are conducted at static conditions, slim tube tests represent a dynamic process where the degree of dispersion in the reservoir is to some extent reproduced in the lab.
12.4.4
Rising Bubble Apparatus
The rising bubble apparatus (RBA) consists of a small-diameter vertical tube mounted in a high-pressure cell. A bubble of solvent is injected at the bottom ofthe tube. The miscibility characteristic is determined by visual observation of the bubble decay as it rises through the reservoir oil. The rising bubble apparatus (RBA) combines the small size and compactness of the visual cell with the dynamic nature ofthe slim tube test. Hence, this method can make miscibility determination much more efficient than the other three methods.
12.5
SCREENING AND FEASIBILITY STUDIES
Screening, design and implementation ofa hydrocarbon miscible project usually involve the following steps: I. Estimate the incremental oil reserves based on the volumetric method. 2. Make a preliminary production forecast based on the break-through ratio (BTR) concept (Section 12.5.2) and preliminary economic evaluation. 3. Use a detailed geological model to evaluate the reservoir characteristics; from this model, provide input data for simulation study. 4. History-match the pool performance with a black oil simulation model under primary and secondary drive mechanisms. This model will provide the saturation and pressure distributions required for a subsequent pseudo-miscible or compositional study. 5. Use a pseudo-miscible or compositional model to generate a production forecast, evaluate the reservoir's performance under miscible flood, and design the project. 6. Use experimental and numerical model studies to determine the optimum slug size and design the composition of the injection fluid. 7. Make an economic evaluation and feasibility design. 8. Obtain regulatory approvals. 9. Design facilities and implement. 10. Develop the data acquisition system, and prepare detailed monitoring programs and detailed field operation guidelines. II. Monitor performance and reservoir management to improve the pool performance under the miscible flood. Prior to implementation ofa hydrocarbon miscible flood, the project goes through several stages. At the early stage of the screening, usually a crude method is used to generate a production forecast and conduct an economic evaluation. At this stage, the incremental reserves may be estimated by the volumetric method, and the production forecasts may be generated by the BTR technique. For the feasibility study, it is imperative to conduct a detailed geological and reservoir simulation study to evaluate the economic viability of the project. Other methods such as statistical models have also been used to evaluate the feasibility of a hydrocarbon miscible flood.
174
______________________cn
ENHANCED RECOVERY BY HYDROCARBON MISCIBLE FLOODING
12.5.1
Volumetric Method
The volumetric equation for estimating incremental oil reserves, RE, is: RE=EHxEvxEDxOOIP where
EH Ev ED OOIP
(1)
aerial sweep efficiency vertical sweep displacement efficiency original oil in place
The reserves target for a miscible flood is the residual oil saturation after waterflooding. Therefore, the displacement efficiency, ED, is defined as: ED= (Socw - So,,)/ (I-S w)
(2)
where Socw= residual oil saturation after waterflood (fraction) residual oil saturation after miscible flood (fraction) Sw = connate water saturation The connate water saturation is the water saturation in the reservoir at discovery, which can be determined from resistivity logs.
Estimation of Residual Oil Saturation The residual oil saturation is the amount of oil left behind in a water-swept zone when the relative permeability approaches zero. The residual oil saturation is a function of wettability, adhesion, and rock properties. Four methods are used to determine residual oil saturation: core flood test, pressure coring, logging, and the single-well tracer method. The core flood test is discussed in Chapter II. Pressure coring is considered to be an accurate method for obtaining a volumetric measure ofremaining oil saturation. However, this method is expensive and requires that a new well be drilled in a waterflooded part of the reservoir. Logging techniques that can be used for obtaining a volumetric measure of remaining oil saturation are log inject log (pulsed neutron, gamma radiation, resistivity), and carbon-oxygen logging. Each method has its own special advantages and limitations. The resistivity logs can be run only in open holes. The single-well tracer method (Deans and Majoros, 1980) measures an average remaining oil saturation that is weighted according to the product of thickness and effective permeability to brine at remaining oil saturation (capacity) for the various strata sampled by injected tracer. In carbonate reservoirs, due to the effect of the
dead end pore volume, this method requires a careful examination of the data obtained and a comprehensive simulation study. For a given formation and interval, the remaining oil saturation obtained from these methods may be different because, besides the question of accuracy, the depth of investigation and vertical resolutions of the various methods are usually different. For example, the single well tracer technique represents a capacity-weighted average while pressure coring or logging gives a volumetric-weighted average. The comparison ofresults provides important measures of the quantity and distribution of remaining oil. For example, if the former is significantly lower than the latter, the interval may be highly stratified or contain dead end pore volume. In all cases, the measured remaining oil saturation will exceed the residual oil saturation.
Estimation of Residual Oil Saturation after Miscible Flooding Theoretically speaking, the residual oil saturation after miscible flooding should be zero due to the lack of interfacial tension between oil and solvent. However, even ifthe miscibility criteria are met, all the oil may not be displaced due to trapping by mobile water or dead end pores. The average oil saturation left behind after hydrocarbon flooding is usually greater than that estimated from the core flood studies. The usual expectation in western Canada carbonates is 5 percent HPV.
Estimation of Areal Sweep Efficiency for Horizontal Miscible Floods Areal sweep efficiency is the fraction ofthe pattern area that has been contacted by solvent and mainly depends upon the mobility ratio of the displacement process in that the areal sweep efficiency decreases as the mobility ratio increases or become more unfavourable. The mobility ratio, M, between an oil bank and the solvent displacing the oil bank in the presence of mobile water is defined by Stalkup (1983) as:
M
w) k, +k' (J..ts u, sW avg
= -'----'--------""'-
k,
( M: +
(3)
k w) J..Lw
oW avg
where k, = effective permeability to solvent (mD) k; = effective permeability to water (mD) k, = effective permeability to oil (mD)
175
~DETERMINATION OF OIL AND GASRESERVES
IJ" = solvent viscosity (m Pa.s) Ilw water viscosity (m Pa.s) Jlo oil viscosity (m Pa.s) sw = solvent/water ow = oil/water Since solvent-oil mixtures have much lower viscosities than oil, the solvent-oil mixing zone becomes less stable than for waterflood, and numerous fingers of solvent may develop and extend toward producing wells. This is one explanation for early solvent break-through and poor sweep efficiency. As can be seen from Equation (3), the concept of injecting water alternately with the miscible fluid improves the overall mobility of the displacement and thus improves the areal sweep efficiency. Areal sweep efficiency in a miscible flood is also a function of areal heterogeneity, geometry of pattern flood, dispersion/diffusion, pore volume of solvent injected, and water alternating gas ratio (WAG). Craig (1971) extensively reviews lab measurements for displacement with a favourable mobility ratio where the displacement front is stable and the effect of viscous fingering is insignificant. For an unfavourable mobility ratio, Habermann (1965), Mahaffey et al. (1966), Dyes et al. (1954), and Kimbler et al. (1969) measured areal sweep efficiency of a homogeneous five-spot pattern for a single-front displacement, where solvent was injected continuously and initially oil was the only mobile fluid. The data indicated that areal sweep efficiency at solvent break-through decreases with increasing mobility ratio. Claridge (1973) developed a correlation for areal sweep efficiency by using Dyes' data and applying Koval's (1963) equations for linear displacement efficiency of an unstable displacement. This correlation applies to a confined five-spot pattern -in a homogenous, single-layer reservoir where the gravity force is negligible compared to the viscous forces and in the absence of movable water or gas. Because of these assumptions, the Claridge method can be used only for agross estimation of areal sweep efficiency. Estimation of Vertical Sweep Efficiency for Horizontal Miscible Floods
In a horizontal miscible displacement, where the density of the solvent is much less than the density of either oil or water, vertical sweep efficiency is substantially reduced as a result of gravity segregation where solvent overrides oil. Parameters that affect vertical sweep efficiency are reservoir stratification, vertical distribution of flow capacity and segregation of
hydrocarbon phases. Obviously, a homogenous reservoir with low ratio of horizontal to vertical permeability has a higher chance of incremental recovery reserves losses. Conversely, diffusion and convective dispersion may allow solvent to liberate the remaining oil from zones where water would not enter. In a WAG process, gravity segregation will cause the injected gas to rise to the top of the formation and water to settle to the bottom. This will result in a low recovery factor since only a thin layer at the top of formation is solvent flooded while the bottom layer is waterflooded. Stone (1982) showed that recovery is primarily a function of the viscous-gravity ratio, VGR, defined as: VGR=
q,
(4)
dp k, a (k,. + k,,) Jl, Jl. where q,
dp k, A ~
Jlw
k,., Jlg
= total flow rate = density difference between water and solvent vertical permeability (mD) reservoir area = relative permeability to water water viscosity (m Pa.s) relative permeability to solvent = solvent viscosity (m Pa.s)
The recovery factor is also a function of water-gas ratio. For the same solvent slug size, higher values of WAG may result in higher recoveries. Based on the Stone (1982) model, Jenkins (1984) presented a solution for estimating the vertical sweep efficiency for a horizontal displacement in a homogenous reservoir with either rectangular or radial geometries. This model will provide only a rough estimate of the vertical recovery due to the limitation in ignoring capillary pressures, nonuniform saturation distribution, and physical dispersion. Hence, this model is recommended only for the screening study. Okazawa et al. (1992) used the Claridge correlation for the estimation ofareal sweep efficiency, and the StoneJenkins model for the estimation of vertical sweep efficiency to predict the performance of large-scale miscible flood. The Okazawa model may be used for the screening study or performance monitoring of a largescale miscible flood. However, for a feasibility study, a detailed geological study and a simulation study are recommended. The volumetric method can also be used to estimate the incremental reserves from vertical miscible floods. In
176
______________________a
ENHANCED RECOVERY BY HYDROCARBON MISCIBLE FLOODING
this estimation it is important to calculate with coning correlations the sandwich loss due to water and gas coning. Field Estimation of Volumetric Sweep Efficiency
Many experimental and mathematical studies of volumetric sweep efficiency have been presented in the literature. However, little attention has been paid to the field evaluation of volumetric sweep efficiency. Asgarpour and Todd (1988) used a radioactive tracer program along with the simulation study to estimate the volumetric sweep efficiency for an ongoing miscible flood in central Alberta. In the simulation study, the historical performance of the primary natural water drive and the first five years of solvent injection were reproduced with a fair degree of accuracy by a pseudomiscible model. This information, along with the results ofthe radio-active tracer program, was used to estimate the volumetric sweep efficiency of 45 percent for this flood. This study concluded that the effect of gravity override and viscous fingering were much more moderate than had been expected from the lab models.
12.5.2
Break-Through Ratio Method
The BTR method can be used to generate a production forecast for the preliminary evaluation of hydrocarbon miscible floods. The incremental hydrocarbon miscible flood estimated from the volumetric method, the waterflood reserves obtained from extrapolation ofWOR vs. cumulative oil, and the economically limiting BTR are used to determine the ultimate point on the BTR curve. The BTR curve is then constructed based on waterflood performance and the performance of a similar pool under hydrocarbon miscible floods. Finally, the production forecast is generated using a trial-and-error procedure (i.e., an oil rate is assumed-the total production rate is obtained from the BTR curve and this rate is compared with the injection rate for appropriate voidage replacement). The BTR method can also be used to estimate the incremental reserves and monitor the performance of ongoing miscible floods (Asgarpour et al., 1988). The incremental reserves are estimated by the extrapolation ofthe BTR curve to the economically limiting BTR provided that the pool is in a mature stage (i.e., BTR >5).
12.5.3
The break-through ratio, BTR, is defined as the water production plus free gas production at reservoir conditions divided by the oil production at stock tank conditions. BTR = (GOR - R,) x Bg + (WORx Bw ) where GOR = gas-oil ratio R, = solution gas-oil ratio B g = gas formation volume factor WOR = water-oil ratio B; = water formation volume factor
ultimate recovery is the incremental oil due to the miscible flood.
(5)
The BTR vs. cumulative oil production plotted on a semi-log graph for most waterfloods is a straight line terminating at the ultimate recovery and the economically limiting BTR. The upward trend of the BTR line for a waterflood is due to the increase in water production. For a miscible flood that is implemented after a waterflood, the BTR curve is expected initially to have a downward trend as a result of a steady decline in water production accompanied by an increase in oil production. This downward trend will be followed by an upward trend primarily due to solvent and chase gas break-through and later by an increase in water production. The upward trend of the BTR will terminate at the ultimate recovery and the economically limiting BTR. The difference between the waterflood recovery and the
Geological Model
For the geological study, structural and stratigraphic cross sections are constructed to evaluate the effect of stratification for the horizontal miscible flood or the impact of shale barrier for the vertical miscible flood. The determination of vertical and horizontal permeabilities and the averaging method is also essential for this evaluation. From this model, input data is provided for the simulation studies.
12.5.4
Simulation StUdies
A black oil simulation study is conducted to reconcile the geological and reservoir data by 'history-matching the pool performance under primary and secondary drive mechanisms. This model provides saturation distribution, pressure, etc. required for a subsequent pseudo-miscible or compositional study. Pseudomiscible or compositional models are used to generate production forecasts and predict reservoir performance under several miscible flood options. Project design is then based on the optimum case. The compositional simulator is capable of evaluating and predicting changes in compositions and pressures of the hydrocarbon phases. Since the model is developed to simulate flow in three dimensions, it considers cross-flow between layers, gravity segregation, channelling, and the effect of variable mobility.
177
DETERMINATION OF OIL AND GASRESERVES
Simulation studies using the compositional simulator can be used to estimate slug size requirements and the effects of mass transfer between phases on miscibility conditions at the leading and trailing edges of the mixing zone. Lack ofphysical dispersion or mixing parameters in the miscible flow calculation is a major drawback. However, this problem can be overcome by adjusting the numerical dispersion to reflect the physical dispersion. The major problem with this model to date has been the cost, which makes a field study impractical. The advent of a high-capacity PC version is reducing this problem. There are two types ofcompositional model: one based on k value correlations, and the other on the equation of state. The first is more cost-effectivebecause ofthe lower computational cost for the flash calculations. Prior to a simulation study, phase behaviour should be studied so that the fluid can be properly characterized and the equation of state can be "tuned" to experimental data. The pseudo-miscible model was developed (Todd and Longstaff, 1972) by modifying an existing three-phase simulator to forecast miscible flood performance. The simulator is capable of modeling the essential features of miscible displacment by a fairly coarse numerical grid. The degree of dispersion rate between solvent and oil, which reflects the degree of viscous fingering, is represented by an input mixing parameter. For most field applications mixing parameters in the range of0.5 to 0.8 are used. For a system exhibiting strong gravity segregation, a large number of grid blocks is required to represent different layers. Production forecasts from the simulation study are used to conduct detailed economic evaluations, make project decisions and obtain management, partner and regulatory approval.
12.5.5
Estimation of Uncertainties
The estimation ofincremental reserves for a hydrocarbon miscible flood depends on several parameters: porosity, pay thickness, areal extent, residual oil saturation to waterflood, connate water saturation, formation volume factor, and residual oil saturation to miscible, areal and vertical sweep efficiency. For most ofthese parameters, only a range may be available. In reserves estimation, the uncertainty associated with these parameters should be taken into account. To properly describe the risk and uncertainty associated with the incremental reserves of a proposed hydrocarbon miscible flood in central Alberta, Asgarpour and Papst (1990) developed a statistical model based on the Monte Carlo simulation technique. The input parameters to this model are the
distribution ofthe parameters required for the reserves estimation and the output is the reserves distribution. Figure 12.5-1 shows a reserves distribution for a proposed miscible flood in central Alberta. Based on this distribution, a different reserves category can be assigned. For example, this figure indicates proved reserves of 1.375 million cubic metres with 80 percent confidence.
100 80
~ 60 ~
:0 m .c
e a.
40
----- - - --
t
20
o -I---.-,r-'-r----,--.---,--;:=;---.-o 1.0 2.0 3.0 4.0 Incremental Reserves (106m3)
Figure 12.5-1 Reserves Distribution
12.5.6
Determination of Solvent and Chase Gas Slug Size
In a miscible flood, the total amount of solvent used should be enough to maintain miscibility conditions at the displacement front in the bulk of the reservoir. Whereas heterogeneity and stratification have a net effect ofincreasing the solvent losses and consequently the solvent requirements, economic considerations dictate optimizing these amounts. The principal parameters determining solvent losses are the dispersion and mixing coefficient. Unfortunately, no simple methods for determining these coefficients are available. The complexity of the flooding process makes the interpretation of data from any of the available methods extremely difficult. These complexities, besides heterogeneities and stratification, could be a result ofthe front displacement, the geometry of flood propagation (Asgarpour et aI., 1989), dead end pores (Asgarpour, 1987), the nature of the miscible flood (Chen et aI., 1986), the presence of mobile water or trapped gas (Asgarpour et aI., 1986, Tiffin, 1982), and wettability. Furthermore, the dispersion observed in a single core sample is different from what is observed a few metres around a wellbore. These dispersions, in tum, could be different from those observed over the inter-well distance in the reservoir.
178
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _.--0
ENHANCED RECOVERY BY HYDROCARBON MISCIBLE FLOODING
The most common method to determine solvent and chase gas slug size is on the basis of dispersiondiffusion calculations at the leading and the trailing edges ofthe solvent-oil mixing zone (Asgarpour, 1987).
mature miscible floods where a large portion of solvent has been injected, horizontal injectors can be drilled to access the unswept layers and improve the volumetric sweep.
12.5.7
In horizontal miscible floods, producers should also be equipped with mechanical packer assemblies to control solvent cycling. In the latter stages of miscible floods, horizontal producers may be used to improve production from layers with low productivity.
Field Performance of Miscible Floods
The performance of hydrocarbon miscible floods has been extensively reviewed in the literature (Griffith and Home, 1975), (Beeler, 1977), (Reinhold et aI., 1992), (Anderson et al., 1992), (Pritchard and Nieman, 1992), (Patel and Broomhall, 1992), (McIntyre et aI., 1991), (Adamache et aI., 1990), (Fong et aI., 1990), (Wood et aI., 1990), (Bennett and Geoghegan, 1990), (Okazawa and Lai, 1989), (Dawson et aI., 1989), (Bilozer and Frydl, 1989), (Sorenson and Griffith, 1988). In general, vertical miscible floods have been more successful than horizontal floods due to the gravity stable displacement. However, in some vertical hydrocarbon miscible floods, poor geological understanding of the reservoir has resulted in lower than expected reserves due to the presence of shale barriers which resulted in poor volumetric sweep efficiency. Sandwich loss due to gas and water coning has also been a major factor in reducing the incremental reserves and economics. Recently, in a few hydrocarbon miscible floods, through innovative completion techniques, the volumetric sweep efficiency has been increased significantly by the reduction of sandwich loss. In vertical miscible floods, horizontal wells can be used as both producers and injectors. Horizontal producers drilled at the water-oil interface can reduce coning problems and improve the volumetric sweep efficiency. Horizontal injectors drilled in vertical miscible floods can provide a stable solvent transition zone, prevent the pure solvent from fingering into oil, and improve the volumetric sweep efficiency. The major problems associated with the horizontal miscible floods are gravity override and viscous fingering. In addition, poor injection profile control has resulted in a low vertical sweep efficiency in many floods. In these, the layer with the highest capacity takes the bulk of the solvent. Only a small portion of the reservoir can be contacted, and solvent is cycled through this layer without improving the incremental reserves from the other layers. Therefore in the early stages of floods, equipment of injectors with mechanical packer assemblies is important to control the amount of solvent injected per layer so each layer receives enough solvent to meet the miscibility criteria. However, for
Poor vertical sweep efficiency caused by gravity override can also be partially improved by increasing the solvent and water injection volumes. The miscibility process also plays a significant role in the success and failure of the hydrocarbon miscible process. Usually, the first-contact and condensing multiple-contact miscible processes have proved to be more successful than the vapourizing process. The poor success of the latter may be due to the presence of pockets of gas in reservoirs with pressures near the bubble-point pressure. This gas may dilute the solvent to the extent that it is no longer miscible with oil.
12.6
CLASSIFICATION OF MISCIBLE HYDROCARBON RESERVES
For the purpose of estimating reserves, it is important to identify the development stage of the project. The stage will signify the degree of confidence in recovery of the reserves and can be linked, therefore, to the timing of allocation of reserves to various categories.
12.6.1
Possible Reserves
It is suggested that prior to classifying any miscible reserves in the possible category, a reservoir engineering study should be completed to identify the reserves. An economic evaluation should also be conducted based on the present, or on a reasonable anticipated, economic condition. A confidence level of 10 to 40 percent probability of the incremental EOR reserves being recovered is required to allocate reserves to the possible category.
Caution should be used in estimating possible reserves based only on analogy to similar pools under miscible flood. The feasibility of a miscible scheme is dependent upon numerous complex parameters, and simple analogy is usually misleading. Miscible flood reserves in the possible category serve for easy identification of EOR potential for business planning. Once the business opportunity is identified, it can set in motion an action plan for implementation of the scheme.
179
DETERMINATION OF OILAND GAS RESERVES
12.6.2
Probable Reserves
When the project is in the implementation phase, tertiary EaR incremental reserves can be allocated to the probable category provided there is sufficient confidence (40 to 80 percent) that these reserves are expected to be recovered.
12.6.3
Proved Reserves
The amount of incremental tertiary EaR reserves allocated to the proved category should be based on the performance of the miscible flood. In the year the scheme is started, a small percentage ofthe incremental reserves could be added to the proved category based on the confidence level (Mukherjee, 1988). Additional hydrocarbon miscible reserves may be allocated to the proved category in a gradual manner over a period of time provided there is sufficient technical confidence in the scheme that the proved reserves figure has a high probability (80 percent) of being recovered.
References Adamache, I., McIntyre, F.J., Pow, M., Lewis, D., Davis, R., Kuhme, A., Bloy, G., Van Regan, N., and Butler, S. 1990. "Horizontal Well Application in a Vertical Miscible Flood." Petroleum Society ofCIMISPE International Technical Meeting, Calgary, AB; Preprints V3, CIM/SPE 90-125, Jun. 1990. Anderson, lH., Laurie, R.A., Loder, W.R. Jr., and Kennedy, P. 1992. "Brassey Field Miscible Flood Management Program Features Innovative Tracer Injection." Paper presented at 67th Annual SPE Technical Conference, Washington, DC, SPE24874, Oct. 1992. Asgarpour, S.S. 1987. "Determination of Slug Size for Carbonate Reservoirs." Paper presented at 38th Annual Technical Meeting of the Petroleum Society of CIM, Calgary; AB, Paper No. 38-09. Asgarpour, S.S., Pope, J.A., and Springer, s.i 1986. "Effect of Mobile Water Saturation on Slug Size Determination." Paper presented at 37th Annual Technical Meeting, Petroleum Society of CIM, Calgary, AB, Paper No. 86-36-35. Asgarpour, S.S., Crawly, A., and Springer, S.J. 1988. "Re-Evaluation of Solvent Requirements for a Hydrocarbon Miscible Flood." SPE Reservoir Engineer, Feb. 1988.
Asgarpour, S.S., and Todd, M.R. 1988. "Evaluation of Volumetric Conformance for the Fenn-Big Valley Horizontal Hydrocarbon Miscible Flood." Proc., 63rd Annual Technical Conference and Exhibition, SPE of AIME, Houston, TX. Asgarpour, S.S., Card, C., Singhal, AX., and Wong, T. 1989. "Performance Evaluation and Reservoir Management of a Tertiary Miscible Flood in the Fenn-Big Valley South Lobe D-2 Pool." JCPT, Nov-Dec, 1989, p. 6. Asgarpour, S.S., and Papst, W. 1990. "A Statistical Model to Evaluate a Hydrocarbon Miscible Flood in an Upper Devonian Field in Central Alberta." JCPT, May-Jun. 1990, p. 61. Beeler, P.F. 1977. "West Virginia CO 2 Oil Recovery Project Interim Report." Proc., US DOE Symposium on Enhanced Oil and Gas Recovery and Improved Drilling Methods, Tulsa, OK, Aug. - Sep. 1977. Bennett, F., and Geoghegan, J.G. 1990. "Monitoring the Performance of Pembina Nisku Miscible Floods." Paper presented at Petroleum Society of CIMISPE International Technical Meeting, Calgary, AB; Preprints V2, CIM/SPE 90-73, Jun. 1990. Bilozer, D.E., and Frydl, P.M. 1989. "Reservoir Description and Performance Analysis of a Mature Miscible Flood in Rainbow Field, Canada." Paper presented at 64th Annual SPE Technical Conference, San Antonio, TX; Proc., G-EOR/General Petroleum Engineering, SPE19656, Oct. 1989. Chen, S.M., Olynyk, L, and Asgarpour, S.S. 1986. "Effect of Multiple-Contact Miscibility on Slug Size Determination." JCPT, May-Jun. 1986. Claridge,.E.L. 1973. "A Trapping Hele-Shaw Model for Miscible-Immiscible Flooding Studies." SPEJ, Oct. 1973, Vol. 1339, pp. 255-261. Craig, F.F. 1971. The Reservoir Engineering Aspects of Waterflooding. SPE Monograph Series, Dallas, TX,No.3. Dawson, A.G., Buskirk, D.L., and Jackson, D.D. 1989. "Impact of Solvent Injection Strategy And Reservoir Description on Hydrocarbon Miscible EaR for the Prudhoe Bay Unit, Alaska." Paper presented at 64th Annual SPE Technical Conference, San Antonio, TX; Proc., G-EOR/ General Petroleum Engineering, SPE-19657, Oct. 1989.
180
2
-I ENHANCED RECOVERY BY HYDROCARBON MISCIBLE FLOODING
Deans, H.A. and Majoros, S. 1980. "The Single-Well Chemical Tracer Method for Measuring Residual Oil Saturation." Final Report for US DOE, Contract No. DE-AS 19-79BC20006 performed at Rice University, Houston, TX, Oct. 1980. Dyes, A.B., Caudle, RH., and Erikson, R.A. 1954. "Oil Production After Breakthrough - As Influenced by Mobility Ratio." Trans., AIME. Vol. 201, pp. 81-86. Fong, D.K., Wong, F.Y., Nagel, R.G., and Peggs, J.K. 1990. "Combining A Volumetric Model with a Pseudo-Miscible Field Simulation to Achieve Uniform Fluid Levelling in the Rainbow Keg River "B" Pool." Petroleum Society ofCIMISPE International Technical Meeting, Calgary, AB, Jun. 1990. Griffith, J.D., and Horne, A.L. 1975. "South Swan Hills Solvent Flood." Proc, 9th World Petroleum Congress, Tokyo, Japan, Vol. 4,1975. Habermann, R 1965. "The Efficiencies of Miscible Displacement as a Function of Mobility Ratio." Trans., AIME, Vol. 219, p. 264; Miscible Processes, Reprint Series, SPE, Dallas, TX, 1965, Vol. 8, pp. 205-214. Jenkins, M.K. 1984. "An Analytical Model for Water/Gas Miscible Displacements." Presented at 4th Symposium on EOR, Tulsa, OK, Apr. 1984, SPE/DOE 12632. Kimbler, O.K., Caudle, RH., and Cooper, H.E. Jr. 1969. "Areal Sweep-out Behaviour in a NineSpot Injection Pattern." JPT, Feb. 1969, pp. 199-202; Trans., AIME, Vol. 231. Koval, EJ. 1963. "A Method for Predicting the Performance of Unstable Miscible Displacement in Heterogeneous Media." SPEJ, Jun. 1963, pp. 145-154; Trans., AIME, Vol. 228. Mahaffey, J.L., Rutherford, W.M., and Matthews, C.W. 1966. "Sweep Efficiency by Miscible Displacement in a Five-Spot." SPEJ, Mar. 1966, pp. 73-80; Trans., AIME, Vol. 237. Mcintyre, FJ., See, D.L., Mallimes, R.M., Burger, D.H., and Tsang, P.W. 1991. "Production Optimization of a Horizontal Well in a Vertical Hydrocarbon Miscible Flood Reservoir." Petroleum Society ofCIMIAOSTRA Technical Conference, Banff, AB; Preprints V2, No. 91-68, 1991.
Okazawa, T., and Lai, F.S.Y. 1989. "Volumetric Balance Method - To Monitor Field Performance of Gas Miscible Floods." 40th Annual Petroleum Society of CIM Technical Meeting, Banff, AB; Preprints VI, No. 89-04-4, May 1989. Okazawa, T., Bozac, P.G., Seto, A.C., and Howe, G.R. 1992. "An Analytical Software for PoolWide Performance Prediction ofEOR Processes." Paper presented at 43rd Annual Technical Meeting of the Petroleum Society ofCIM, Calgary, AB, CIM 92-89. Patel, R.S., and Broomhall, R.W. 1992. "Use of Horizontal Wells in Vertical' Miscible Floods, Pembina, Nisku, Alberta, Canada." 8th SPEIDOE Enhanced Oil Recovery Symposium, Tulsa, OK; Proc., VI, 1992, SPE/DOE-24124, Apr. 1992. Pritchard, D.W.L., and Nieman, R.E. 1992. "Improving Oil Recovery through WAG (WaterAlternating-Gas) Cycle Optimization in a Gravity-Override-Dominated Miscible Flood." 8th SPEIDOE Enhanced Oil Recovery Symposium, Tulsa, OK; Proc., V2, SPE/DOE24181, Apr. 1992. Reinbold, E.W., Bokhari, S.W., Enger, S.R., Ma, T.D., and Renke, S.M. 1992. "Early Performance and Evaluation of the Kuparuk Hydrocarbon Miscible Flood." 67th Annual SPE Technical Conference, Washington, D.C; Proc., Reservoir Eng., SPE-24930, Oct. 1992. Sorensen, L.E., and Griffith, J.D. 1988. "Evaluation of Solvent and Chase Gas Bank Sizes in the South Swan Hills Hydrocarbon Miscible Flood." 39th Annual Petroleum Society of CIM1CGPA 2nd Quarterly Technical Meeting, Calgary, AB; Preprints V3, No. 88-39-100, Jun. 1988. Stalkup, F.I. 1983. MiscibleDisplacement. SPE Monograph Series, Dallas, TX. Stone, H.L. 1982. "Vertical Conformance in an Alternating Water-Miscible Gas Flood." Paper presented at 57th Annual Fall Technical Meeting, SPE of AIME, New Orleans, LA, Sep. 1982, SPE 11130. Tiffin, D.L. 1982. "Effects of Mobile Water on Multiple-Contact Miscible Gas Displacement." Paper presented at the SPE/DOE Enhanced Oil Recovery Symposium, Tulsa, OK, Apr. 1982, SPE/DOE 10687.
Mukherjee, D. 1958. Internal File Note, Gulf Canada Resources Ltd., Calgary, AR
181
...
.""'" DETERMINATION OF OIL AND GASRESERVES
Todd, M.R., and Longstaff, W.J. 1972. "The Development, Testing, and Application of a Numerical Simulator for Predicting Miscible Flood Performance." 1PT, Jul. 1972, pp. 874-82.
Wood, K.N., Cornish, R.O., Lal, F.S., Taylor, H.O., and Woodford, R.B. 1990. "Solvent Tracersand the Judy Creek Hydrocarbon Miscible Flood." Petroleum Societyof CIM/SPE International Technical Meeting, Calgary, AB; Preprints V2, No. CIM/SPE 90-79, Jun. 1990.
182
c
Chapter 13
ENHANCED RECOVERY BY IMMISCIBLE GAS INJECTION
13.1
INTRODUCTION
Immiscible' gas injection (gasflood) was first used to enhance oil recovery before the turn of the century and actually predates the use of water as an injectant. As with water, gases are used for both their pressure maintenance and their fluid displacement properties. In the case ofgas injection, however, displacement takes a decidedly secondary role. A further difference can also be attributed to the fact that gases can have a substantial degree of mutual solubility with crude oil and hence can, to some extent, do the following:
Notable exceptions can occur, however, when the subject reservoir has any of the following: • A sizable original gas cap • A remote location where gas sales are not feasible • A location that lacks a suitable or economically attractive water source • Extremely low permeabilities, making waterflooding impractical • Water-sensitive minerals
• Vapourize various oil components
• Extreme attic oil losses (e.g., due to adverse coning characteristics)
• Cause contacted oil to expand and mobilize
• Substantial vertical relief
• Reduce viscosity of contacted oil
If any of these conditions are present, the feasibility of employing a gasflood for enhancing the primary recovery mechanism should be considered.
All three phenomena may enhance oil recovery beyond that expected from a simple gas-liquid displacement process. The gasflood injectant that is most commonly used is hydrocarbon-based, not necessarily due to its effectiveness, but rather to its availability and relatively low cost. Other gases that have been or could be successfully employed include (but are not limited to) nitrogen (N,), carbon dioxide (CO,), sulphur gases, flue gas, and air. Despite the fact that the performance of a gas injection scheme can, under some circumstances, compete with or even surpass that of a waterflood, the use of gasflooding has diminished with time particularly during the last 25 years, when natural gas became an increasingly valuable commodity. In addition to the cost of foregone gas sales, high-pressure gas injection schemes also carry added costs associated with high pressure injection flowlines, compression, reprocessing, and possibly the purchase of gas external to the project. These additional costs often make waterflooding or even primary recovery more economic than gasflooding.
• Immiscible refers to gas and oil existingas separate phases in all concentrations everywhere withinthe system.
13.2
TYPES OF FLOODS
Gasfloods have historically been categorized as being one of two types according to where the gas is injected in relation to the oil zone. Figure 13.2-1 schematically illustrates an "external" or updip injection scheme, and a "dispersed" or pattern-style flood. Although both types are subject to similar physical processes and principles, they have by design, different primary gas flow directions (vertical for external; horizontal for dispersed). This can cause them to have very different performance characteristics and hence different prediction technique requirements. External injection schemes are more popular and effective as they are often used to assist a primary gas cap drive, and because they take advantage of the natural phenomenon of gravity segregation or "override," a process that is detrimental to the effectiveness of horizontally flooding with gas. External gravity stable injection projects have exhibited incremental recoveries as high as 40-50 percent. Dispersed gas injection schemes are relatively rare, and when used in the absence of a gravity stable process,
183
DETERMINATION OF OIL AND GASRESERVES
EXTERNAL GAS INJECTION
Oil Gas Production Injection
Oil Production
Gas Injection
- -DISPERSED GAS INJECTION
Oil Production
Gas Injection
Oil Production
--~--------------------------------Figure 13.2-1
Gas Injection
they have historically shown themselves to be only marginally effective, with typical recoveries of only a few percent. This is due to the adverse impact of the strong tendency for gas to find the path of least resistance, either vertically (override) or areally (fingering). Furthermore this tendency is considerably aggravated by the existence of almost ever-present geological heterogeneities. In addition to this distinction, further subsets can occur due to the degree ofpressure maintenance invoked (full or partial) and, in the case of vertical schemes, the existence or nonexistence of a gas cap. Combining both ofthese variables results in vertical floods in which the original gas-oil contact will (I) advance, resulting in a true gas displacement process; (2) remain stable, allowing for some other mechanism to be used to deplete the
184
reservoir; or (3) recede at a controlled rate with some other mechanism employed to deplete the reservoir. It should be noted that for a vertical gas displacement configuration, gas need not necessarily be injected directly into the gas cap or even the structurally highest point as the gas will migrate to these locations of its own accord. This is a particularly useful attribute when flooding dipping reservoirs where considerations such as surface topography may limit access to the structural highs.
13.3
PERFORMANCE PREDICTION
The flood stage at which one of five basic prediction techniques is most appropriate is treated in considerable detail in Chapter II. The reader is encouraged to review this passage for the rationale behind the recommended methods shown in Table 13.3-1.
___________________a
ENHANCED RECOVERY BYIMMISCIBLE GASINJECTION
Table 13.3-1
Recommended Performance Prediction Methods
Stage Exploration/discovery Delineationthrough early life Middle through late life
Prediction Technique Analogies, volumetrics Analogies,numerical simulation, volumetrics, analytical methodsI Numerical simulation. decline analysis
A word of caution in the use of these recommendations is warranted: regardlessofthe depletion stage and techniqueemployed, it is wisewheneverpossible to use more than one procedure as a cross-check or validation process. When gas injection is used primarily for pressure maintenance anddisplacement," production performance prediction methods are eitherfor external injection methods, which are an extension of gas cap drive prediction techniques; or for dispersed injection schemes, which arean extension of solutiongas drivemethods withmany elements similar to horizontal waterfloods. Due to these and previously noted similarities, the various analysis techniques will not be described in as completea manner as they are elsewhere.To avoid repetition, only those aspects that need to be emphasized or that are unique to gasflooding are described here.
13.3.1 External Injection Schemes As noted in Chapter 9, the preferred techniquesinvolve the use of material balance or numerical simulation methods. The analytical Welge (1952) method is also recommended as a shortcut approach. Further to this, however, often it is important to include the effects of gravitydrainage as reportedby ShreveandWelch(1956) and Craig et al. (1957). If decline analysis techniques are to be used for performance prediction purposes, it should be noted that gas-drive-only reservoirs, after an initial period of sustained oil production, often exhibit harmonic declines; the initial rapid decline is caused by the adverse
1
A concise set of examples utilizing classicalanalytical prediction techniques for both external and dispersed injection witheither complete or partial pressure maintenance can be found in Roebuck (1987).
2
If phase behavior effects playa significant role, compositional numerical simulation must be given serious consideration as the preferred prediction technique.
mobilityratio, and the long oil-production tail is caused by gravity drainage.
13.3.2 Dispersed Gas Injection Schemes As with external gas injection projects, the preferred methodfor estimatingrecovery and future performance is numerical simulation-not an easy task as the rapidity and degree of gas break-through are often difficult to simulate. This is a direct consequence of the low viscosity and density of the gas, and its nonwetting characteristics, which combine to generate very high mobility ratios (50 to 100 times that of water) and, as a result, poor sweep efficiencies. Should the lack of time or data not permit a simulation to be undertaken, the analyticalPirson (1958) technique for solution gas drive can be utilized, as can Craig's horizontal displacement technique (Craig et al., 1955). These methods are discussed and recommended in Chapters 9 and II, respectively. The volumetric analysis technique described in Section 11.5 is also applicable, but particular care must be paidto the estimation of horizontaland verticalsweep efficiencies. In addition to the expected mobility-ratioinduced reduction in areal sweep (Dyes et aI., 1954), and the layering-induced reduction in vertical sweep (Stiles, 1949),furtherefficiency losses can occur due to both override and fingering. Analogous reservoirs and mechanistic numerical models may be used to evaluate the significance of these two phenomena. References Craig, F.F. Jr., Geffen, T.M., and Morse, RA. 1955. "Oil Recovery Performance of Pattern Gas or Water Injection Operations from Model Tests." JPT, Jan. 1955, pp. 7-14; Trans., AIME, Vol. 204. Craig, F.F. Jr., Sanderlin, J.L., Moore, D.W., and Geffen, T.M. 1957. "A Laboratory Study of Gravity Segregation in Frontal Drives." JPT, Oct. 1957,pp. 275-81; Trans., AIME, Vol. 210. Dyes, A.B., Caudle RH., and Erikson, R.A. 1954. "Oil Production after Breakthrough as Influenced by Mobility Ratio." JPT, Apr. 1954, pp. 27-32; Trans., AIME, Vol. 201. Pirson, SJ. 1958. Oil Reservoir Engineering. McGraw-HiIl Book Co. Inc., New York, NY, pp. 484-532. Roebuck, J.F. Jr. 1987. "SPE Petroleum Engineering Handbook." SPE of AIME, Chapter 43, Appendix A, pp. 10-13. 185
DETERMINATION OF OIL AND GASRESERVES
Shreve D.R., and Welch, L.W. Jr. 1956. "Gas Drive and Gravity Drainage for Pressure Maintenance Operations." JPT, Jun. 1956, pp. 136-43; Trans., AIME, Vol. 207. Stiles, W.E. 1949. "Use ofPerrneability Distribution in Water Flood Calculations." Trans., AIME, Vol. 186,pp. 9-13.
Welge HoI. 1952. "A Simplified Method for Computing Oil Recovery by Gas or Water Drive." Trans., AIME, Vol. 195, pp. 91-98.
186
s
Chapter 14
ENHANCED RECOVERY BY THERMAL STIMULATION
14.1
INTRODUCTION
The thermal recovery processes that have been used extensively for the recovery of heavy oil and bitumen from the oil sands have met with mixed success. The term "heavy oil" is used to designate crude oils having an API gravity range of IS to 25 degrees. Heavy oil is literally heavier, thicker and slower to pour than the conventional light and medium crudes. Heavy oil, however, is relatively mobile at reservoir conditions and can be successfully produced by primary recovery methods. Thermal recovery processes are then used to further increase the recovery of heavy oil. The bitumen found in the oil sands deposits is a viscous mixture of hydrocarbons with an API gravity of less than 15 degrees and a viscosity of several thousand centipoise at room temperature. Thus, bitumen is not economically recoverable in its natural state by conventional primary or secondary recovery methods. The ultimate objective of any thermal process is to improve the mobility ofthe crude by reducing its viscosity through the introduction of heat into the reservoir. In addition, steam pressure and thermal expansion also enhance the driving forces present in the reservoir: gravity drainage, solution gas drive, and reservoir compaction. The following are the thermal processes most commonly used for the recovery of heavy oil and bitumen: • Cyclic steam stimulation • Steam flood • In situ combustion • Electromagnetic oil heating These are discussed in the sections that follow.
14.2
CYCLIC STEAM STIMULATION
Cyclic steam stimulation is probably the most widely used thermal recovery process at the present time. The popularity of this process is mainly due to its relative ease of application, the low initial capital required,
and the quick return on investment. The ultimate oil recovery from this process (15·20 percent) is generally much lower than recovery from steam flood (20·50 percent). However, most steam stimulation processes may be converted to steam flood once inter-well heat communication has been established. Cyclic steam stimulation is a single well process with injection and production carried out at the same well. Steam is injected into the well for a certain length of time, usually at a rate and steam quality that are relatively constant (60.80 percent cold water equivalent at wellhead). Generally, the steam injection rate is the maximum rate obtainable at bottom-hole pressures below the formation fracture pressure. The bottom-hole steam quality and pressure may be predicted using wellbore models ofthe type discussed by Fontanilla and Aziz (1982) and others (Farouq Ali, 1981; Durrant and Thambynayagam, 1980; Willhite, 1966). The well is allowed to soak for a period (of at least a few days) that depends on the volume ofsteam injected; soaking allows the injected steam to condense and distribute the heat more evenly. At the end of the soak period, the well is put on production. The reservoir pressure during the initial production period is very high, and fluids are able to flow back under the reservoir pressure alone. The production during this flowback period consists mostly of hot water, flashed steam, formation gases, and traces ofoil. Upon completion ofthe flowback period, the reservoir pressure will have dropped and a bottom-hole pump will be required to lift the reservoir fluids. These injection-production cycles are repeated until the oil production rate drops below the economic limit. At this stage, other thermal recovery methods such as steam flooding or in situ combustion may be considered.
14.2.1 Process Variation The cyclic steam stimulation process is sometimes modified in order to improve its sweep and thermal
187
.,-1 C.',
!
DETERMINATION OFOILANDGASRESERVES I
efficiencies. Laboratory studies and field tests have been conducted to investigate the addition of chemicals or gases to the injected steam. These include surfactants, carbon dioxide, ethane, naphtha, methane, propane, butane, heptane, natural gas, air, and oxygen (Kular et al., 1989; Ploeg and Duerkson, 1985; Ivory et al., 1991; Pursley, 1974; Waxman et aI., 1980). The following are the major mechanisms by which steam additives improve oil recovery: I. The diversion ofsteam to higher oil saturation zones improves sweep efficiency.
2. The reduction in surface tension ofthe oil improves displacement efficiency. 3. Gas expansion and flashing of solution gases provide an additional driving force in the reservoir. Although steam additives seem to offer some potential under certain reservoir and operating conditions, further research and testing are needed to improve the recovery of the cyclic steam stimulation process.
14.2.2 Field Examples The following are some of the steam stimulation and steam flood projects in Canada and the United States: • Shell Peace River Thermal Pilot (Waxman et aI., 1980) • Husky Paris Valley Cyclic Gas-Steam Pilot (Meldau et aI., 1981) • Petro-Canada PCEl Steam Stimulation Project (Towson and Khallad, 1991) • Amoco Gregoire Lake In Situ Steam Pilot (Kular et aI., 1989) • Chevron Keen River Steamflood Project (Oglesby et al., 1982) • Esso Cold Lake Thermal Project (Mainland and Lo, 1983) • Athabasca Oil Sands Project
14.2.3 Recovery Mechanisms Cyclic steam stimulation and steam flood recovery mechanisms are as follows: I.
Reduction ofoil viscosity due to increased temperature
2.
Steam pressure providing the drive energy for oil to flow towards the producing well
3. Gravity drainage of the liquid phases (Denbina et aI., 1987; Cardwell and Parson, 1949; Farouq Ali, 1982)
4. Thermal expansion ofoil providing energy for fluid flow (Denbina et aI., 1987) 5. Reservoir compaction maintaining reservoir pressure (Denbina et al., 1987) 6.
Steam distillation causing the lighter hydrocarbons to separate from the heavy ends and form a miscible oil bank ahead ofthe steam front
14.2.4 Design Considerations Screening guidelines have been developed by many researchers (Adams and Khan, 1969; Belyea, 1956; Boberg and Lantz, 1966; Buckles, 1979; Bums, 1969; Crawford, 1971; Doscher, 1966; Gontijo and Aziz, 1984; Prats, 1978; Shepherd, 1979; Williams et aI., 1980) in order to define the reservoir and fluid properties under which steam stimulation processes are most likely to be economical. The following guidelines are based on the results of some successful projects: Formation thickness (m) Depth (m)
400 to 1000
Porosity (% PV)
> 30
Permeability (mD)
250 to 1000
API gravity (degrees)
10 to 34
Oil viscosity at reservoir conditions (mPa.s)
< 15,000
Initial oil saturation (% PV)
> 40
> 10
The mechanisms involved in a steam stimulation process are very complex. Methods used to predict performance are only approximate at best because of the many simplifying assumptions that have to be made. Nevertheless, there are certain prominent factors that may affect the oil recovery in a steam stimulation process: the volume of steam injected, steam quality, injection pressure, and reservoir thickness. The amount of heat injected determines the volume of heated reservoir and ultimately the percentage of oil recovery. A thick pay zone is also desirable for effective gravity drainage (Butler et aI., 1981; Dykstra, 1978). The depth ofthe reservoir is another important factor. Deep reservoirs (2000-3000 m) may not be suitable for steam stimulation, because of the large heat losses from the wellbore. On the other hand, shallow reservoirs (200250 m) may not allow high enough injection pressures to maintain reasonable steam injection rates and provide sufficiently high steam temperatures for the reduction of oil viscosity. The well patterns most commonly used for the cyclic steam stimulation process are the 5-spot and the 7-spot, which allow the conversion of cyclic steam stimulation
188
_______________________a
ENHANCED RECOVERY BYTHERMAL STIMULATION
to steam flood later if desired. Single weIl tests (Dillabough and Prats, 1974) are generally conducted to obtain preliminary data on recovery potential, operating costs, and other design factors. Well spacing may vary from 0.4 to 2 hectares. Infill drilling has also been used to exploit developed heat zones and achieve early inter-well communication. Another common practice in commercial projects is to drill clusters ofdeviated wells from a single well pad in order to optimize the use of land and surface facilities.
14.3
STEAM FLOODING
. In the steam flood process, steam is injected into the reservoir on a pattern basis, much like a waterflood. Various well patterns, including the 5-spot and 7-spot, have been employed. The injected steam reduces the viscosity of the oil and provides the driving force required to move the oil towards the producing wells. In the application ofthe steam flood process to oil sands deposits, it is essential to achieve flow communication between the injector well and the producers prior to flooding. Frequently, the wells are produced by steam stimulation for a few cycles until communication between wells has been established, and then steam flooding is started. Other naturally existing communication paths in the oil sands deposits, such as bottom water and high permeability layers, may provide valuable means of improving injectivity for effective reservoir heating. Lack of steam injectivity may require hydraulic fracturing ofthe wells before steaming. Steam flooding with continuous steam injection can recover significantly more oil (up to 50 percent) than steam stimulation alone (10-25 percent). However, there are disadvantages associated with steam flooding. It generally results in higher steam-oil ratios than cyclic steam stimulation because of the much larger volume of reservoir that must be heated before any significant oil recovery is realized. The amount of reservoir heating required in cyclic steam stimulation is confined to the near-wellbore region, and oil production is therefore realized much earlier.
variety of surfactants. CommerciaIly available surfactants are now chemically stable at temperatures up to 300°C. However, the in situ behaviour offoam is still not fully understood, and field tests (Kular et al., 1989; Patzek, 1988; Sander, 1991) indicate that its propagation in the porous medium is very slow. In most cases the cost of surfactants offsets the possible benefits. Field pilots have been conducted to test the injection of gas atthe boundaries ofthe steam zone to improve steam confinement and to maintain the pressure in the steam zone. The injection of air with steam provides another less expensive alternative to the use of surfactants. The steam-air process works on the assumption that low temperature oxidation produces coke particles that tend to plug the pore throats and provide resistance to flow (Ivory et al., 1991). Thus, steam is diverted to other parts of the reservoir, and the result is an improved sweep efficiency.
14.3.2 Design Considerations The following guidelines may be used to screen reservoirs for potential steam flood applications. However, these guidelines are only approximate as geological heterogeneities specific to each reservoir cannot be accounted for. Formation thickness (m)
6 to 20
Depth (m)
100 to 500
Porosity (% PV)
>30
Permeability (mD)
>500
API oil gravity
< 25
Reservoir oil viscosity (mPa.s)
5000 to I 000 000
Initial oil saturation (% PV)
> 50
A number ofadditives have been injected with steam to improve the oil recovery by the steam flood process. These additives improve the thermal and sweep efficiencies of the injected steam by diverting it towards the colder regions of the reservoir.
Recovery with steam flood is approximately 40 to 50 percent ofthe original oil in place, with steam-oil ratios in the range of 5 to 7. The steam-oil ratios are dependent upon the nature of the reservoir. Very deep reservoirs (1200-1500 m) may be impractical for steam flooding due to the excessive heat losses in the wellbore and the very high steam pressure required at the surface. The reservoir should be at least 6 to 10 metres thick to minimize heat losses to the overburden and underburden. Successful steam flood processes are generally in shallow (300 m) reservoirs having reasonably high porosity (30-40 percent pore volume), permeability less than I darcy, and oil saturation of 85-90 percent pore volume.
Patzek (1988) and others (Kular et al., 1989; Ploeg and Duerkson, 1985; Sander, 1991; Suffridge, 1991; Butler, 1986) have reported mixed success using a
Injection rates for steam flood are generally designed to compensate for heat losses to the adjacent formations while providing effective heating of the reservoir.
14.3.1 Process Variation
189
DETERMINATION OF OIL ANDGASRESERVES
During pilot testing, steam injection rates should also compensate for the heat flowing out of the pattern due to the lack of steam confinement. Steam flood processes are usuaIly started at high injection rates, which are later optimized once steam gravity override or steam breakthrough occurs (Myhill and Stegeimeier, 1978; Chu and Trimble, 1975; Ali and Meldau, 1979; Bursell and Pittman, 1975; Vogel, 1982; Belvins, 1978; Stokes, 1978; Van Dijk, 1968). Where fluid communications have already been developed through cyclic stimulation of the wells, injection rates can be optimized at start-up.
consists of a thin highly water-saturated zone near the bottom of the formation and a fining upward sand sequence. This results in good thermal efficiency and high oil production rates. Gas Cap. The presence of a gas cap will tend to channel injected steam to the top of the formation, resulting in excess heat loss and poor thermal efficiency. However, the extent of the gas cap is a critical factor especially if gravity drainage is the predominan; production mechanism (Kular et a!., 1989). Blocking agents may be used to improve the vertical sweep efficiency (Sander, 1991).
However, many factors must be taken into account in designing a steam flood process: the mineral content of reservoir rock, the availability of fuel and water, the analysis of crude oil, sand production, water disposal wells, water treating requirements, production facilities to handle hot fluids, emulsion treating, and transportation of heavy crude.
Shale. The presence of a substantial and impermeable shale layer near the middle of the formation may prevent the rise of the steam zone, resulting in poor volumetric sweep and heat efficiencies.
14.4
Lack of Steam Confinement. If the oil sand deposit contains natural fractures (e.g., the Carbonate Trend in northern Alberta), a significant fraction of the injected steam may be lost. Poor steam confinement may significantly reduce the energy available in the heated zone to drive the fluids towards the producing well.
CAUSES OF FAILURE FOR CYCLIC STEAM STIMULATION AND STEAM FLOOD PROCESSES
The following reservoir and operating limitations may cause the cyclic steam stimulation and steam flood processes to become uneconomical: Lack of Injectivity. Some oil sands deposits have such a high saturation of bitumen that the steam has great difficulty penetrating the highly viscous oil bank. As a result, the steam tends to channel to the poorer part of the formation, which has lower oil saturation and higher water saturation. Clay swelling due to incompatibility between the injected water and the formation water may also limit steam injectivity. Bottom Water. The term "bottom water" refers to sand layers containing mobile water that account for more than 20 percent of the formation thickness. Such bottom-water layers are detrimental to the cyclic steam stimulation process. Due to the much higher mobility of steam in the water zone, most of the injected steam will be lost to the water zone, resulting in very poor thermal efficiency. During the production cycle, the cold water is much more mobile than the bitumen and will tend to be produced first. In addition, the cold water will tend to cool the oil around the wellbore and reduce the volume of the heated zone. On the other hand, a thin bottom-water sand can be used effectively to heat the formation. For example, the steam stimulation process is very successful at the Peace River Pilot (Waxman et aI., 1980) where the oil sand deposit
Thin Formation. Very thin formations may result in excessive heat loss to the overburden and underburden, leading to poor heat efficiency.
Low Porosity and Permeability. Some heavy oil deposits such as oil shales have such low porosity (less than 20 percent by volume) and low permeability (less than 100 mD) that the steam injectivity may be seriously limited. Hydraulic fracturing is required to exploit such heavy oil deposits (Kular and Chinna, 1988). Poor Reservoir. Due to the high initial capital investment and operating costs of the steam processes, reservoirs with less than 40 percent oil saturation are not likely to be economically recoverable by these processes. _. Shallow Reservoirs. Shallow reservoirs with insufficient overburden will tend to limit the steam injection pressure, and thus reduce oil productivity. Deep Reservoirs. Very deep reservoirs have such high reservoir static pressure that the steam injectivity may be limited. The oil sand deposits in Alberta generally require fracturing before steam can be injected at a reasonable rate. A deep reservoir means a higher steam injection pressure, which requires the added expense of high pressure steam generators. Also, deep reservoirs cause excessive heat losses from the wellbore, resulting in the injection of poor quality steam.
190
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _1
ENHANCED RECOVERY BY THERMAL STIMULATION
14.5
FORECASTING MODELS
K,b = overburden thermal conductivity
A number of options are available to engineers for predicting the performance of thermal recovery processes. These include numerical simulation models, analytical models and simple correlation equations. Ideally, reservoir simulation models will provide the most accurate answer. However, these models cannot be utilized in cases where only limited data is available. Time and, to some extent, cost limitations may also work against the use of numerical simulation models. As an alternative, analytical models may be used quite effectively for process design and forecasting oil recovery. The analytical models for steam recovery processes are generally divided into three types. Figure 14.5-1 illustrates the distribution of fluids as assumed in these analytical gravity drainage models. Frontal Displacement Model. This model assumes a cylindrical steam zone, with displacement of oil over the full thickness of the oil zone. Steam Overlay Model. This model assumes that the steam lies directly over the oil zone, and the principal direction ofsteam zone growth is vertically downwards.
~T
=
D = t = P = A =
M, = h =
Marx and Langenheim's solution to Equation (I) is given in Equation (2): A(t)
14.5.1 Marx and Langenheim Model Marx and Langenheim developed a frontal displacement model in which the growth of the steam zone depends on the rate of steam injection and the loss of heat to the overburden and underburden. The heat balance equation used in this model is written as: Hi = heat loss + heat accumulation =
where
2f'[o "ltD(t-P) Kob~T J(dA) dP + M,hdT dA dP dt Hi = constant heat injection rate (kJ/d)
(I)
=
(HiMrhD) f(x)
(2)
4Kob~T
where A(t) = area of steam zone (rn-) x = dimensionless parameter
Conical Steam Zone Model. This model assumes that the steam zone has the shape of an inverted cone. The steam not only rises upward but also expands outward due to heat conduction and the drainage of the heated oil toward the wellbore. Some of the most commonly used models for predicting the production rates of the cyclic steam stimulation and steam flood processes include those by Marx and Langenheim (1959), Myhill and Stegeimeier (1978), Vogel (1982), Butler (1986) and Butler et al. (1981). The development and applications of these models are presented in the following subsections.
(kJ/m/d/°C) injection temperature minus initial formation temperature (0C) overburden thermal diffusivity (m2fh) time (d) time (d) area of steam zone (m-) volumetric heat capacity of formation (kJ/m3jOC) pay thickness (m)
f(x)
2x I = e,2 (erfcx) +--
(4)
.,fit
2 erfcx = I - erfx = I - -
Ie·' .,fit, x
2
dP
(5)
Applying D' Arcy's law for conditions of gravity drainage, the rate of oil displacement, qo' in m3/d from the steam zone may be written as: qo = [
where
Hi
(6)
porosity (fraction) initial oil saturation (fraction) residual oil saturation (fraction) temperature difference between steam and initial reservoir temperature (T, - To) (0C) Ti = injection temperature eC) To = initial formation temperature (0C) erf = error function
191
DETERMINATION OF OIL AND GASRESERVES
::::::: Cold Oil·-------------------~----- Zone
Frontal Displacement Model
T
. ,' .
. Steam Zone:"
::::::::::::::: Cold Oil Zone :::::::::::::::::::::::::::
---------------------------------------------------------Steam Overlay Model
--,, , , ,,
, ,,, ,,, ,, ,, -
-------~--
-------- ---
,
.-~-----
,
~'
,,
, ,,, ,, ,
, ,,
__ -
0°
,,
". t· e~('('\ ·
S"',one
0;
_\
Q
C,/.- P,
.:.:
o'
_~oo
-
_-_,
. ~/-:.:: =--~\O 0\\: -:. z.One
----
-
<:- .-§,'-::~- --::--0
0
-v-z>: _-::--
"'("-,,.-
,.
-- --:::.-- _"°0 -_-_::t.. -::--::.-_-::~o -;:'-::::::::--
e 0/1,.:::
_- _-_-- _ _-\\ 0 e 0 - - _- - _-"\00
-::::-- -
e ;-::---0',_ -
~.I
. J'-_-_ _-
:: - - : : ~ o __ -_\,,0
:: -
--~
<:
c a,i.. :: - -- _ - _ _
:'&~-:::
_-::_.-.. °0
_- __ -- _ - _'0 ----
.:
0,' --::.::, .. _ ,~ - : : -
I
0
c;L
,
\
-'.
" ......
Q
c
f..l--
-
_::.:: - -
0
Conical Steam Zone Model
Source: After Gontlio and Azlz, 1984.
Figure 14.5-1
Types of Analytical Gravity Drainage Models
192
c
ENHANCED RECOVERY BY THERMAL STIMULATION
erfc
complementary error function (Abramowitz and Stegun, 1964)
=
14.5.2 Myhill and Stegeimeier Model Myhill and Stegeimeier presented an analytical model using a simple energy balance to calculate the steam zone size. This energy balance approach is based on the assumption that the oil ultimately produced from both steam stimulation and steam flood processes is proportional to the steam zone volume. Other assumptions made in developing the model are that the steam zone is cylindrical in shape, and that the thermal properties in the reservoir, the heat losses, and the steam injection rates are constant. It is also assumed that the oil-steam ratios of any thermal process can be expressed in terms of a thermal efficiency term, E hs, that is defined as the ratio of heat remaining in the steam zone to the total heat injected (Figure 14.5-2).
s, = 1. (e·n erfc-F,; + 2 ~ n
tn
1)
tD =
1.0
.
~
w oj
c:
0.8
-\
'\
I\.
~
'"
c7J 0.6
I\-
1>' c
-, ""'"
-,
'0
E
~ I-
h D is the ratio of latent heat to total energy injected l-,
-,
i'--
~
-,
'0 0.4
-:::: l-, .,f;o. <,
<, <,
'iii
0.2
r---...
Snurce: After Prats, 1969.
0.667 ' I 0.5
,,~
- - ,o.~ ~ --<;
<,
0.091
0.0 0.01
(9)
(8)
-, I'\.,
Q)
f"L, CwLlT + 1 ---:=.....:...-
Ifthe thermal efficiency and enthalpy ratios are known, it is possible to calculate the maximum oil-steam ratio, OSR, using the following equation:
I::-..
-,
<,
-,
ffi
D
<,
-,
-,
E
h =
(7)
Z; (M I ) '
-
Figure 14.5-2 is a graph of the thermal efficiency, E hs, vs. the dimensionless time, tD' which can be used to estimate the thermal efficiency of the steam processes (Prats, 1986). The ratio, hD, oflatent heat to total energy injected is given by:
where fsd = downhole steam quality (fraction) L, = latent heat of vapourization of steam (kJlkg) LlT = injection temperature minus initial formation temperature (0C)
The dimensionless time parameter, tD' is given by: 35 040 kh,M,t yn
where khz = heat conductivity of steam zone (kJ/m/d/°C) M 2 = volumetric heat capacity of cap rock (kJ/m3/ 0C) t)TS = time of injection (years) Z, = gross thickness of reservoir (m) M 1 = average heat capacity of steam zone (kJlkg/°C)
0
0.1
~
<;
---- -10
r:::: ::::
100
Dimensionless Time, tD
Figure 14.5-2 Thermal Efficiency of Steam Zone as a Function of the Dimensionless Time Parameter 193
DETERMINATION OFOILAND GASRESERVES
PwCw(l + hD ) E"q>dS OSR=
(~,) (I 0)
I
M,
where Pw = density of water (kg/m') C; = specific heat of water (kJ/kg/°C) h o = ratio oflatent heat to total energy injected ~, = thermal efficiency (fraction) q> = porosity (fraction) dS = difference between steam temperature and initial reservoir temperature CC) Z, = net thickness of reservoir (m) Z, = gross thickness of reservoir (m)
condensate and heated oil flow by gravity to a horizon_ tal production well located at the bottom ofthe chamber and are removed continuously. The expression for the oil drainage rate, Q, is based on the gravity drainage theory and is given by: Q=2 where Q q> So k g
= = = = a. =
14.5.3 Vogel Model Vogel's steam overlay model is based on ultimate heat requirements determined from simple two-dimensional heat flow equations. The total heat requirement is equal to the sum of the heat lost from the reservoir, the heat conducted to the produced fluids, and the heat that remains in the steam zone. The heat requirement, Q,o,al' is given by:
Q"", = Ah (p,C,)
I "'Ifitii;
+ 2K,A
(II) where A h
x,
= = = =
K2
= time (d) = thermal diffusivity of overburden (m2/d) = thermal conductivity of underburden
PsC,
project area (m") thickness of steam zone (m) heat capacity (kJ/mlfOK) thermal conductivity of overburden
(kJ/m/°K/d)
(kJ/m/°K/d) 0. 2
= thermal diffusivity of underburden (mvd)
14.5.4 Butler Model The conical steam zone model developed by Butler (Butler et al., 1981; Romney et al., 1991; Dugdale, 1986) is based on the assumption of continuous steam injection into a growing steam-saturated volume or chamber. Steam flows to the boundary ofthe chamber, condenses, and gives up its heat to the surrounding oil sands. The
h = m = v, =
14.6
2q>S,kgo.h my,
(I 2)
oil drainage rate (ml/d/m length of horizontal well) porosity (fraction) initial oil saturation (fraction) effective permeability to oil (um") gravitation constant (9.81 m/s") thermal diffusivity of reservoir material (rnvd) pay zone thickness (m) bitumen viscosity exponent (usually = 3) kinematic viscosity of oil at steam temperature (m2/d)
IN SITU COMBUSTION PROCESSES
In a combustion process, air is injected into one well and the formation is ignited. As the burnt front moves through the reservoir, a portion of the bitumen is consumed as fuel and combustion gases and steam are generated. These hot fluids raise the temperature and reduce the viscosity ofthe bitumen, which is then driven towards the production wells. In situ combustion projects in Canada and the United States include the following: • PetroCanada Viking-Kinsella Wainwright B Oxygen Fireflood Pilot (Dugdale, 1986; Dugdale et al., 1985) • Panf'anadian Countess Fireflood Pilot (Metwally, 1991) • BP Cold Lake Pressure-Up Blow-Down Wet Combustion Pilot (Mehra, 1991) • Murphy Eyehill In Situ Combustion Pilot (Farquharson and Thornton, 1985) • Amoco Athabasca In Situ Combustion Project (Jenkins and Kirkpatrick, 1979) • Mobil Kern County South Belridge In Situ Combustion Project (Gates et al., 1978) • Home Oil Silverdale Water Alternating Gas Project (Hanna, 1987)
194
-
..a
ENHANCED RECOVERY BY THERMAL STIMULATION
• Texaco Caddo Pine Island In Situ Combustion Pilot (Horne et al., 1979)
14.6.1 Recovery Mechanisms The following are the major recovery mechanisms of the in situ combustion process: Oxidation of Crude. The temperature at which oxidation takes place depends on the concentration of oxygen. High-temperature oxidation uses up the oxygen and generates heat. Low-temperature oxidation promotes the formation of fuel and spontaneous ignition. Thermal Cracking. Thermal cracking or pyrolysis of the crude generates light hydrocarbons and leaves coke behind as fuel. Steam Distillation. Steam generated by oxidation at the combustion front evaporates the light hydrocarbons from the crude. These are displaced ahead of the steam front to form an oil bank. Steam Drive. Steam provides the energy to drive the heated oil ahead of the combustion front. Thermal Expansion. Thermal expansion of crude, combustion gases, and light hydrocarbons also provide the driving force to drive the heated oil towards the production well. Gravity Override. Steam, combustion gases, and light hydrocarbons are lighter than the crude oil and tend to rise to the top of the formation, bypassing some of the crude oil in the middle or lower part of the formation. Viscosity Reduction. Heat generated by combustion raises the temperature ofthe formation and significantly reduces the viscosity of the crude.
14.6.2 Process Variations Although the in situ combustion process is more energy-efficient than cyclic steam stimulation or steam flood and can be used in thinner pay zones, the heat efficiency of the dry combustion process is still very low. About 70 percent ofthe heat generated at the high temperature combustion front is left in the burnt zone. The following modifications are required to improve the heat efficiency of the dry combustion process: Thermal Wave Process. This technique involves the dilution of the injected air with combustion flue gas to increase the heat capacity of the injected air. Combined Thermal Drive. This is a wet combustion process designed to improve the sweep efficiency and reduce the volume of air required. It involves the simultaneous injection of air and water and results in
lower air requirements and higher oil recovery. Field results show that the simultaneous injection of air and water is more effective than the injection of a slug of water following air injection. The most important consideration in this process is to ensure that sufficient water is injected for conversion to steam without quenching the combustion. The required water-air ratio (WAR) for a given reservoir is calculated from a material and heat balance. Combination of Forward Combustion and Waterflooding. In this process, referred to as COFCAW, the water-air ratio is high enough to quench the combustion. Low temperature oxidation occurs in the steam zone to maintain the steam temperature. Steam Stimulation Followed by Wet Combustion. In reservoirs containing a very viscous crude oil (i.e., bitumen), the mobility of the crude is too low to allow economic production rates for the combustion process. Cyclic steam stimulation has been used in a number of fields to increase the mobility of the crude, create a communication path between wells and allow the combustion front to move towards the production well more rapidly. Enriched Air Combustion Process. Oxygen-enriched air and pure oxygen are being used in this process. The following are the potential advantages of using pure oxygen instead of air: • High displacement rate • Lower gas injection volumes resulting in fewer operating problems for the compressor • Increased mobility of the cold oil due to the dissolution of carbon dioxide in the oil • Higher recovery factors • Larger well spacing, which reduces the infill drilling • Flammable produced gases may be separated and used as fuel An alternative to this process is to gradually increase the oxygen content of the air from about 30 percent to 95 percent. Laboratory results show that the injection of 99.5 percent oxygen should result in a combustion gas primarily composed of carbon dioxide. This may reduce the oil viscosity and cause some swelling of the crude.
14.6.3 Design Considerations Factors influencing the selection of well patterns include the reservoir dip angle and the utilization of existing wells. Because of the high mobility of air compared to that ofoil, usually a few injection wells are sufficient to 195
DETERMINATION OFOILAND GASRESERVES
sustain the fireflood with a large number of production wells. In situ combustion pilots usually experiment with different well patterns and spacings. The inverted 9-spot pattern, inverted 7-spot pattern, confined 5-spot pattern, line drive, and single well injection have all been commonly used. For example, Amoco's in situ combustion pilot (Jenkins and Kirkpatrick, 1979) in Athabasca started with a two-well test with a distance oDO m (100 feet) between the wells. Then different well patterns, ranging from a 0.2 ha (1/2 acre) 5-spot to a 4 ha (10 acre) 9-spot, and finally a 1 ha (2.5 acre) 5-spot, were tested. The design criteria for in situ combustion processes are as follows: Formation thickness (m) Depth (m) Porosity (% PV) Permeability (mD) Oil gravity ( degree API) Initial oil viscosity (mPa.s) Initial oil saturation at reservoir conditions (% PV) Type of formation
3 to 15 < 3500 > 35 > 100 10 to 35 < 10,000
> 10 Sand or sandstone and carbonates with high porosity, no gas cap or bottom water
14.6.4 Causes of Failure An in situ combustion process may fail for any of the following reasons: 1. Low oil saturation in the formation may not deposit enough fuel to support combustion. Incomplete oxygen consumption due to the lack offuel or early break-through of combustion gases at the production well may limit inflow into the wellbore and cause reduced pump efficiency. 2. Low air injectivity may be caused by a water zone near the wellbore, formation plugging, or oil droplets present in the compressed air. Low permeability zones in the formation also cause problems in the removal of the combustion gases, which consist mainly of nitrogen and carbon dioxide. 3. Reservoir heterogeneities that cause channelling and leaking of the injected air from the burnt zone will result in poor sweep efficiencies. 4. Low gravity oils characterized by high fuel content may require a large volume of air for combustion. 196
5. Explosions could occur in injection lines, injection wells, and air compressors. Tubulars may be destroyed by high temperatures due to the breakthrough of fire front at the production well or backburn at the injection well. Corrosion may reduce the life of pumps and surface facilities. 6. Tight emulsions are often created during in situ combustion. Emulsified fluids cause rod fall problems and high flowline pressure because of their high viscosities. The operation of the skim tanks and separators may be affected because the tight emulsions are very difficult to break. 7. Sand production problems caused by large volumes of combustion gases may result in operating and erosion problems in pumps and surface equipment. Severe gas locking may also lead to dry stroking and will accelerate pump failure due to the lack of lubrication.
14.7
ELECTROMAGNETIC HEATING
Two different methods of electrical stimulation have been field-tested in Canada. Both use the reservoir as a resistive element that heats up as electrical power is applied. This reduces the oil viscosity, thus improving oil production rates. In the first method (Romney et aI., 1991), electrical current at a frequency of 60 Hz is delivered from one well to another. In the second method, a single well acts as the electrical injector and ground return well. This model has been applied to a number of field tests both in Canada and worldwide, with varying degrees of success. The mechanics ofthe second method require electrical current to be transmitted through the formation-pay zone and back up the production casing. Short-circuiting is prevented by using nonconductive materials on the casing and production strings. Romney et al. (1991) discusses the design of single well electromagnetic stimulation in detail.
References Abramowitz, M., and Stegun, LA. 1964. Handbook of Mathematical Functions with Formulas. Graphs. and Mathematical Tables. US Department of Commerce, National Bureau of Standards, Applied Mathematics, Series 55, Jun. 1964. Adams, R.H., and Khan, A.M. 1969. "Cyclic Steam Injection Project Performance Analysis and Some Results of a Continuous Steam Displacement Pilot." JPT, Jan. 1969, pp. 95-100. Ali, S.M., and Meldau, R.F. 1979. "Current Steam Flood Technology." JPT, Oct. 1979, pp. 13321342.
ENHANCED RECOVERY BYTHERMAL STIMULATION
Belvins, T.R. 1978. "Analysis ofa Steam Drive Project, Inglewood Field, California." JPT, Sep. 1978, pp. 1141-1150. Belyea, H.R. 1956. "Grosrnont Formation in the Loon Lake Area." Journal ofAlberta Society of Petroleum Geology, Vol. 4, p. 66. Boberg, T.C., and Lantz, R.B. 1966. "Calculation of the Production Rate of a Thermally Stimulated WelL" JPT, Dec. 1966, p. 1613. Buckles, R.S. 1979. "Steam Stimulation Heavy Oil Recovery at Cold Lake, Alberta." Paper presented at AIME-SPE California Regional Meeting, Ventura, CA, Apr. 1979, SPE 7994. Bums, J.A 1969. "A Review of Steam Soak Operations in California." JPT, Jan. 1969, pp. 2534. Bursell, C.G., and Pittman, G.M. 1975. "Performance of Steam Displacement Kern River Field." JPT, Aug. 1975, pp. 997-1004. Butler, R.M. 1986. "Thermal Recovery." Course notes (copyright 1986), The University of Calgary, Calgary, AB, pp. 2.1-4.46. Butler, R.M., McNab, G.S., and Lo, H.Y. 1981. "Theoretical Studies on the Gravity Drainage of Heavy Oil During In situ Steam Heating." Canadian Journal ofChemical Engineering, Vol. 59, Aug. 1981. Cardwell, W.T., and Parson, R.L. 1949. "Gravity Drainage Theory." Trans., AIME, Vol. 179, p.199. Chu, C., and Trimble, AE. 1975. "Numerical Simulation of Steam Displacement-Field Performance Applications." JPT, Jun. 1975, pp. 765-776. Crawford, P.B. 1971. "Thermal Recovery Guide Helps Select Projects." World Oil, Aug. I, 1971, pp. 47-48, 53. Denbina, E.S., Boberg, T.C., and Rotter, M.B. 1987. "Evaluation of Key Reservoir Drive Mechanisms in the Early Cycles of Steam Stimulation at Cold Lake." Paper presented at 62nd Annual SPE Technical Conference, Dallas, TX, SPE 16737. Dillabough, J.A., and Prats, M. 1974. "Proposed Pilot Test for Bitumen Recovery From the Peace River Tar Sand Deposit, Alberta." Paper presented at Symposium on Heavy Crude Recovery, Maracaibo, Venezuela, Jul. 1974.
Doscher, T.M. 1966. "Factors Influencing Success in Steam Soak Operations." Petroleum Industry Conference on Thermal Recovery, Los Angeles, CA, Jun. 1966, pp. 76-80. Dugdale, PJ. 1986. "Comparison of Recovery and Economics For Oxygen and Air Fireflood in Canadian Heavy Oil Areas." Paper presented at SPE-DOE Fifth Symposium on Enhanced Oil Recovery, Tulsa, OK, Apr. 1986, SPEIDOE 14921. Dugdale, PJ., Fabes, L., Saizew, H., and Khallad, K. 1985. "Design and Operation of the VikingKinsella Wainwright B Oxygen Fireflood Pilot." Paper presented at the South Saskatchewan Section CIM Petroleum Conference, Regina, SK, Sep.1985. Durrant, AJ., and Thambynayagam, C. 1980. "Wellbore Heat Transmission and Pressure Drop for SteamlWater Injection and Geothermal Production: A Sample Solution Technique." SPE Reservoir Engineering, Mar. 1980, pp. 148-162. Dykstra, H. 1978. "The Prediction of Oil Recovery by Gravity Drainage." JPT, May 1978, p. 818. Farouq Ali, S.M. 1981. "A Comprehensive Wellbore Steam/Water Flow Model for Steam Injection and Geothermal Applications." SPEJ, Oct. 1981, pp. 527-534. - - - . 1982. "Elements of Heavy Oil Recovery." Course notes (copyright 1982), University of Alberta, Edmonton, AB, pp. 47-60. Farquharson, R.G., and Thornton, R.W. 1985. "Lessons From Eyehill." Paper presented at the First Annual CIM Technical Meeting, South Saskatchewan Section, Regina, SK, Sep. 1985. Fontanilla, J.P., and Aziz, K. 1982. "Prediction of Bottom-Hole Conditions for Wet Steam Injection Wells." JPT, Mar. 1982, pp. 80-88. Gates, C.F., Jung, K.D., and Surface, R.A 1978. "In Situ Combustion in the Tulare Formation, South Belridge Field, Kern County, CA." JPT, May 1978, pp. 798-802. Gontijo, J.E., and Aziz, K. 1984. "A Simple Analytical Model For Simulating Heavy Oil Recovery by Cyclic Steam in Pressure Depleted Reservoirs." Paper presented at SPE Annual Technical Conference and Exhibition, Houston, TX, Sep. 1984, SPE 13037.
197
DETERMINATION OFOILAND GASRESERVES
Hanna, M. 1987. "The Silverdale Water Alternating Gas Project." Paper presented at the First Annual CIM Technical Meeting, South Saskatchewan Section, Regina, SK, Oct. 1987. Horne, J.S., Bousaid, I., Dore, T.L., and Smith, L.B. 1979. "Initiation of an In situ Combustion Project in a Thin Oil Column Underlain by Water." JPT, Oct. 1979, pp. 2233-2245. Ivory, 1., Derocco, M., and Scott, K. 1991. "Comprehensive Analysis of the Mechanisms by which Air Improves Bitumen Recovery in Steam Injection Processes." Paper presented at CIM Conference, Banff, AB, Apr. 1991, CIM 91-106. Jenkins, G.R., and Kirkpatrick, J.W. 1979. "Twenty Years' Operation of an In situ Combustion Project." JCPT, Jan-Mar. 1979, pp. 60-65. Kular, G.S., and Chinna, H. 1988. "Multiple Hydraulic Fracture Propagation in Oil Sands." Paper presented at SPE Rocky Mountain Regional Meeting, Casper, WY, May 1988. Kular, G.S., Lowe, K., and Coonibie, D. 1989. "Foam Application in an Oil Sands Steam Flood Process." Paper presented at 64th Annual SPE Technical Conference and Exhibition, San Antonio, TX, Oct. 1989, SPE 19690. Mainland, G.G., and Lo, H.Y. 1983. "Technology Basis for Commercial In situ Recovery of Cold Lake Bitumen." Paper presented at II th World Petroleum Congress, London, UK, Aug. 1983. Marx, 1.W., and Langenheim, R.W. 1959. "Reservoir Heating by Fluid Injection." Trans., AIME. Vol. 216, p. 312. Mehra, R.K. 1991. "Performance Analysis of In situ Combustion Pilot Project." Paper presented at SPE International Thermal Operations Symposium, Bakersfield, CA, Feb. 1991, SPE 21537. Meldau, R.F., Shipley, R.G., and Coats, K.H. 1981. "Cyclic Gas/Steam Stimulation of Heavy-Oil Wells." JPT, Oct. 1981, pp. 1990. Metwally, M. 1991. "Recovery Mechanisms: Fireflooding a High-Gravity Crude in a Waterflood Sandstone Reservoir, Countess Field, Alberta." Paper presented at SPE International Thermal Operations Symposium, Bakersfield, CA, Feb. 1991, SPE 21536.
Myhill, N.A., and Stegeimeier, G.L. 1978. "Steam Drive Correlation and Prediction." JPT, Feb. 1978, pp. 173-182. Oglesby, K.D., Belvins, T.R., Rogers, E.E., and Johnson, W.M. 1982. "Status of the l O-Pattem Steamflood, Kern River Field, CA." JPT, Oct. 1982, p. 2251. Patzek, T.W. 1988. "Kern River Steam Foam Pilots." Paper presented at SPEIDOE EOR Symposium, Tulsa, OK, Apr. 1988. Ploeg, J.F., and Duerkson, J.H. 1985. "Two Successful Steam/Foam Field Tests, Section 15A and 26C, Midway Sunset Field." Paper presented at SPE California Regional Meeting, Bakersfield, CA, Mar. 1985, SPE 13609. Prats, M.A. 1969. "The Heat Efficiency of Thermal Recovery Processes." JPT, Mar. 1969, pp. 323332. - - - . 1978. "Current Appraisal of Thermal Recovery." JPT, Aug. 1978, pp. 63-69.
-----.1986. ThermalRecovery.SPE Monograph, Vol. 7, pp. 43-50. Pursley, SA 1974. "Experimental Studies of Thermal Recovery Processes." Paper presented at Heavy Oil Symposium, Maracaibo, Venezuela, Jul. 1974. Romney, G.A., Wong, A., and McKibbon, 1.H. 1991. "A Preview of Ari Electromagnetic Heating Project." Paper presented at CIM Annual Meeting, Banff, AB, Mar. 1991, CIM 91-109. Sander, P.R. 1991 "Steam - Foam Diversion Process Development to Overcome Steam Override in Athabasca." Paper presented at Annual SPE Conference, Dallas, TX, Oct. 1991. Shepherd, D.W. 1979. "Predicting Bitumen Recovery from Steam Stimulation." World Oil, Sep. 1979, pp.68-72. Stokes, D.D. 1978. "Steam Drive as a Supplemental Recovery Process in an Intermediate Viscosity Reservoir, Mount Poso Field, CA." JPT, Jan. 1978, pp. 125-131. Suffridge, F.E. 1991. "Foam Performance Under Reservoir Conditions." Paper presented at SPE Annual Conference, Dallas, TX, Oct. 1991. Towson, D., and Khallad, A. 1991. "The PCEJ Steam Stimulation Project." Paper presented at the CIM! AOSTRA Technical Conference in Banff, AB, Apr. 1991, CIM/AOSTRA 91-108.
198
g
ENHANCED RECOVERY BY THERMAL STIMULATION
Van Dijk, C. 1968. "Steam Drive Project in the Schoone-beck Field, The Netherlands." JPT, Mar. 1968,pp.295-302. Vogel, J.V. 1982. "Simplified Heat Calculations for Steam Floods." Paper presented at 57th Annual Fall SPE Technical Conference, New Orleans, LA, SPE 11219. Waxman, M.H., Closmann, PJ., and Deeds, C.T. 1980. "Peace River Tar Flow Experiments Under In Situ Conditions." Paper presented at 55th SPEAIME Annual Fall Technical Conference and Exhibition, Dallas, TX, Sep. 1980, SPE 951 I.
Williams, R.L., Brown, S.L., and Ramey, HJ. Jr. 1980. "Economic Appraisal of Thermal Drive Projects - A New Approach." Paper presented at SPE Technical Conference and Exhibition, Dallas, TX, Sep. 1980, SPE 9358. Willhite, G.P. 1966. "Overall Heat Transfer Coefficients in Steam and Hot Water Injection Wells." Paper presented at Rocky Mountain SPE Regional Meeting, Denver, CO, May 1966, SPE 1449.
199
Chapter 15
ENHANCED RECOVERY BY CARBON DIOXIDE FLOODING
15.1
INTRODUCTION
Carbon dioxide flooding, in both the miscible and immiscible modes, is one of the most widely used enhanced oil recovery techniques today. There are over forty carbon dioxide floods in operation throughout the world. In Canada, several pilot and experimental floods have been tried or are currently in operation. In addition, single-well "huffand puff' stimulations have been tried in various fields. Carbon dioxide flooding has now been proven in both the laboratory and the field as a viable technology when applied to selected reservoirs. Carbon dioxide flooding may be in miscible, nearmiscible or immiscible modes and may be implemented before, in combination with, or post-waterflood. Completely miscible (low tension) processes are usually considered those in which recoveries ofgreater than 90 percent occur in slim tube tests and in which there is no visible two-phase flow in lab tests. Carbon dioxide (C02) is a very powerful vapourizer of hydrocarbons and, as a dense-state gas, it possesses a dissolving power for light to intermediate petroleum fractions that is superior to hydrocarbon, nitrogen or flue gases. This dissolving power can be utilized for in situ fractionation of oil to develop high concentration banks of light and intermediate components that have high displacement efficiencies (up to 95 percent) and lower minimum miscibility pressures (MMP). Miscible carbon dioxide floods may also recover oil beyond lowtension effects because ofthe extraction of components from nonmobile oil in heterogenous rock. Immiscible carbon dioxide gas drives are useful for both oil and condensate reservoirs because of the effects of swelling, viscosity reduction, vapourization, and efficient gravity drainage. Medium heavy oils that may not waterflood well and that have high intermediate fractions may also be candidates for immiscible flooding. Some evidence also exists that oil- carbon dioxide mixtures may improve waterflood behaviour by resulting in phases that are rich in resins and asphaltenes. These may stabilize fines and clays and alter wettability.
The following are crucial for determining reserves and evaluating a carbon dioxide flood: • The availability and cost of the CO 2 supply • The classification of the process as miscible or immiscible for recovery purposes • The efficiency ofthe process in terms ofunits ofCO2 injected for each incremental unit of oil recovered (utilization rate) In Canada, the use of carbon dioxide has been limited by the location, size, and development and transportation costs ofthe CO 2 supplies. The use of hydrocarbon light ends for miscible floods has been preferred in the past because of low prices, proximity, oversupply, and government incentives.
15.2
PROCESS REVIEW
The three classifications of carbon dioxide floods are miscible (including near-miscible), immiscible, and carbonated waterfloods. The latter are not currently of interest. Miscible processes are the most common and are characterized by phase behaviour effects that cause a stable miscible bank with microscopic displacement efficiencies near 100 percent. In comparison with waterflooding, this increase in displacement efficiency more than. offsets the adverse mobility ratios between the CO 2 and the oil, especially if gravity effects, alternating water gas injection, or horizontal wells can be used to advantage. Miscible and near-miscible processes are typically implemented in reservoirs containing oils with API gravities greater than 27, with reservoir temperatures less than 105°C (220°F) and pressures greater than 9650 kPa (1400 psi). Miscibility pressures decrease as t?e C2-C I 0 fraction of the oil increases, and increase With decreasing API oil gravity and with reservoir temperature. Miscibility pressures typically range from 26 200 to 9650 kPa (3800 to 1400 psi) as API gravity increases.
200
c
ENHANCED RECOVERY BY CARBON DIOXIDE FLOODING
Miscibility with oils having API gravities less than 27 have also been reported. For these low-gravityoils, estimates of MMP become scattered, but range from a low ofl4 620 kPa (2120 psi) to over 27600 kPa (4000 pSI). If carbon dioxide is available, it is often the "solvent of choice" for miscible flooding because it is a powerful extractor of intermediate components from crudeoil and can lead to a reductionin MMP of as much as 6900 kPa (1000 psi). Displacement efficiencies for the miscible process in laboratory core floods with connate water saturations are over 95 percent of the original oil in place (OOlP). Lab tests on water- flooded cores may recover 70 percent of the residual oil in place. In field applications implemented before waterflooding, overall oil recovery factors typically vary between 45 and 65 percent for horizontal floods and 55 and 90 percent for vertically directed gravity stable floods. For miscible CO2 flooding in the tertiary mode, from 20 to 30 percentofthe residual oil to waterfloodmaybe recovered in horizontal floods. In vertically directed floods, the presence of water may inhibit fingering and aid the areal spread of CO2 , resulting in recoveries of 40 to 70 percent ofthe residual oil. Immiscible processes are generally less favoured than miscible processes where a choice is possible. In the immiscible version of the process, mass exchange between the oil and the injected CO2 , while not sufficient to cause a 100 percent flush of oil, may result in displacement efficiencies that are higher than either waterflood or inert gas flood. However, as a rule of thumb,immiscibleprocessesare chosenfor lower gravity oils in the 18·24 API range at temperatures where swelling and viscosity reductionare considered themain recovery mechanisms.
15.3
RECOVERY MECHANISMS
The following mechanisms contribute to enhanced recoveryby the use of carbon dioxide flooding: Viscosity Reduction, which improves the flow characteristics of the oil and improves the mobility ratio in the flood Swelling, which reducesthe amount of stocktank oil in the residual oil saturationand may improvethe relative permeabilityto oil Reduction in Interfacial Tension (1FT), which allowsthe oil to be released from the rock; in a miscible flood, the 1FT is reduced to below 0.1 dynes/em, allowing displacement efficiencies of over 90 percent, but
significant reductions in 1FT can also occur in immiscible CO2 floods ~xtractio~/Vapourization, which is especially ~mporta~t In
oils that have high percentages of
intermediate componentsthat can be extracted into the
CO2 phase; the amount of 1FTreduction that occurs is increased and the MMP is lowered; extraction also allows the recovery of a portion of the nonswept oil Dissolved Gas Drive, in which the dissolved CO will help recoveries in the blowdown phase of the flood Injectivity improvements can also occur because of removal of oil saturation around the wellbore and because of interaction between the carbon dioxide and the rock.
15.4
DESIGN CONSIDERATIONS
15.4.1 Phase Behaviour The recoveryof oil by carbon dioxidefloodingis highly dependent upon the phase behavior between carbon dioxide, water and oil. The phase behaviour strongly affectsfluid flow by altering mobility ratios, interfacial tensions, relative permeabilitity, and rates of mass transfer mixing. Carbondioxide in the dense gas state is a verypowerful dissolverfor lightand intermediate petroleumfractions. The extraction and concentration of these fractions is highly pressure-dependent and causes the formation of a stable miscible or near-miscible bank. Typically, the pressures required for the MMP are 10 350 to 13 800 kPa (1500 to 2000 psi) lower than for a methane highpressure gas drive. In reservoirs with lowerpressuresand temperatures, the process is more complex as more phases develop. Miscibilitymay not occur, but there will be significant benefits due to a reduction in 1FT and viscosity, and swelling and solution gas effects.
15.4.2 Displacement Efficiency The estimation of the microscopic sweep for gas or solventdrives in reservoirswith low or immobilewater saturations is usually based on measured or simulated oil recoveries that are obtained from multiple contact displacement tests in composite cores or tubes packed with sand (slim tubes). In reservoirs that have been previously waterflooded, or where connate water saturations are mobile, corefloods maybe required to ensure that the oil is not shielded from the CO2 in high water saturationzones. The choice of an optimum flooding pressure or solvent composition is usually estimated from correlations 201
..
'ilt' DETERMINATION OF OILAND GASRESERVES
based on a combination of calculated and measured laboratory data. In floods where shielding does occur, optimum operating pressures may be lower than MMPs measured by slim tubes. For CO2 floods, the decisions should also take into account questions such as possible reduction of injectivity by precipitation of heavy ends and potential flow interference effects that could benefit the sweep efficiency in partially miscible floods, as well as the presence or absence of mobile water or gas saturations.
15.4.3 Volumetric Sweep Efficiency Miscible processes, including CO2 floods, unfortunately can suffer from poor volumetric sweep efficiencies as a result of the high mobilities of the low viscosity solvents (less than 0.1 mPa.s) and chase gases. Unfavourable mobility ratios coupled with reservoir heterogeneities can be disastrous to miscible flood processes that rely on maintaining the integrity of small slugs of solvent during the course ofthe flood, not only because low volumetric sweep efficiencies may result, but also because fingering may cause premature dissipation of the slug and result in greatly diminished displacement efficiency between the (immiscible) chase gas or water and the reservoir oil. Techniques used to improve the volumetric sweep efficiency of miscible floods include alternate gaswater injection (WAG), pre solvent water injection to reduce permeability contrasts, infill drilling to alter patterns, and blocking and diverting agents. Various short-cut methods of estimating volumetric sweep efficiency may be used by considering areal and vertical sweep efficiencies separately. A final design will require more sophisticated numerical models. The estimation of sweep efficiency considerations for CO2 floods is similar to that for other floods. Areal sweep efficiency is a function of the mobility ratio (relative permeability, viscosity ratio), permeability trends, saturation distributions, well pattern effects, solvent throughput, and production rates. In vertically directed floods, areal sweep is also affected by density ratios between carbon dioxide, oil, gas, and water. Vertical sweep efficiency is often a result of stratification in the reservoir rock in a direction parallel to the main flow direction. The strata are swept in order of descending permeability sequence, with the lowest permeability being unswept at the project termination. Other causes of poor vertical sweep include gravity override or underride in reservoirs with little or no stratification.
15.4.4 Slug Sizing Carbon dioxide floods may be operated with essentially horizontal displacement or, in high dip or reef reservoirs, with gravity stabilization. In either case, the process usually entails the injection of a slug of CO followed by or co-injected with water or flue gas. Th~ volumes of CO 2 necessary for a particular application depend on the level of gravity stabilization during displacement. For gravity stable floods, the slugs of CO 2 range in size from 10 to 20 percent hydrocarbon pore volume (HPV). In horizontal floods, the slug sizes may range from 20 to 60 percent HPV depending on factors such as water saturation, heterogeneity,well patterns and spacing. A typical formation volume factor for CO2 at 20 690 kPa (3000 psi) and 60 °C (140 OF) is 266 m3/res. m3 (1500 scfper reservoir barrel).
15.5
RESERVE EVALUATION
In reserves evaluation the following are important considerations with respect to carbon dioxide flooding: I. The availability and cost of the supply must be evaluated. Carbon dioxide is available from natural sources, from fertilizer plants, and as a combustion by-product (such as from electric power generation plants). The use of carbon dioxide in Canada has been limited by the location, size, and development and transportation costs of the CO2 supplies. It is important to consider the following when evaluating a supply: • The maximum available rates and total volumes. • Contract terms: length, price escalators, performance clauses, and royalty payments. • The reliability ofsupply and availability ofbackup or alternative volumes. • Purity: nitrogen and methane will raise the MMP, and propane, butanes and H2S will lower it. Combustion by-products like oxides may have to be removed to avoid corrosion. • The capital required to develop the source of carbon dioxide. Dehydration and compression will be needed for a raw source, and the removal of combustion products is expensive. • Transportation costs: These are a limiting factor, either as pipeline length (capital cost) or as trucking costs. Generally, trucking is practical only for small pilots, one-well huff-and-puff, small slugs such as a small vertical scheme, or as a short-term supplement to lower cost supplies.
202
c
ENHANCED RECOVERY BYCARBON DIOXIDE FLOODING
2. The viability and the estimation of economically recoverable reserves for a CO 2 flood depend on the combinations of the following relationships: • The cost of CO 2 vs. the netback price received for the oil • The CO 2 utilization or flood efficiency, i.e., the amount of CO2 injected to recover each incremental unit of oil-estimated values range from 530 to 2670 mJ/mJ (3 to 15 mcflbbl) • The incremental production rates-the amount and timing of the oil production vs. the capital, the operating and injection costs, and the timing The recovery efficiency depends on the following: • The oil composition and type-these affect miscibility and the recovery mechanisms • The stage of the flood-the CO 2 injection can occur before, during or after the waterflood • The normal factors that affect all floods, e.g., water saturations and geology Some particular situations that may cause problems with CO2 floods are reservoirs with large gas caps and water legs, depleted pools, zones with high permeability streaks, and low permeability reservoirs with lower gravity asphaltic crudes. 3. Incremental production rates are more difficult to accurately forecast than recovery factors for many EOR projects and, especially, early in the life ofthe project. The base production rates may be affected by such factors as wellbore problems, injection rates, permeability streaks, break-through, recompletions, and infill drilling. Computor simulators improve the ability to handle all the variables, but may not significantly improve the accuracy of the forecasts. Good lab test results should be used in the simulators to help define the effects the recovery processes will have on the incremental production rates as well as on the overall recovery factors. 4. To obtain the highest recovery efficiency, it is important to provide the maximum contact between the oil and the CO 2 (both timewise and areally). The earlier in the life of the pool that the CO 2 can be injected, the higher the target oil saturation will be and the lower the potential water blockage. Provisions to increase the conformance, such as alternate gas-water injection and diverting agents, may be necessary. Injectors and producers should be equipped, if possible, to shut off high permeability zones, and
producers should be able to handle sporadic slugs of gas (high and low gas-oil ratios). 5. Gas or miscible floods such as CO 2 floods can be subject to early break-through, so provisions should be made for the separation and re-injection of the produced or break-through CO 2 , The re-injection may also reduce the overall CO 2 requirements, especially if the pool is being flooded in stages, and it will provide the maximum contact time with the oil over the life of the flood. 6. Corrosion is a major problem in CO 2 floods. In producers, the CO 2 can make metal water-wet and accelerate corrosion by stripping off the protective film ofoil. Also, water and CO2 form carbonic acid, which is corrosive. Chemical inhibitors and coated tubing should be used. Injected CO 2 should be handled in a dry state as much as possible and if a WAG (alternating water and gas injection) is being used, an alcohol slug should be used between the water and the CO 2 to clean up and dry out the injection lines and tubing. 7. Carbon dioxide flooding can cause asphaltenes to precipitate from the crude oil and result in plugging in the formation, downhole equipment and surface treating facilities. This problem would require a flush-squeeze treatment with an aromatic solvent such as toluene to restore production or injection. In some floods, calcium carbonate plugging at the high water cut production wells is a problem. This can be treated with acid jobs and the injection of scale inhibitors. 8. Because carbon dioxide is a "greenhouse gas," possible goverrunent incentives may improve the viability of a project.
15.6
FIELD APPLICATIONS
More than forty miscible and immiscible CO 2 pilot, experimental, and mature field applications are in operation worldwide. Several noteworthy ones are described here. The Wertz Tensleep Miscible CO 2 Project This project was undertaken in a reservoir in Wyoming that had previously been waterflooded to 45 percent of OOIP. The recovery of an estimated additional 10 percent incremental oorp (or 22 percent of remaining oil in place) has been attributed to the injection of CO2 and water to repressure to above MMP, and the drilling ofnew injectors and producers at key locations. Carbon dioxide utilization is estimated at 2500 mJ/m3 (14 mcflbbl). 203
DETERMINATION OF OIL ANDGAS RESERVES
The SACROC Miscible CO 2 Flood This project in Texas is one of the earliest and largest applications of miscible CO 2 flooding in the world. Despite many pioneering difficulties, including controversy regarding the MMP, this flood continues. It is expected to yield incremental recoveries of 7.5 percent OOIP in selected sections of the pool with CO2 utilization of 1780 m3/m3 (10 mcflbbl) of incremental oil. The Lick Creek Meakin Sand Immiscible CO 2 Flood This immiscible version of the process has used a combination of cyclic stimulation, continuous CO 2 injection, alternating water and CO2, and continuous water injection to recover the 160 mPa.s reservoir oil. This project in Arkansas is currently working well and is anticipated to yield an incremental recovery of 13 percent OOIP with CO2 utilization of roughly 1780 m3/m3 (10 mcflbbl). The Hansford Marmaton CO 2 Flood This project was initiated in an immiscible mode in a pressure-depleted reservoir containing a secondary gas cap. Recovery from primary was estimated at 13 percent OOIP. After the reservoir was repressured,
204
miscibility was developed, and a further 9 percent of OOIP was recovered during an 8-year period with an estimated utilization of 1246 to 1780 ml/m3 (7 to 10 mcflbbl). The literature contains textbooks and papers that contribute to the understanding ofcarbon dioxide flooding (Holm, 1982; Mungan, 1981, 1982; Stalkup, 1978; Klins, 1984).
References Holm, L.W. 1982. "C02 Flooding: Its Time Has Come." JPT. Dec. 1982, pp. 2739-2745. Klins. M.A. 1984. Carbon Dioxide Flooding - Basic Mechanisms and Project Design. International Human Resource Development Corporation, Boston, MA. Mungan, N. 1981. "Carbon Dioxide FloodingFundamentals." JCPT. Jan.-Mar. 1981, pp. 87-92. - - - . 1982. "Carbon Dioxide Flooding Applications." JCPT. Nov.-Dec. 1982, pp. 112-117. Stalkup, F.r. 1989. "Carbon Dioxide Flooding: Past, Present, and Outlook for the Future." JPT. Aug. 1978, pp. 1102-1112.
Chapter 16
RESERVES ESTIMATION FOR HORIZONTAL WELLS
16.1
INTRODUCTION
~orizontal
wells provide an alternative way of draining . 011 and gas from a pool. They allow drainage from a larger reservoir volume (than vertical wells in the same setting), along with production at increased rates or reduced pressure drawdown. Various performance analyses and theoretical studies have shown that in certain situations, horizontal wells can yield significantly higher (more than three times) oil rates and reserves than vertical wells; however, they also entail higher drilling, completion, and workover costs. Although to date, the technical and economic success of horizontal wells has ranged from spectacular to very disappointing, there is a growing consensus about their potential to provide significant additions to the world's oil and gas reserves (up to 2 percent of the initial in-place volumes). The most popular uses of horizontal wells have been in offshore operations, pools that are prone to coning, naturally fractured reservoirs, medium- to heavygravity pools, low productivity pools, and waterflood or enhanced oil recovery. In many cases, in addition to an increase in the drainage area, the recovery factors are also improved. From a recent study of Canadian horizontal wells, it has been concluded that the profitability of horizontal wells is directly linked to the reserves drained. The increased production rate helps to offset the increased cost of placing the horizontal wells (Bowers and Bielecki, 1993). Other factors, such as heterogeneities, damage, and lateral pressure drops within the well, may retard drainage, and offset the advantages mentioned. Thus, drainage hydrodynamics (within the reservoir, and especially in and around the well) have an important influence on the reserves. The hydrodynamics around a horizontal well, in turn, depend upon the geological features and dominant production mechanisms. The hydrodynamics also depend upon operationally induced features such as prevailing pressure and saturation distributions due
to prior depletion, damage, well length, undulating well trajectory, diameter, and flow rate. The interactions between these factors are extremely complex and not fully understood at the present time. It may be fair to say that theoretical developments regarding anticipated production declines under various real life reservoir settings, production mechanisms, and completion conditions are still in their infancy. In addition, industry's database in terms of performance history cost-effective trouble-shooting, and success rates for remedial measures is extremely limited despite the fact that in early 1993 nearly 5000 horizontal wells were producing worldwide, including more than 1000 in Canada. The net effect of these problems is to lower confidence in reserves estimates for horizontal wells (as compared to vertical wells), whether they are based on volumetric determinations, performance, analogies, correlations, or simulation. The challenge is not only to come up with independent corroboration of reserves estimates, but also to quantify uncertainty. An ideal procedure would be to project performance to the economic limit and verify reserves by volumetric determination. However, sufficient data may not always be available to accomplish both of these to the desired level of confidence. The volumetric method involves determination of the range ofareas and volumes drained by a horizontal well and recovery factors. The drainage volume would depend upon the length, orientation and location of the well; production mechanism; stratification; and fractures. The recovery factors would depend upon the co~pletion parameters, prior depletion, nature of operations, and reservoir variability. In practice, even after placement of a horizontal well, many of the parameters involved may not be known to the desired accuracy. The same would be true for the other methods of reserves determination. Besides, various diagnostic and remedial measures for poorer-than-expected performance are slowly being evolved. As experience is gained, they are gradually improving, but there are stilI 205
DETERMINATION OF OIL AND GASRESERVES
significant uncertainties in reserves determination. A procedure would therefore have to be essentially iterative to incorporate reasonable and consistent estimates of various parameters and their implications on drainage. The evaluator would require good geological and hydrodynamic models of the drainage volumes of a horizontal well. One way to quantify the range of uncertainties on production projections and reserves would be to use a detailed Monte Carlo computer simulation (Springer et aI., 1991). This, in tum, requires prior knowledge of statistical distribution of various input parameters. The drainage to a horizontal well could be improved by certain geological features (e.g., fractures) and impeded by others (e.g., stratification, previously depleted regions, and damage). Therefore, detailed geological and hydrodynamic models for the drainage area of a horizontal well are essential for understanding and quantifying production performance. Interpretation of logs and cores, well tests, or pressure data for the horizontal well and any offsetting wells would assist in the preparation of these models. The examination of flow distribution within and around a wellbore (as is done during the design of horizontal wells) is of great importance. Significant implications to reserves could be due to vertical location, stratification, orientation, undulations, prior depletion, effectiveness of completions, formation damage, and lateral pressure drops within the well. The overall depletion mechanism or the nature of the production decline is not altered by the use of a horizontal well. However, some changes to decline rates may occur over time due to the effects ofchanging flow regimes, heterogeneities, cross-flow, and interference from different boundaries ofthe drainage area. The use of smaller pressure drawdown (i.e., a coning situation) or increased flow rates may help to prolong the economic life and hence the reserves in some situations. These may also be helped by gravity drainage to the horizontal wells. At low pressure drawdown, gravity may be contributing significantly to the production from horizontal wells. The impact on recovery of regulations concerning horizontal wells may be hard to quantify. Depletion strategy and economic reserves may change due to factors such as allowables, spacing, offset distances, and royalty regulations, so these must all be considered in the determination of reserves. Due to higher initial productivities of horizontal wells, production curtailment or fiscal (royalty, tax) relief during their early
life would have significant impact on the Overall economics. In situations of marginal economics, incen_ tives could have a major impact on probable reserves. Also, the role of horizontal wells in the overall deple_ tion strategy for the pool must be defined prior to reserves determination. In view of the uncertainties, reserves determination would involve several iterations to ensure consistency, followed by a quantification of confidence levels (Springer et aI., 1991).
16.2
RESERVES DETERMINATION TECHNIQUES
16.2.1 Performance Projection Horizontal wells, as already mentioned, mainly provide increased access to the reservoir. Placement of a horizontal well by itself does not change the basic reservoir mechanism or the type of decline to be expected, although some variations could occur. Producibility and declines for horizontal wells depend upon the nature of the reservoir, the state of depletion, and the dominant production mechanisms. Theoretical discussions are available for only a few idealized horizontal well systems. Using these as guides, it is possible to project the behaviour of horizontal wells. Usually, the performance of a vertical well provides important clues to the performance ofa horizontal well in the same setting. Several methods for determining rates under steady-state conditions have been proposed. Ofthese, Joshi's method is the most widely used (Mutalik and Joshi, 1992). Oil rate, qh' in barrels per day is expressed as:* 0.007078 khh Llp J.l,B,
(1)
where kh = horizontal permeability (mD) h = net pay thickness (ft) Llp = pressure drop (psi) u, = viscosity of oil (cp)
*In metric units, the constant is 542.9 and the units are as follows: permeability, J.lm': pressure, MPa: flow rate, mJ/d.
206
c
RESERVES ESTIMATION FOR HORIZONTAL WEllS
B, = formation volume factor (res. bbl/stb) a = (L/2){O.5 + [0.25 + (2r'h/L)4]O.5}O.5 r,h = the drainage radius for the horizontal well (ft) L = length of the horizontal well (ft) 13 = anisotropy = -VkH/kv rw = well radius (ft) It may be noted that the equation is valid only for singlephase flow and uses single values for various input parameters. The value of drainage distance, r'h' for a horizontal well may not be known a priori. As a first approximation, the drainage distance, r,v' for vertical wells could be used for r,h' For horizontal wells in reservoirs under solution gas drive, producibility under unsteady and semi-steady conditions has been projected by Poon (1990), Mutalik and Joshi (1992), Babu and Odeh (1989), and others. Poon's analysis uses an analogy between horizontal wells and vertical fractures for projecting performance. It is particularly useful since it provides "type curves" for certain idealized conditions. For other situations, flow equations could be combined with material balance and the semi-steady state treated as a succession ofsteady states. The procedure would involve alternately obtaining estimates of average reservoir pressure (material balance) and flow rates (steady state) for different periods until the economic limit was reached. It must be kept in mind that, in some situations, uncertainties in many ofthe parameters may render these projections of little practical value. Another approach could be to use Babu's method for projecting performance and study various sensitivities to evaluate the impact of uncertainties. In coning and cresting situations, operations would be discontinued at certain minimum oil rates or at certain water cuts or gas-oil ratios. The latter parameters may be based upon safety, equipment, economic or regulatory considerations. Theoretically, cresting can be avoided by producing below certain critical rates (Freeborn et aI., 1990), which themselves may change with the changing pressures or fluid levels. Chaperon (1986) presented an approximate method for computing critical rates for horizontal wells. This method is generally accepted and used by the industry. Critical rates for horizontal wells are usually much higher than for vertical wells. In practice, only a few kinds ofreservoirs can produce "clean" oil or gas for an extended period. These include gas pools under active water drive or some offshore operations with limited
platform space that do not permit installation of equipment to handle large volumes ofwater or gas production. In these cases, oil or gas reserves would be those obtained prior to significant break-through. Breakthrough may be delayed by operating at sub-critical rates. This would involve continuously altering rates with changing fluid contacts until the rates become uneconomic. In other cases where facilities are not major constraints, large gas-oil ratio or water cut may result in an uneconomic oil rate. The nondrained part of the oil column is known as the "cresting loss" or, in the case of both bottom water and gas cap, as the "sandwich loss." These can be estimated from the design features for.a horizontal well, as well as from operational and reservoir parameters (Chaperon, 1986; Joshi, 1991). It is generally recognized that horizontal wells could significantly reduce these losses (by 20 to 40 percent). Most often, the bulk of oil production would occur under increasing water cuts or gas-oil ratios or both. Under these conditions, reserves would again be the sum of oil drained by the mean change of fluid contacts in the drainage area (ignoring the effects of the crest) and the volume of mobile oil within the crest. Correlations are available to estimate the time for the crest to break through at the horizontal well (Papatzcos et aI., 1991; Yang and Wattenburger, 1991). Estimates of breakthrough time would help in estimating the amount of clean oil production. Oil cuts would harmonically decline thereafter (until interference from offsetting wells was experienced), yielding a straight line on a semi-log plot ofoil cut vs, cumulative oil. For passive water drive cases, reserves would essentially be due to fluid expansion and drainage of the movable oil within the crest. The latter can be estimated by a method suggested by Butler (1989). He suggested it would be equal to movable oil within half a cylinder between the horizontal well and the fluid contact. * For an anisotropic reservoir, this would be modified to a half ellipsoid (Figure 16.2-1). The distance between the interface and the well is called "stand-off', h. This would be the vertical axis of the ellipsoid, and the horizontal axis would be given by the expression h(k H!kv)o.5. For an undulating well or a tilted fluid contact, the minimum distance between fluid contacts and well trajectory would be the effective stand-off. Similarily, ifthe lateral pressure drop caused the rates to exceed the critical in some parts of the well, localized cresting would tend to reduce reserves for the entire well. In such situations, if heterogeneities could
* Butlersubsequently published more sophisticated theoretical models. 207
DETERMINATION OFOILAND GASRESERVES
completions are not available to fully assess the reasons for these increments. Viscous fingering, heterogeneities or hydrodynamics within and around horizontal welIs promoting water or gas channelling could be some of the causes resulting in poorer recoveries.
Pi
(a)
I I I I I
:
oil
,
~rev-
,
t
Once the performance after break-through can be projected, a summation of oil production will provide estimates for reserves.
r
(b)
., h
I·
L~
Source: Joshi, 1991.
Figure 16.2-1 Schematic of Horizontal and Vertical Well Drainage Areas*
be adequately characterized, detailed numerical modelling might be the only way ofobtaining reliable reserves estimates under different completion and operating conditions. For optimizing reserves, it may be necessary to ascertain that the flow along a horizontal well is evenly distributed. At this time, no methods other than correlations (Yang and Wattenburger, 1991) are available in the public domain for estimating post-break-through production of oil and water (or gas) via a horizontal well. As a first approximation, coning correlations of Kuo (1989) for the vertical wells or Butler's method for horizontal wells (Butler and Suprunowicz, 1992) may be used. Computer-generated projections for the Suffield Jenner pool in Alberta appear more optimistic than these correlations. The actual decline ofoil cuts with cumulative oil was not unlike that for a vertical well after allowances were made for increased drainage area due to length, and reductions in crest volume due to heterogeneities (Russell and Espiritu, 1992). For some horizontal wells in the Provost Dina pools ofAlberta (Heysel, 1992) very modest increments over vertical wells have been reported. However, data on well trajectories and • It can be assumed that drainage distances for vertical
wells (r,,) and horizontal wells (r'h - U2) are equal. However, experience with partiallydepleted Canadian pools indicates that r" could be larger than r,h - U2.
Whereas horizontal wells have proven to be effective in minimizing water production, their effectiveness in controlling gas cresting has only provided mixed results. If gas cresting is a limiting factor, usualIy the reserves are . much lower than the method as described would indicate. The reasons could be a sharp drop in effective oil permeability at high gas saturations or viscous fingering as the result ofunfavourable mobility ofoil compared to that of gas. The foregoing discussion pertains to the improved reservoir drainage by horizontal wells under solution gas drive and water and gas coning situations. Horizontal welIs can also significantly improve reserves drained from waterfloods as well as thermal and nonthermal enhanced oil recovery. The improvement could be the result of increased access, injectivity or productivity, and increased volumetric sweep efficiencies. However, fractures or previously drained regions could seriously limit the incremental reserves. Careful engineering of horizontal well length, orientation, vertical placement, and operation is needed to obtain optimal reserves under these conditions. As in the case of primary production, the key factors controlling the reserves would be the hydrodynamics within the drainage region and the economics. The role of reservoir variability must be taken into account in all situations. Sufficient details on certain heterogeneities may not be known, even after a horizontal well starts producing. Due to this variability, the performance of horizontal wells tends to be site-specific. Another consequence is the difficulty in identifying the "average" reservoir parameters. At this time, in terms of length of performance history and available geological and operational details, industry's database is extremely limited for use in deriving meaningful analogies and correlations. Well test data and performance histories, besides confirming production mechanisms, can help to quantify certain reserves parameters. Otherwise, they do not seem to be definitive enough for reserve estimation. In a few cases where the data are available for a long enough duration to be definitive, the decline curve and material balance
208 .t,~:
RESERVES ESTIMATION FOR HORIZONTAL WELLS
methodologies for conventional wells could be extended for horizontal wells. Generally, the most fruitful techniques for reserves in vertical wells would also be applicable to horizontal wells. A methodology for horizontal wells is suggested in Section 16.3.3.
16.2.2 Volumetric Method Detailed flow distribution around a well is the most important consideration in identifying the drainage area for a horizontal well, which would drain a much larger portion of a reservoir than a vertical well, depending upon its length. Other factors determining drainage area would be the distance to the nearest pool boundaries and the distance to offsetting wells as well as the rate of drainage by them. For homogeneous reservoirs under solution gas drive, Joshi (1991) has presented methods for estimating drainage areas based upon estimating the time to reach semi-steady state for different drainage geometries. From these, effective drainage area can be estimated. Limited experience to date suggests that drainage distance for horizontal wells (reh - Ll2 in Figure 16.2-1) would, in many cases, be smaller than that for vertical wells (rev)' The reasons could be heterogeneities and prior depletion. As a rule of thumb, a 300 m well would drain the equivalent oftwo vertical wells, and a 600 m well-the equivalent of three vertical wells. However, this rule of thumb must be used with extreme caution. It has been observed from the performance of several Canadian oil wells that the reserves for sandstone pools are generally proportional to their lengths (Bowers and Bielecki, 1933). Corresponding correlations between well lengths and reserves drained for fractured carbonate pools are rather weak. It is possible that this is caused by water influx via some of the relatively larger fractures. By and large, horizontal wells in Estevan light oil pools were draining 250 to 300 m in the lateral direction whereas for Lloydminster heavy oil, this distance is less than ISO m and could be as low as 50 to 70 m (Springer and Flach, 1993). In some Alberta light oil pools, very disappointing reserves were noted (Bowers and Bielecki, 1993), implying small drainage areas or poor recovery factors.
16.2.3 Role of Heterogeneities In a heterogeneous reservoir, a horizontal well is likely to traverse many more prolific regions than a vertical well. For a given pressure drawdown, most of the inflow would be from these more prolific regions. Thus,
one horizontal well would be equivalent to several individual vertical wells placed in the path ofthe horizontal well. The increased producibility as well as the increased reserves would be similar to those expected for closely spaced vertical infill wells. The accelerated drainage may induce faster declines (as well as interference with offsetting wells). Extreme examples of such prolific zones are fractured regions in Austin Chalk in Texas, Bakken Shale in North Dakota, and karstic regions in the Raspo Mare oil field off the Italian coast in the Adriatic Sea. Variable fracture or vug density in the dolomitic reefs of Alberta and Saskatchewan may also constitute prolific regions ("sweet spots"), but with less dramatic impact on reserves. On the other hand, these sweet spots may also act as pathways for water or gas to break through at the wells and thus reduce volumetric sweep and recovery factors. The vertical and lateral extent of the drained region would mainly depend upon geological features such as stratification, fractures, barriers to flow, and lateral variations. Effective drainage volume for a horizontal well would thus be smaller than the hydrocarbon pore volumes contained within the drainage area if these exist. In order to identify the drainage volume of a horizontal well, a geological model would be very helpful. It may be noted that even in pools with good geological control, horizontal wells usually reveal unanticipated features. A geological model, updated with data from horizontal wells, would greatly aid in determining drainage volume for the well.
16.2.4 Importance of Channelling in Reserves Performance In certain geological settings, it becomes apparent that the production is dominated by water channelling rather than the classical water coning. For instance, several Mississippian pools in the Estevan area ofthe province of Saskatchewan contain no bottom-water leg, and yet they produce large quantities of water. They must certainly be receiving pressure support via numerous fractures present in the region. Besides this postdepositional fracturing, these carbonate deposits have been witnesses to several events of replacement of calcium carbonate by dolomite and anhydrite. Whereas fractures act as conduits for the active waters to invade the oil zone, dolomitization increases storage (porosity), and vugs and micro-fractures increase permeability. In addition, site-specific 3-D configuration of the reservoir (intercalation of porous and dense intervals occasionally traversed by fractures, and poor continuity of dense and porous features over inter-well 209
DETERMINATION OF OIL AND GASRESERVES
distances) characterize sweeping ofthe pay zone by the influxing water. Therefore, the reserves drained by horizontal wells depend upon factors such as the prior exploitation of underlying zones within the pool (timing), the level ofheterogeneity and occurrence of dense zones, and the stand-off above the water-oil contacts or the base of the pay. Contrary to what might be anticipated in a classical coning situation, most horizontal wells in developed pools fail to drain significant amounts of incremental reserves over and above what two or three vertical infill wells might drain under similar conditions.
1. Features that improve drainage: • Enlarged drainage volume
In this area, the advantage of higher initial oil rates for the horizontal wells is often negated by sharp declines as the water production increases. Water rates and cumulative water production are seen to increase disproportionately to the corresponding increases in oil production because of the existence of numerous vertical fractures and the prevailing distribution of the invaded water (due to prior operations). Under these circumstances, lateral pressure drops within the horizontal well due to two-phase (or three-phase) flow assume special significance. Consequently, horizontal wells may be doing a poor job of draining oil around their "toes." The situation may be further complicated by the specific reservoir description (porous or tight zones and fractures along the length of the well) and near-wellbore formation damage.
2. Features that hinder drainage:
It follows then that for projecting performance, a detailed knowledge of reservoir description and a proper understanding ofthe geology and hydrodynamics ofthe drainage region around a horizontal well (within the oil pool, including any supporting aquifer) are absolutely essential. Viscosity (temperature) of oil plays an important role by way of causing viscous fingering and limiting volumetric sweep by the invading water.
16.2.5 Recovery Factors Once the drainage volume has been estimated, the next step is to estimate the upper and lower limits of recovery factors for drainage via horizontal wells. An understanding of the behaviour of vertical wells in the same pool in terms of the dominant production mechanisms and the factors limiting production provides important clues to the production behaviour of horizontal wells. As previously mentioned, some features would help in improving recovery whereas others might hinder efficient drainage. The three lists that follow give some of the more important of these characteristics.
• Heterogeneities (sweet spots) within the drainage area; barriers to the flow of bottom water or gas into the horizontal well • Reduced pressure drawdown, which may help to mitigate drainage restrictions (e.g., cresting, fines production) • Effective lowering of the economic oil rate limit (one horizontal well replacing several vertical wells) • Heterogeneities (stratification, barriers to substantial drainage in depletion drive, by-passing of oil in water- or gas-drive flooding) • Previously drained regions within the drainage volume that may be at lower pressures, or higher pressures (watered-out regions). . • Wellbore damage (lower effective well radius) • Lateral pressure drops (turbulence, multi-phase flow, sediments or debris present in the hole) causing effective drainage from only a part ofthe well • Undulating well trajectory or "porpoising" (some sections may get closer to fluid contacts or the tops or bottoms of the pay zones; in some instances, some sections of wells may even be outside the pay zone, reducing the effective well length in a good part of the pay) 3. By examination of geological and hydrodynamic models, some of the questions about the impact of less than ideal conditions on recovery factors may be clarified. These questions could be as follows: • Are small intervals contributing the bulk of the flow? • If SO, will they continue to be recharged adequately? • Is there more severe skin in certain parts of the well? • Could a lateral pressure drop within the well be restricting drainage from some parts ofthe well? • Would early break-through of water or gas be promoted by the dominant flow routes? • Once break-through occurs at any point in the well, would it seriously restrict subsequent drainage by the well?
210
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _sd
RESERVES ESTIMATION FOR HORIZONTAL WELLS
A quantification of these effects on recovery factors could be obtained by a quick, coarse-grid simulation study.
16.3
DETERMINATION OF RESERVES
16.3.1 Determination of Reserves Parameters Average reserve parameters would be difficult to determine without closely examining a geological model of the drainage region around a horizontal well. These parameters could be porosity, permeability (vis-a-vis orientation of the well), characterization of the aquifer and the gas cap, net pay thickness (Reisz, 1992), thickness above or below the well in the case of undulating well trajectory, location ofpay tops and bottoms within the drainage region, fractures, effective well length, reservoir pressure, saturations, damage, and drainage distances.
16.3.2 Key Elements All of the elements of reserves determination for horizontal wells are similar to those applicable to vertical wells. However, the required analysis is usually more rigorous because a detailed analysis of the hydrodynamics of the drainage around each horizontal well must be included. The procedure is iterative to ensure consistency between reserves obtained from volumetric as well as performance analysis and all available geological, reservoir, and production data. The procedure calls for sound engineering judgement regarding appropriate values of parameters to be used for performance projections and reserves estimation and, in addition, requires a clear understanding of the dominant recovery mechanism and the parameters that limit reserves for exploitation of the pool by conventional wells. The possible relaxation of the limiting conditions on drainage using horizontal wells is estimated based on these. A hydrodynamic model for the drainage area incorporating reservoir variations, current state of depletions, and qualitative visualization of flow distribution within the drainage area of horizontal wells is required. Finally, the implication of operational and economic factors on reserves must be explicitly included.
16.3.3 Steps Involved in Reserves Determinations The proposed procedure involves iterations of the following steps until an acceptable determination is achieved: I. Prepare a geological model for the drainage region of the horizontal well. The model should address questions regarding the boundaries, the limits of the drainage area due to any barriers to flow, heterogeneities and facies changes, fluid contacts, anisotropy, directional trends, preferred fracture orientations, micro-fractures, and sweet spots. 2. Prepare a qualitative hydrodynamic model incorporating data on the current state of drainage, the well trajectory, the pressure and saturation distribution prior to the placement of the horizontal well, the effective drainage region, and the flowing pressure distribution around the horizontal well, including any possible interference with offsetting wells. 3. Obtain estimates of various drainage and reserves parameters such as effective pay thickness, shape of the drainage area, sweet spots, drainage distance, porosity, pressure distribution, saturation distribution, compressibility, permeability, kH/kv, and skin. 4. Estimate the hydrocarbons in place in the drainage volume and the range ofthe associated uncertainty. 5. Estimate the range of recovery factors for horizontal wells from data on recovery factors for conventional drainage, and possible relaxation of parameters controlling production. The roles ofvarious influences may be quantified using coarse-grid simulation or engineering judgement. 6. Estimate the initial productivity from the estimates ofdrawdown, permeability (vertical as well as horizontal), compressibility, and saturations. Actual performance or test data may be used for validating estimates of various parameters. 7. Project the production forecast for the specific situation. Performance data, equations, material balance, and simulation results, if available, may be used for validating decline performance. In the absence ofany better data, initial productivity along with volumetric reserves may be used for projecting performance. This data may then be input into economic analysis for obtaining economic reserves.
211
DETERMINATION OF OIL AND GASRESERVES
Depending upon the situation, curves of rate-time, rate-cumulative production, volumeratios, and cumulative volume of gas or water vs. cumulativeoil or gas may help to determine the reserves. Care must be exercised to ascertain that there is adequate history,thatthe performance is determined by reservoir and geological factors only, and that the performance is consistent with the known mechanisms. Data on the performance of horizontal wells in analogous situations, if available, could be useful. Some statistical data on performance of horizontal wells in different oil zones from certain Canadian producing areas over the first 12monthsof production has recently been published (Springer et aI., 1993). Where uncertainty is high, the production forecast should be based upon estimates of initial productivity and volumetricallydetermined reserves. 8. Identify any enhancementpotential to reserves due to prudent operational changes, recompletions, facilities or equipment upgrades. These datacanthen be used for further refining the productionforecast. Another fine-tuningcould be required due to interference with offset wells, if such interference could be established fromtheirperformance (Springer and Flach, 1993). 9. Ensure consistency between reserves based on volumetricdeterminationand productionforecasts. A few iterations may be required to achieve this. 10. Evaluate the range of uncertainties in the reserves estimates and relevant confidencelevels. This will depend upon geological control, the amount of historical data from the pool, the success of costeffective diagnostic or remedial operations, and the length of time the horizontal well has been producing. The Monte Carlo computer simulation method for quantifying confidence levels is described in Section 22.4.4 (Springer et aI., 1991).
References Babu, D.K., and Odeh, A.S. 1989. "Productivity ofa Horizontal Well." SPE Reservoir Engineering, Vol. 4, No.4, Nov. 1989, pp. 417-421.
212
. Bowers, B., and Bielecki, J. 1993. "Horizontal Oil Wells: Economicsand Potential Impact on the Reserves and Supply of Canadian Convential Oil." WorkingDocument, Horizontal Well Committee of the National Energy Board, Calgary, AB, Jun. 1993. Butler, R.M. 1989. "The Potential for Horizontal Wells for Petroleum Production." JCPT, Vol. 28, No.3, May-Jun. 1989, pp. 39-47. Butler, R.M., and Suprunowicz, R. 1992. "Vertical ConfinedWater Drive to Horizontal Well - Part I: Water and Oil of Equal Densities." JCPT, Vol. 31, No.1, Jun. 1992, pp. 32-38. Chaperon, 1. 1986. "Theoretical Study of Coning Toward Horizontal and Vertical Wells in Anisotropic Formations." Paper presented at 61st Annual Fall Meeting, SPE of AIME, New Orleans, LA, Oct. 1986, SPE 15377. Freeborn, R., Russell, B., and MacDonald, A.J. 1990. "South Jenner Horizontal Wells: A Water Coning Case Study." JCPT, Vol. 29, No.3, pp. 41-46. Heysel, M. 1992. "Horizontal Well Performancein the Dina Sandstonein the Provost Area of Alberta." Presented at Annual CIM Technical Meeting,Calgary, AB, Jun. 1992,CIM-ATM 92-34. Joshi, S.D. 1991. Horizontal Well Technology. Pennwell Publishing Co., Tulsa, OK, p. 34. Kuo, M.C.T. 1989. "Correlations Rapidly Analyze Water Coning." O&GJ, Oct. 1989, pp. 87-90. Mutalik, P., and Joshi, S.D. 1992. "Decline Curve Analysis Predicts Oil Recovery from Horizontal Wells." O&GJ, Sep. 1992, pp. 42-48. Papatzcos,P., Herring, T.R., Martinsen, R., and Skjaeveland, S.M. 1991. "Cone Break-through Timefor Horizontal Wells." SPE Reservoir Engineering, Vol. 6, No.3, Aug. 1991, pp.311-328. Poon, D.C. 1990. "Decline Curves for Predicting Performance of Horizontal Wells." JCPT, Vol. 30, No. I, pp. 77-81. Reisz, M.R. 1992. "Reservoir Evaluation of Horizontal Bakken Well Performanceon the Southwestern Flank of the Williston Basin." Paper presented at SPE InternationalMeeting, Beijing, China, Mar. 1992, SPE 22389.
RESERVES ESTIMATION FOR HORIZONTAL WELLS
Russell, B., and Espiritu, R. 1992. Personal communication. Springer, S.1., Mutalik.P; Asgarpour, S., and Singhal, A.K. 1991. "Risk Analysis for Horizontal Wells." Paperpresented at 4th Saskatchewan Symposium, CIM, Regina, SK, Oct. 1991, PaperNo. 13. Springer, S.1., and Flach, P.D. 1993. "A Review of the Drainage Area/lnterwell Spacing Used in Some Established Horizontal Well Projects." DEA44/DEA 67 International Forum"Horizontal Technology - Living With Reality," Calgary, AB, Jun. 1993.
Springer, S.1., Flach, P.D., Porter, K.E., Christie, D.S., and Scott, G.C. 1993. "A Review of the First Five Hundred Horizontal Wells Drilled in Western Canada."Paper presented at 44th Annual Technical Meetingof the Petroleum Societyof CIM, Calgary, AB, May 1993, CIM 93-19. Yang, W., and Wattenburger, R.A. 1991. "Water Coning Correlations for Vertical and Horizontal Wells." Paperpresented at 66th Annual SPE Technical Conference and Exhibition, Dallas,TX, Oct. 1991, SPE 22931.
213
Chapter 17
NUMERICAL SIMULATION
17.1
INTRODUCTION
Numerical simulation is the most sophisticated tool for estimating hydrocarbon reserves and determining methods to use for optimizing the recovery ofhydrocarbons from a reservoir. Numerical simulation has been used in reservoir studies since 1960. The rapid development of digital computer technology in the early seventies stimulated the widespread development and application ofreservoir simulation computer programs. At first, the high cost of software development and computing limited the use ofnumerical reservoir simulation; however, the recent availability of powerful low-cost personal computers and work stations has made it much more accessible to petroleum engineers. Today, numerical reservoir simulators are more efficient and more accurate. This section provides an overview of numerical simulation practice. Readers who wish to gain an in-depth knowledge ofthe mathematical aspects ofsimulation should read the book by Aziz and Sattari (1979). Excellent discussions on practical applications of reservoir simulation may be found in books by Crichlow (1977) and Mattax (1990).
17.2
TYPES OF RESERVOIR SIMULATORS
Reservoir simulation is based on the physical principles of mass conservation, fluid flow, and the conservation of energy. From these come a set of partial differential equations describing the behaviour of fluids in a reservoir. According to the type of process and the number of components required to be modelled, reservoir simulators may be categorized as follows: Black oil simulators, which model multi-phase flow in a reservoir without consideration for the composition of the hydrocarbon fluids. The liquid phase consists of water and the oil and gas in solution. The gas phase consists of only free gas. Mass transfer of the oil component from the liquid to the gas phase is not taken into account.
214
Compositional simulators (Coats, 1980a; Nolen, 1973; Thele et aI., 1983), which account for mass transfer between liquid and gas phase. The hydrocarbon phase is represented by"n" components; k-values and flash equilibrium are used to represent phase behaviour. Enhanced oil recovery simulators, which include in situ combustion (Youngren, 1980; Coats, 1980b), steam stimulation, (Coats, 1978) hydrocarbon miscible (Todd and Longstaff, 1972), carbon dioxide flooding (Chase and Todd, 1984), and chemical injection (Todd and Chase, 1979). These simulators apply the basic concepts of both black oil and compositional simulators with added features to model a particular enhanced oil recovery process. Reservoir simulators have also been developed to model naturally fractured reservoirs. In addition to modelling the processes described, a naturally fractured reservoir simulator must also model the complex flow behaviour in a matrix-fracture system. Naturally fractured reservoirs are characterized by two systems: a matrix system which has low permeability and high capacity, and a fracture system which has high permeability and low capacity. The bulk of the fluid is contained in the matrix system, and fluid flow occurs primarily in the fractures. A comprehensive review of naturally fractured reservoirs is given by Aguilera (1980). The general approach in naturally fractured reservoir simulation is the dual-porosity formulation shown in Figure 17.2-I(a), in which the rock matrix is considered as a series ofdiscontinuous blocks within a continuous fracture system. The matrix blocks act-as the source and feed into the fracture system. The fractures can be thought of as a system of connected pipes. This model was proposed by Warren and Root (1963). Recent developments allow a more vigorous treatment of fluid flow in naturally fractured reservoirs to be incorporated into simulators. In addition to fracturematrix interaction, matrix-matrix flow is permitted; this
NUMERICAL SIMULATION
gives rise to the dual-permeability formulation (Gilman and Kazami, 1988) shown in Figure 17.2-1(b).
I
I
•
l-
•
1
•
I -
•
•
I
•
l-
I
•
l-
1
•
fracture matrix (a) Dual Porosity
• I
•
----
...
•
--
-l-. Flow Out
Lly Llx
I l-
•
Figure 17.3-1 Mass Balance on Reservoir Element
I
that will affect the decisions made by a simulation engineer will be discussed.
fracture matrix (b) Dual Permeability
Figure 17.2-1 Schematic Diagram of MatrixFracture Connectivity
17.3
Flow In
I
I l-
•
Llz
MATHEMATICAL FORMULATION
Mathematical functions for all the cases discussed have been presented in detail in the literature, and so will not be repeated here. In general, the formulations involve the use of partial differential equations that are solved using finite difference schemes. Figure 17.3-1 shows a small volume element ofthe reservoir with dimensions dx, liy, and liz. Simulation involves a mass balance over many elements similar to the one shown. The exact solution to the partial differential equations is rarely available. In practice, numerical techniques are used to obtain approximate solutions to those equations. The finite difference method is the one most commonly used for reservoir simulation. The method transforms the continuous differential equation into a discrete form in both time and space. The reservoir region is subdivided into elements or grid blocks similar to the block shown in Figure 17.3-1. The solution to the system of flow equations is obtained for each grid node. The dependent parameters obtained for each grid node represent the average value for the element. Detailed discussion of the finite-difference method is available in the literature (Aziz and Settari, 1979) and will not be provided here. However, certain concepts
The early approach to' solving the multi-phase flow equations was the Implicit Pressure Explicit Saturation (IMPES) Method, in which the flow equations were combined into a single pressure equation. After the pressure has been advanced in time, the saturations are updated explicitly. This approach assumes that the capillary pressure and transmissibility terms do not change substantially within a timestep. The advantages of the IMPES method are its low computer memory requirement and reduced computation per timestep. The IMPES method has been found to be satisfactory for many problems; however, in situations where high flow rates exist, such as in water coning, gas percolation problems and naturally fractured reservoir simulation, a more stable solution method is required. The fully implicit method, on the other hand, requires the simultaneous solution ofthe multi-phase flow equations (Au et al., 1980).This method requires substantially more computing time and data storage. Increased stability ofthe fully implicit method allows larger timesteps to be used. Most commercial simulators allow the user to specify the method of solution. More advanced simulators offer semi-implicit and dynamic implicit methods. The semi-implicit method solves a subset of the flow equations simultaneously whereas the dynamic implicit method switches between the IMPES and fully implicit methods on an individual grid block according to flow conditions. Unless computer memory and run time limitations present a problem, it is advisable to use the fully implicit method of solution to avoid unnecessary numerical problems.
215
DETERMINATION OFOILAND GASRESERVES
17.4
ANATOMY OF RESERVOIR SIMULATION
Reservoir simulation is a complex engineering task. A simulation study must be planned and organized to ensure that useful results are obtained. The objectives of the simulation study must be clearly defined. The engineer should have a list of specific questions the study should answer, and preliminary reservoir engineering calculations should have been completed. Before carrying out a simulation study, an engineer should be thoroughly familiar with previous reservoir studies. The results and conclusions ofprevious studies may be useful to fine-tune current study objectives and help save time. Once the objectives and scope of the study are clear, a reservoir simulation study generally involves the following phases: 1. Data collection 2. Model grid design 3. Sensitivity tests 4. History matching 5. Performance prediction The following sections describe these phases of the simulation activity.
17.5
DATA REQUIREMENTS
A numerical simulator may be used to model any reservoir. The input data to the simulator describe a unique model for a particular reservoir. The data required to construct a reservoir model may be grouped as follows: Reservoir geometry, which describes the size, shape, internal and external boundaries of the reservoir Rock and fluid properties, which affect the dynamics of fluid flow in the reservoir Production and well data, which describe the well locations, perforation intervals, skin factors, and flow rates
17.5.1 Reservoir Geometry A geometric description of a reservoir is usually derived using a team approach involving geologists, geophysicists and reservoir engineers. A good understanding of regional geology and depositional environment is necessary. Seismic sections are useful in preparing structural maps and positions of faults. Formation top and thickness of zones to be simulated may be obtained from well logs and drilling records.
17.5.2 Rock and Fluid Properties The important petrophysical properties of rock required in reservoir simulation include porosity, absolute permeability, relative permeabilities, capillary pressure data, rock compressibility, and fluid saturations. The average porosity can be determined from core analysis. The porosity is also calculated from well logs. Porosity logs calibrated against core porosity are generally more reliable than log data alone. Absolute permeability is one of the most difficult reservoir properties to define. It is also critical to the prediction of fluid migration in a reservoir. Integrated permeabilities from cores and well-test data should be used in reservoir simulation. In a reservoir where more than one fluid is present, the relative permeability of individual fluids as a function offluid saturation is required. Relative permeability data are usually obtained from laboratory measurements on core samples. The relative permeability relationships are obtained for gas-oil, oil-water, and gas-water systems. Most reservoir simulators use Stone's (I970) model to approximate three-phase relative permeability behaviour. The capillary pressure data are determined from laboratory analyses. Rock compressibility data are obtained from laboratory analyses of the reservoir rock or from published correlations. Formation fluid saturation distributions can be derived from log analysis. Another option is to calculate fluid saturation distributions based on the positions of the water-oil and gas-oil contacts. The fluids may be assumed to be initially either fully segregated (no transition zone) or dispersed (with a transition zone). The capillary pressure curves are used to determine the saturation in the transition zone. Fluid properties include formation volume factors, fluid viscosity, solution gas-oil ratio, and fluid density. The source ofthese data is usually laboratory PVT analysis. Iflaboratory data are not available, correlations can be used to generate them. For compositional simulation, the equation of state is used for calculating fluid pr?perties. The effects of temperature on viscosity, density, relative permeability and capillary pressure are also required for thermal simulation.
17.5.3 Production and Well Data The data required to specify well operation include well locations, perforation intervals, well productivity index, skin and flow rates for each well. Sources ofproduclion
216
_________________.za
NUMERICAL SIMULATION
and well data are pressure tests, drilling records, and well production records. The constraints imposed on wells due to surface facilities or economic limits must also be available. Typical well constraints are water-oil ratio, gas-oil ratio, bottom-hole pressure, and maximum and minimum flow rates.
17.6
RESERVOIR MODEL GRID DESIGN
A reservoir can be modelled with !D, 2D or 3D grid systems. Depending on the objectives of the study, one of the following may be used:
ID models, which have limited applications including material balance, simulation of experiments, and interaction between two wells. In a vertical or dipping !D model, the effect of gravity override, updip gas injection, and bottom water injection can be evaluated. 2D areal models (Figure 17.6-1), which are commonly used in field simulation. The model is suitable when areal flow pattern dominates reservoir performance, and the vertical variation in rock and fluid properties in the reservoir is small.
Figure 17.6-1 20 Areal Model
Figure 17.6-2
20 Vertical Model
2D radial models, which are a special type of 20 models. While most simulation models are defined by cartesian coordinates, the 2D radial models are defined using a cylindrical coordinate system (Figure 17.6-3) and have special applications in the study of near-well effect. The 20 radial models are often called coning models because they are used principally to study water and gas coning behaviour. This type of model is useful in studying single well operations to determine the optimal completion interval, critical flow rate to avoid coning, well deliverability, and well test analysis.
.
,• ,
I I
--
z
I
2D vertical models, which are used to model vertical cross sections of a reservoir (Figure 17.6-2). The applications include gravity segregation effect, effect of stratigraphy, frontal displacement, effects of well completion intervals, and flow into a horizontal well. 2D vertical models are also used to generate pseudo-functions, which reduce three-dimensional simulation to two-dimensional areal simulation (Jack et aI., 1973; Coats et al., 1971).
Figure 17.6-3
20 Radial Model
217
DETERMINATION OF OIL ANDGASRESERVES
3D models, which are used to study large multiple well reservoirs with thick reservoir pay sections, significant vertical variation in rock and fluid properties, faults, and partial communication between layers (Figure 17.6-4). 3D models are also used to study large reservoirs with several noncommunicating producing horizons, multiple completions with or without commingled production, aquifer influx, and horizontal well development.
In areal simulations where the effect ofwell pattern and infill wells is studied, sufficient grid blocks should be used so that all the wells in the reservoir model are separated by several grid blocks. Ifpossible, the orientation of the grid system should parallel trends of high permeability. A full field reservoir simulation may not be necessary to satisfy the study objectives. In many cases, a study of an element of symmetry from a reservoir with repeated well patterns may be sufficient.
17.7
RESERVOIR MODEL INITIAL,IZATION
Preceding sections have indicated that reservoir data are available mainly at well locations. The reservoir simulator, however, requires reservoir parameters for each grid node in the reservoir model. The common practice is to construct contour maps of the reservoir parameters. The reservoir model grid is then overlaid on the contour map, and values are assigned manually to each grid node. Figure 17.6-4
3D Model
An efficient reservoir model is one that satisfies the study objectives at the lowest cost. Since the cost of a simulation study, including engineering person-time costs and computing costs, is proportional to the complexity of the model, it is desirable to employ the simplest model possible. The model, however, must be able to represent reservoir geometry and positions offaults and wells, and be able to show fluid migration patterns. It is difficult to design an optimal grid system for a reservoir. However, the following guidelines may be useful. Since the parameter values for each grid node in a reservoir model are the average values for the block, the number of grid nodes should be increased in the area of interest or where reservoir parameters are expected to change rapidly. Typically, smaller grid blocks are required around wells. One caution is that abrupt changes in grid sizes introduce truncation errors. As a general rule, the ratio of the grid lengths for two adjacent grid blocks should be less than two. Local grid refinement features are available in most reservoir simulators. This feature allows any grid in a reservoir model to be subdivided into smaller grids without adding extra blocks in other parts ofthe model. Local grid refinement can be very useful in areas with wells and faults. By subdividing a well block vertically into more layers, local grid refinement provides a means to specify completion intervals more precisely.
Some simulators utilize reservoir parameters at well locations and generate the distribution ofparameters for each grid node using a second or higher order interpolation scheme. The number of wells and their locations can affect the quality of interpolation. The reservoir parameters assigned to each grid node by this method should be examined carefully and any anomalies corrected. Most reservoir- simulators have the ability to display the reservoir model and initial conditions on a computer display screen for visual inspection. The initial pressure and fluid saturation distributions in the reservoir model can be defined using the interpolation scheme described. Alternatively, the simulator can be used to calculate pressure and saturation distribution based on specified water-oil and gas-oil contacts and reference pressure.
17.8
MODEL SENSITIVITY ANALYSIS
The numerical truncation error associated with timestep size and grid size can affect the accuracy of simulation results. Before a detailed history match is performed, the sensitivity of a reservoir model to truncation error should be analyzed. The effect of grid size on simulation results can be evaluated with a simple model of a representative portion of the reservoir that includes an injection and production well. A series of simulation runs with decreasing grid size is performed. When the reduction in grid size does not change the simulation results beyond
218
__________________________c.
NUMERICAL SIMULATION
the accuracy required, the grid size is considered acceptable. A smaller 2D model is often used to perform grid sensitivity tests because it is cumbersome to change the grid block sizes in a complex 3D field scale model. The effect of timestep size should also be investigated in the reservoir model sensitivity analysis. The timestep size used in field scale simulation is indirectly controlled by how often the well rates are changed. However, when there is no change in well rates and the maximum timestep size is not controlled, the numerical truncation error can be significant. A few simulation runs should be made with different timestep sizes to determine the maximum timestep size that will produce no adverse effect on the results. Most reservoir simulators use automatic time step selection algorithms to determine the appropriate timestep size. The algorithm selects a timestep size that will maintain pressure, saturation or temperature change over a timestep at the level specified by the user. If automatic timestep selection is used, the maximum time step size determined from the sensitivity study should be imposed.
17.9
HISTORY MATCHING
The data available to construct a reservoir model is often limited, so it is very unlikely that the initial reservoir model will provide a good representation of the reservoir. However, this data represents the best estimates of the engineers and geologists. The predictions obtained from a reservoir model are thus not very useful unless the model is able to produce a performance similar to the historical data. History matching is a process in which the parameters of the model are adjusted until the computed results are similar to the historical data. The adjustment of parameters should be carried out within reasonable orders ofmagnitude; the input of unrealistic data for the sake of obtaining a good history match is never justified. The historical data usually includes observed pressures, gas-oil ratio, and water-oil ratio. In cases where a well is produced at a constant pressure or total fluid rate, the match variable can be the oil or gas rates. In some cases break-through time may be an important match parameter. Before any of the historical data is used in the history match, an engineer should analyze the data to confirm the accuracy ofthe recorded information. The engineer must make sure that the data is in comparable units and
the pressure data has been corrected to the proper datum. When long production history is available, it is customary for simulation engineers to specify monthly, quarterly or semi-annually averaged daily rates as input to the simulator. These daily production rates are obtained by dividing the recorded cumulative production during the selected period by the number ofdays in that period. The production data to be matched should also be averaged in the same fashion. History matching is a time-consuming exercise. It can take more than fifty percent of the time allocated to a reservoir SImulation study. There is no system for changing the reservoir parameters that would result in a good history match, so engineers must rely on their reservoir simulation experience and their knowledge of the reservoir. The general rule in history matching is to change the parameters that have the largest uncertainty and also the largest effect on the results. The engineers must constantly check to make sure the parameters are within reasonable limits.
17.10 FORECASTING RESERVOIR PERFORMANCE Following a satisfactory history match, the reservoir model may be used to predict reservoir performance. From the objectives of the simulation study, a list of prediction cases is developed. It is always useful to establish a base case for comparing different proposed development strategies. The base case is usually the continuation ofthe existing operating strategy. The following are typical questions a reservoir model may answer: • Estimate of reserves • Well pattern and spacing • Injection well location • Drilling schedule • Critical production rates • Well completion strategy • Well deliverability • Vertical vs. horizontal well performance • Migration of fluid • Recovery mechanisms The model provides estimates of fluids in place at initial and current conditions. Ideally, these estimates should compare with the results from volumetric and material balance calculations. The results of the base case forecast should match reasonably well with the
219
DETERMINATION OFOIL AND GAS RESERVES
results from decline curve analysis. Any significant differences in these results should be investigated and an explanation included in the engineering report. If the objectives of the simulation study include the determination of ultimate recovery for a number of reservoir development alternatives, these cases are simulated to the economic limit in order to estimate reserves. It is necessary to define the appropriate criteria for reservoir abandonment conditions, such as minimum producing rates, maximum water cut, minimum pressure, and other factors that determine the economic limit. When a simulation model is used to estimate ultimate recovery, it is important to recognize that results are subject to considerable uncertainty, especially if the model is developed for a reservoir with limited production history. However, the comparison of ultimate recoveries from different development strategies can be very meaningful and an excellent basis for choosing between alternative development methods for a field.
17.11 USE AND MISUSE OF RESERVOIR SIMULATION The discussions in the preceding sections highlight some applications of numerical reservoir simulation. One major advantage of simulation models is that it can be used to evaluate different field development strategies at very small cost and without irreversible damage to the reservoir. The misuse of reservoir simulation, however, can lead to erroneous conclusions and costly mistakes.
A reservoir model should not be treated as a "black box" for turning out numbers. Reservoir simulation is no substitute for good reservoir engineering. Only intelligent use of reservoir simulation can avoid costly mistakes.
17.12 SUMMARY Reservoir simulation is a very useful tool for studying reservoir behaviour, for comparing alternative field development strategies, and for forecasting production and estimating reserves. Reservoir simulation involves the use of complex mathematical formulations, numerical approximations, and reservoir descriptions, all ofwhich contain many uncertainties. It is necessary to use good engineering judgement in conducting simulation studies and in interpretating the results obtained. The advances in computer technology show no signs of slowing. This trend will facilitate widespread applications of reservoir simulation technology to petroleum reservoir engineering problems in the future.
References Aguilera, R. 1980. Naturally Fractured Reservoirs. PennWell Publishing Company, Tulsa, OK. Au, A.D.K., Behie, A., Rubin, B., and Vinsome, P.K.W. 1980. "Techniques for Fully Implicit Reservoir Simulation." Paper presented at the 1980 SPE Annual Technical Conference and Exhibition, Dallas, TX, Sep. 1980, SPE 9302. Aziz, K., and Settari, A. 1979. Petroleum Reservoir Simulation. Elsevier Applied Science Publishers, New York, NY.
The use ofreservoir simulators requires at least as much experience and engineering judgement as routine reservoir calculations. In considering the results obtained from reservoir simulation, three questions must be asked:
Chase, C.A., and Todd, M.R. 1984. "Numerical Simulation of CO, Flood Performance." SPEJ, Dec. 1984, pp. 597-605.
I. Are the final parameters used to obtain a good history match reasonable?
Coats, K.H. 1978. "A Highly Implicit Steamflood Model." SPEJ, Oct. 1978, pp. 369-83.
2. Is the simulator used in the study appropriate for the process under consideration?
- - - . 1980a. "An Equation of State Compositional Model." SPEJ, Oct. 1980, pp. 36376.
3. Are the simulation results consistent with other engineering calculations? Forecasts of reservoir performance are more reliable during the first few years. Longer term prediction tends to be less reliable because errors caused by uncertainties in reservoir description become more significant with time. Predictions of absolute value of recovery, for example, will be less reliable in the long term. However, comparison ofrelative differences between similar prediction cases are less likely to change.
- - - . 1980b. "In Situ Combustion Model." SPEJ, Dec. 1980, pp. 533-54. Coats, K.H., Dempsey, J.R., and Henderson, J.H. 1971. "The Use of Vertical Equilibrium in TwoDimensional Simulation of Three-Dimensional Reservoir Performance." SPEJ, Mar. 1971, pp. 63-71. Crichlow, H.B. 1977. Modern Reservoir Engineering - A Simulation Approach. Prentice-Hall, Inc., Englewood Cliffs, NJ.
220
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _71
NUMERICAL SIMULATION
Gilman, lR., and Kazemi, H. 1988. "Improved Calculations for Viscous and Gravity Displacement in Matrix Blocks in Dual-Porosity Simulators." JPT, Jan. 1988, pp. 60-70. Jack, H.H., Smith, OJ.E., and Mattax, C.C. 1973. "The Modeling of a Three-Dimensional Reservoir with a Two-Dimensional Reservoir SimulatorThe Use of Dynamic Pseudo Function." SPEJ, Jun. 1973, pp. 175-85. Mattax, C.C., and Dalton, R.L. 1990. Reservoir Simulation. SPE Monograph, Vol. 13. Nolen, J.S. 1973. "Numerical Simulation of Compositional Phenomena in Petroleum Reservoirs." Paper presented at the 1973 SPE Symposium on Numerical Simulation of Reservoir Performance, Houston, TX, Jan. 1978, SPE4274. Stone, H.L. 1970. "Probability Model for Estimating Three-phase Relative Permeability." Trans., SPE of AIME, Vol. 249.
Thele, KJ., Lake, lW., and Sepehrnoori, K. 1983. "A Comparison of Three Equation-of-State Compositional Simulators." Paper presented at the 1983 SPE Symposium on Reservoir Simulation, San Francisco, CA, Nov. 1983, SPE 12245. Todd, M.R., and Chase, C.A. 1979. "A Numerical Simulator for Predicting Chemical Flood Performance." Paper presented at the SPE Symposium on Reservoir Simulation, Denver, CO, Feb. 1979, SPE 7689. Todd, M.R., and Longstaff, WJ. 1972. "The Development, Testing, and Application of a Numerical Simulator for Predicting Miscible Flood Performance." JPT, Jul. 1972, pp. 874-82. Warren, r.s., and Root, PJ. 1963. "The Behaviour of Naturally Fractured Reservoirs." SPEJ, Sep. 1963, pp. 245-55. Youngren, G.K. 1980. "Development and Application of an In Situ Combustion Reservoir Simulator." SPEJ, Feb. 1980, pp. 39-51.
221
Chapter 18
DECLINE CURVE METHODS
18.1
INTRODUCTION
The decline curve is a basic tool for estimating remaining proved reserves, and can be applied once there is sufficient history to show a trend in a performance variable that is a continuous function of either time or cumulative production. Forecasts are made by extrapolating trends to an endpoint where production is expected to cease (i.e., an economic limit or a related parameter such as water-oil ratio). Such forecasts are particularly useful in the latter stages of depletion when trends are clearly evident and there is insufficient revenue to justify a more comprehensive analysis. The origin of decline curves is uncertain, but their usefulness to monitor day-to-day operations likely predates their use as a forecasting tool. Indeed, prior to the general trend to centralize and use computers for production accounting and engineering functions, it was common practice for field offices to maintain production graphs to assist with day-to-day operations. Decline curve methods have a universal appeal because they provide a simple visual representation of a complex production process. In some cases a visual interpretation is too simplistic, and some background knowledge is needed in order to draw reliable conclusions. In particular, it should be appreciated that forecasts are usually based on linear extrapolations of historical trends. Such extrapolations are strongly affected by any transformation used to obtain a linear relationship. It is also implicitly assumed that the factors causing the historical decline will continue during the forecast period. Some factors causing the decline are physical processes (e.g., pressure depletion, coning, interface movement) that are not easily changed. However, other factors such as regulatory environment (e.g., well spacing, gas-oil ratio penalties, maximum rates) and operating practices (e.g., type and size of artificial lift, hours of operation, frequency ofwork 0 vers, gas gathering system pressure) can quickly change from time to time and from lease to lease.
18.2
SOURCE AND ACCURACY OF PRODUCTION DATA
It is worthwhile to review the source of the basic data used to prepare decline curves. Production accounting functions such as royalty payments, allocation ofgroup production to individual wells, gas plant balances, and reports to regulatory agencies usually have a monthly reporting and reconciliation period. Daily records of hours ofproduction, test rates, system pressure and other operating variables are kept to make these monthly reports, but they are often discarded or placed in dead files after a few months. The permanently accessible record ofproduction and injection data usually consists ofmonthly totals for gas, oil, and water production (injection), operated hours, and wellhead pressure. Monthly totals are usually converted to daily rates for graphing purposes because facility capacity, contract rates, and economic limits are usually expressed as daily rates. The frequency and quality of well tests are the most important factors affecting official production records. For gas wells, it is common practice to measure raw gas production for each well and to run annual deliverability tests. The measured production helps to ensure reliable well-by-well cumulatives; however, the seasonal and variable demand for gas can result in highly variable rates, and this tends to complicate decline analyses. For oil wells, it is common practice to measure group production, and test individual wells monthly. The dayto-day demand for oil is less affected by markets, and many oil wells are produced at capacity, which tends to simplify decline analyses. The least complex production facility is a single well served by a single-well battery. In such a facility, there is no doubt about the source of the production. The measurement accuracy will also be reliable if the facility is properly sized. Among the most complex facilities are central treating facilities that serve several multiwell satellite batteries equipped with three-phase test separators and operating at high pressure. In this case, the total (group) production of oil, gas, and water is
222
5
DECLINE CURVE METHODS
allocated to individual wells on the basis of their operated hours and test rates. The accuracy ofthis allocation depends upon the frequency of well testing and the variation of oil, gas, and water rates among wells. A good indication of the allocation accuracy is given by the quotient of theoretical production and measured production for each fluid (e.g., oil, gas, and water). Theoretical production is the piece-wise sum of the product of test rate and time interval. These quotients (called proration or allocation factors) are usually considered to be acceptable if they are in the range of 0.95 to 1.05. It should be noted that errors in test rates or producing hours will cause misallocations among wells and pools. Errors in gas-oil and water-oil ratios can be somewhat larger because the allocation factor for each fluid may differ (e.g., an oil allocation factor less than 1.0 and a water allocation factor greater than 1.0). Thus, gas-oil ratio and water-oil ratio curves often show more "noise" than their corresponding rate curves.
18.3
TERMINOLOGY
The following are definitions of terms used in this chapter: Decline curve: the generic label applied to many different types of charts, graphs, and data representations. The most basic decline curves show the change in oil, gas, or water production rate with time (rate-time graphs). The production rate is usually expressed as volume per day to facilitate understanding; however, hourly, weekly, monthly, and yearly rates may also be used. Graphs with time as the independent variable are easily understood and the rate-time data is directly applicable to economic evaluations. The other common independent variable is cumulative oil or gas production (rate-cumulative graphs). The advantage of these is that an extrapolation to the economic limit yields a direct estimate of the proved reserve. Calendar-day rate: the monthly total production (injection) divided by the number of days in the month. Operated-day rate: . the quotient of monthly production (injection) and actual operated hours in the month multiplied by 24. If calendar-day and operatedday rates are plotted on the same graph, any separation of the curves is a measure of the shut-in or down time. Ifthere are no rate controls, the area between the graphs is a measure of "lost production." Operated-day rates may define a better decline trend than calendar-day rates because they smooth out the variation caused by down time.
Ratio curves: the gas-oil ratio (GOR) and water-oil ratio (WOR) curves that are commonly plotted for oil wells, These ratios are a measure of the efficiency of the oil production process. An increase in either ofthese ratios is usually accompanied by a decrease in the oil rate. GOR penalties are often applied as a rate control measure to limit the amount ofreservoir voidage caused by high-GOR wells. For gas wells, the corresponding ratios are condensate-gas ratio (CGR), liquid-gas ratio (LGR), and water-gas ratio (WGR). The CGR is a measure of the richness of the raw gas. In gas cycling schemes, the CGR decreases with increasing dry gas break-through. The WGR ratiois a measure of'production problems associated with liquid buildup in wells, hydrates, and water coning. Cut curves: the fraction of oil or water cut in the liquid production from oil wells. These curves are another measure of the efficiency of the oil production process. Their fixed range (i.e., 0 to I) provides an alternative criterion for an economic limit. Reservoir performance charts: the composite presentation ofrate-time graphs supplemented with reservoir data (e.g., reservoir pressure, interface depth) and performance variables (gas-oil ratio, water cut, number of producing wells, water injection, cumulative oil). These charts are often maintained for a lease or unit by the operator, and for a field or pool by a regulatory agency. Figure 18.3-1 is an example of a reservoir performance chart for the gas-cycling and gas-cap operations for a pool in Alberta. In the figure, IR is the injection rate and cd is the calendar day. WGR is the water gas ratio. Production performance charts: charts graphed on semi-log paper which utilize the fact that the product or quotient of two straight lines on semi-log paper is another straight line with a slope related to the slopes of the other two. The idea is to use this slope interdependence to help estimate the decline rate. The advantage is that the decline should be more reliable because more ofthe data has been used to estimate it. Figure 18.3-2 is an example of a production performance chart for a pumping well where the production is controlled by the artificial lift. The gas production is not shown because it is not a factor in the decline. The slopes of the oil, water and WOR curves are interrelated.
18.4
SINGLE·WELL VS. AGGREGATED· WELL METHODS
Decline curve methods may be classified many ways; however, any classification should recognize the difference between analyses for a single well and analyses
223
DETERMINATION OF OIL AND GASRESERVES
Me
'C
.Q
100
•
Me
24
~
~
£!; ~
16
''C"
8
.Q
M e 8
~o
~
q. (rna/d)
:E' ••
',E" o
M e
24
·0
16
o
G
~
10
,E
- - - - - - - Operated-day Rate - - - Calendar-day Rate
8 o
40
a:
~
n:I7::'.7"91""" ' . ",, 0 6"',T""6""26"'3T""64"T6=-=5T""9."T6=7T""66:r6=-9f:"90:r9""f:" 92:r9::1 3
01:: 7-';47=5T-7"".
Figure 18.3-1
Reservoir Performance Chart
for aggregated production from a group of wells. Decline curve analyses for single wells are widely used and readily interpreted because they have the following advantages: • All the raw data can be displayed. • Decline trends are easy to recognize and often correlate with the total fluid production rate. • The economic limit can be directly applied to estimate reserves. • The conventional decline equations have been shown to have a strong foundation based on reservoir engineering principles. On the other hand, decline curve analyses and forecasts for aggregated production from a group of wells are also widely used, but may be misinterpreted for the following reasons: • Only part of the raw data can be displayed. • Decline trends may be masked by the number and variability ofthe wells contributing to the aggregated production. • The economic limit cannot be directly applied to estimate reserves. • The analyses are largely empirical (may be enhanced by statistical analysis).
Figure 18.3-2
18.5
Production Performance Chart
DECLINE CURVE METHODS FOR A SINGLE WELL
Decline curves are a visual tool, and it is easy to overlook that trends and extrapolations (linear or curved) are defined by mathematical equations. The most common equations were given in classic papers by Arps (1945,1956). Table 18.5-1 summarizesArps'rate-time and rate-cumulative equations along with dimensionless time and production groups proposed by Gentry (1972). The decline relationships in Arps' first paper were based on the loss ratios between equal time intervals. While these relationships were useful for tabular data, they are ofless interest today with the easy access to computers and graphing programs. Mead (1956) refined the loss ratio and series methods and was among the first to attempt to associate the type of decline with the drive mechanism. The equations in Table 18.5-1 are solutions to the following differential equation:
(dq~dt)
d -b= ---'-dt
(I)
224
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _1
DECLINE CURVE METHODS
18.5.1 Exponential Decline
where b decline exponent q = producing rate t = time At the time they were formulated, the equations were considered to be empirical and were classified as exponential, hyperbolic or harmonic. The classification was based on the value of the exponent, b, used to characterize the change in decline rate with the rate of production. The classification is still widely used, but it is now recognized that the value of b is not limited to the range 0 ~ b ~ 1.
Table 18.5-1
Exponential decline is most commonly used because both the rate vs. cumulative and the log (rate) vs, time graphs are linear. Figure 18.5-1 is an adaptation of a normalized rate, q/q., vs. a normalized cumulative relationship, N'/(NP)l yr s by Schoemaker(1967) showing both decline rate, d, and time, t, as parameters. The diagram uses one year as the reference time, and decline rates are expressed as percentage per year. The chart illustrates a subtle difference between the slope, a, and the annual decline rate, d. Various combinations of decline rate and time (such as d = 5%, t = 10 years;
Decline Curve Equations Type of Decline Exponential
Characteristics
Decline is constant.
Exponent
b=O
Hyperbolic Decline varieswith instantaneous rate raised to power "b."
Harmonic Decline is directly proportional to the instantaneous rate. b = 1.0
Rate-time relationship
q = qJI
-1 + ba, t)'
Rate-cumulative relationship b
(~) -I a.t> - - , b
Dimensionless time, td
a,t=(~)-I
-]
Dimensionless production, q"
N,= q, t
I_(~qi) In
(~)
In (~)
~---qqt
m-
l
where a = decline as a fraction of producing rate (slope of line) ai = initial decline rate b = decline exponent e = natural logarithm base 2.71828 Np = cumulative production q = producing rate at time (t) qi = producing rate at the beginning of the decline t = time Source: After Arps, 1956;Gentry, 1972.
225
-----=-r-/ ;;"', I DETERMINATION OF OILAND GAS RESERVES
d = 10%, t = 5 years; and d = 50%, t = I year) do not result in the same final rate, q/qi' Values of 0.6,0.59 and 0.5 can be read from Figure 18.5-1.The difference in rate is due to the number of times the annual decline rate is applied(becauseofthe similarityof declinecurve calculations to compound interest or depreciation calculations). The slope, a, corresponds to very short compound periods,and inthe mathematical limitingprocessis calledcontinuous compounding. The decline rate, d, is related to the decline slope, a, by the expression: (2)
d = I - e'
Schoemaker shows how Figure 18.5-1 can be used to solve many practical exponential decline problems. He points out that five parameters are used in the equations (q, q, N p, t, and either a or d) and, when any three are known, the other two can be determined from the figure. For example, if a new well has a capacity of
100 m3/d and is expected to decline at 10 percent per year, what will the rate and cumulative production be after 10 years? The answers can be read from the intersection of the 10 percent decline and 10-year lines (i.e., q/q; = 0.35 N/(Np)1 = 6.18). Thus, after 10 years, the rate will be 35 m /d and the cumulative will be 100 x 365 x 6.18 = 225660 m3•
r
18.5.2 Hyperbolic Decline With hyperbolic decline, the decline is proportional to the productionrate raised to the power b. Unfortunately, hyperbolic decline does not plot as a linear relationship on common graph papers (i.e., linear, semi-log, or loglogco-ordinates). Priorto the widespread use ofpersonal computers, this lack of linearity was the main reason for the restricted use of hyperbolic declines. Slider (1968) preparedtransparent overlays (each overlayhad a fixed b-value and a family of decline rates) that could be visually matched to log (rate) vs. time graphs. Once
0.3 1-----hI-+--\-iL.IJ-~:.+¥W~_",._¥~¥~_J!."___A.,._,,A.,._,,jL~.L-_bA=Lj__r::::::,"5<""1
\'0 'lIS
ZO'lIS
2
3 4 567 Normalized Cumulative Production, Np/(Np) 1 y'
Source: After Schoemaker, 1967.
Figure 18.5-1
226
Exponential Decline Chart
8
9
10
DECLINE CURVE METHODS
the overlays were prepared, the visual matching technique could be applied with about the same ease as an exponential decline extrapolation. The next development in handling hyperbolic declines was based on the dimensionless groups shown in Table 18.5-1. Gentry used these groups to develop the generalized rate-time and rate-cumulative graphs shown in Figures 18.5-2 and 18.5-3, respectively. Figure 18.5-2 is simply the family of rate-time graphs for a unity decline rate (a, = 1.0). It should be noted that the exponential decline (b = 0) plots as a straight line as expected on the log rate vs. time co-ordinates. Figure 18.5-3 is more difficult to understand because the transformation to dimensionless production changes the character of the graph. This figure is not directly comparable to standard rate-cumulative graphs. It should be noted that the harmonic decline (b = I) does not plot as the expected straight line on the log rate vs. cumulative co-ordinates. Figure 18.5-3 shows that the cumulative production is strongly related to b, but some calculations are required to quantify the sensitivity in every case. To apply Gentry's method, two rate-time pairs are read from a decline graph, and these values along with cumulative production over the time period are applied to Figure 18.5-3 to determine b. With a knowledge of b and the time period between points, Figure 18.5-2 is used to calculate the decline rate. Agbi and Ng (1987) showed that the dimensionless production equation for hyperbolic decline in Table 18.5-1 can be expressed as a nonlinear equation with "b" as the only unknown.
Source: After Gentry, 1972.
Figure 18.5-2
Decline Curve Analysis Chart Relating Production Rate to Time
10.0
~~~,y
1.0:f-
0.8
0.6
0.4
0.2
o
q, = Np/(q,t) Source: AlterGentry, 1972.
Figure 18.5-3
f(b)
Decline Curve Analysis Chart Relating Production Rate to Cumulative Production =
Y(x" - I) (1- b) - b (1 - x··
I )
(3)
where x = q/q y = N/(qjt) b = decline exponent The authors then used standard numerical techniques to find the roots of this equation (i.e., f(b) = 0). The equation was demonstrated to have at least two roots, one at b = 0 and another at b = 1. In general the equation behaves as a cubic equation with three real roots including b = 0 and b = I. If the decline is truly exponential or harmonic, then the data will also satisfy the dimensionless production equations for these declines in Table 18.5-1. The value of "b" need not lie between oand I. Because this is a general solution, it shows that negative values are also possible. The method is actually a numerical equivalent of Gentry's graphical solution based on Figures 18.5-2 and 18.5-3. It should be noted that both methods assume that any type of decline is specified by two rates, the cumulative production, and the actual time on production (i.e., qj' q, Np ' t). Ifthe production rigorously followed the Arps' equations and there were no measurement or reporting errors, then everyone using the method would get the same answer. Unfortunately, because real data does not rigorously follow the equations and has some noise, the method is data-dependent. When different data pairs are used, different values may be calculated for both a and b. The method does not give a quality or "goodness-of-fit" criterion, but if the theoretical curve
227
DETERMINATION OFOILANDGASRESERVES
is plotted on the same scale as the raw data, a visual comparison is always possible. In many cases the purpose of decline analysis is to estimate the value of future production. Experienced evaluators avoid extrapolating hyperbolic declines over long time periods because they frequently result in unrealistically high reserve and value estimates. The characteristic of hyperbolic decline (i.e., continuously decreasing decline rate) can result in extremely long producing lives that are incompatible with experience elsewhere and with expectations for equipment life. Many wells are observed to trend toward an exponential decline in their later life. Figure 18.5-4 is a log rate vs. time overlay developed by Long and Davis (1988) to cope with this problem. Each line on Figure 18.5-4 is for a fixed b-value. The range extends from 0.3 to I. 7, which allows handling of those wells where b-values greater than 1.0 have been observed, e.g., in Alberta tight gas (Milk River) and fractured and heterogeneous reservoirs (Austin Chalk and Spraberry). The numbered dots on Figure 18.5-4 correspond to tangent points where an exponential decline would start with
the specified exponential decline rate (slope). The power of the method is that it uses all of the data to establish the nature ofthe decline, but allows selection of a point at which the decline is expected to hold the decline rate and follow an exponential decline. The method is particularly suited where monthly production rate is plotted on standard three-cycle graph paper. Robertson (1988) developed the following production rate equation, which is hyperbolic initially, but asymptotically exponential with time:
(4) where ~
= a dimensionless constant to control how strongly hyperbolic the initial decline is before asymptotically becoming exponential
The value of ~ ranges from 0 to 1.0 and is related to the abandonment pressure and the rock and fluid properties. This equation provides for another slack
10' , - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - ,
95 80 60 50 40 30
Hyperbolic Decline Type Curves 3-Cycle Semilog x 20 Years
b=1.7
4
3
b= 1.5 b = 1.3
5 6
b = 1.1
b = 1.0 5
10 Li-L..1-.LJe-.J.-L..1-l.2l.-L...L.LJe-.J.-L...LL..l-l..-L-L..L.l-L...L.J.:::J=--....L-L...LL.l-L...LL..l..-LJ-J o 12 24 36 48 60 72 84 96 108 120 132 144 156 168 180 192 204 216 228 Source: Long and Davis. 1988.
Figure 18.5-4
Time (months)
Hyperbolic Curve Overlay
228 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
I,
DECLINE CURVE METHODS
parameterto fit actualdataby an analytical decline equation. Although the author did not recommend it, the equation provides a mathematical framework to perform a fully numerical curve fit with results very similar to the manual and visual process of Long and Davis (i.e., similar to the Agbi and Ng extension to the Gentry graphical solution). Lack of linearityof hyperbolic declines is no longer an obstaclewhendataare displayedandprocessedby cornputer. There are numerouslow-cost softwareprograms for performing decline curve analysis. Most programs apply a least-squares criterion to find the value of "b" which best fits the reported production data. Once "b" has been determined, the production data are displayed on the samegraph with the theoretical declineand forecast. Some programs allow manual changes to the least-square parametersto obtain a personalized visual fit to the data.
18.5.3 Harmonic Decline Harmonic decline is a special case of hyperbolic decline in which the decline rate is directly proportional to the instantaneous production rate. The ratecumulative relationship from Table 18.5-1 shows that harmonic declinewillplot as a straightline on a log rate vs. cumulative plot. Figure 18.5·5 is a production performance graph for a Blairmore oil well, which illustrates severalsegmentsof harmonic decline(Purvis, 1987). The rate and ratio curves are somewhat erratic due to measurement andbatteryprorationerrors. Thetotal fluid production rate, qw + qo' is determined by the size and operating speed of the artificial lift equipment. During the past 30 years the total fluid rate has varied widely with a few periodsof relatively stablerate. It is interesting to note that during these dramatic changes in rate, the WOR + I graph has shown moderate sensitivity to rate.This is surprisinggiven that duringthe period 1974 to 1976, the qw + ~ rate was about 10 times the initial rate. Anotherinteresting feature of Figure 18.5-5 is the shape of the cumulative water plus cumulative oil (Qw+ Qo) graph. Thegraphisinitiallyconcavedownward, butafter a periodof continuous waterproduction, the graphtrends toward a straight line. In fact, three rectilinear sections of this graph are evident. The slopes of these three segments were transferred to the WOR + I graph. This is a useful characteristic of harmonic decline, which can be applied after a period of continuous water production. The reason for this particularcharacteristic is that the derivative, d(Q w + Qo)/ d(Qo)' equals
10'
0------------------, 2
,
Qw+Qo(10 mid) _
q. + q, (m'/d)
10
10
o
20
40
60
80
Cumulative Oil Production (10' m') Source: AfterPurvis. 1987.
Figure 18.5-5
Production Performance Graphs
WOR + I. The functional relationship can be demonstrated as follows. Any linear segment of the Qw+ Qo graph has the functional form (5)
which, when differentiatedwith respect to Qo yields d(Q. + Q,) = WOR + I = C 10"· '2Q, d(Q,) 3
(6)
where c t , c2' cl = data-specificcoefficients Consequently, the dash line approximations of the ~ graphwere drawnto honour the usual slope interdependence among the graphs. Because the well is part of a multi-well pool, no conclusions can be drawn on the effect the well rate has on pool recovery. However, it is clear that increasedrates have increased recovery from this particularwell.
229
~I :~,
I ,
DETERMINATION OF OIL AND GAS RESERVES
18.5.4 Dimensionless Solutions and Type-Curve Matching
Figure 18.5-6 shows his analytical transient type curves combined with Arps' empirical depletion type curves. The depletion type curves are essentially the same as those proposed by Gentry; however, Fetkovich plotted q/qj instead of q/q and used log-log coordinates to facilitate type-curve matching. It is apparent from Figure 18.5-6 that the transition from transient to depletion behaviour occurs at a dimensionless time of approximately 0.3. Figure 18.5-6 also shows that until the dimensionless time exceeds 0.3, it is impossible to know the type of decline that ultimately develops. Thus, the safest approach to extrapolating trends early in the life of a well is to assume an exponential decline.
Fetkovich (1980) used simplified material balance and inflow performance relationships for both gas and oil wells to show that the Arps' empirical equations match up with some ofthe classical solutions to the radial flow diffusivity equation. Exponential decline was shown to be the long-time solution to the constant terminal pressure case (constant bottom-hole pressure). The short-time (transient) solution is a function of the reservoir size expressed as r/rw ratios (r, = external boundary radius, rw = wellbore radius). Fetkovich demonstrated that for oil wells (slightly compressible single-phase flow) the type of decline does not change with the drawdown. On the other hand, for gas wells (compressible single-phase flow) it was demonstrated that a change in back pressure changes the type of decline. This finding helps explain the reliability ofdecline analysis for oil wells. In many practical cases, wells are produced at capacity and the bottom-hole pressure does not change significantly over time (i.e., the well is pumped off). Fetkovich demonstrated that empirical decline curve analysis has a solid theoretical base. 10
So
<0 Transient
<00 100 100 1.0
Type-curve matching was first used to interpret pressure buildup and drawdown data. The procedure involves comparing the pressure-time data from a well with a family ofdimensionless solutions. The same general procedure is used for decline data. Fetkovich summarizes the procedure as follows: I. The actual rate-time production data are plotted on a log-log tracing paper of the same size as the type curves to be used. Any convenient units can be used for rate or time because a change in units
+
Depletion
1000
r'/rw = 100000 Exponential Common to Analytical and Em irical Solutions Analytical Type Curve Solution
10.1
10 tdO
Source: After Fetkovich, 1980.
Figure 18.5-6
Composite of Analytical and Empirical Type Curves
230
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _6
DECLINE CURVE METHODS
simply causes a uniform shift of the raw data on a logarithmic scale. 2. The tracing paper with the data curve is placed over a type curve and shifted until a good match is obtained. The axes of the two curves must be kept parallel during this process. Several different type curves should be tried to obtain the best fit ofall the data. 3. To make a forecast, the type curve is traced onto the tracing paper overlay. Future rates are then simply read from the real-time scale on which the raw data was plotted. 4. To evaluate deciine-curve constants or reservoir variables, the type of decline is noted and a match point is selected anywhere on the overlapping portion of the curves. With a knowledge of the type of decline and the coordinates of the match point on both sheets, the constants or variables are evaluated from the appropriate dimensionless relationship. Many software programs for performing decline analysis facilitate a computerized Fetkovich overlay procedure. These programs greatly facilitate matching actual data to any of the numerous type curves and are particularly useful for analyzing gas wells and other wells with extended transient behaviour.
18.6
DECLINE CURVE METHODS FOR A GROUP OF WELLS
Estimating the reserves for a group ofwells could be an onerous task if a decline analysis were performed for each well. Consequently, it is common practice to perform one decline analysis for the aggregated production from all the wells in a lease or pool. While this is common practice, it is not as reliable as one might assume. When the production from a group of wells is aggregated (summed), only the total is available for plotting, and much of the raw data is omitted from the analysis. Sometimes the average-well rate is plotted to make the analysis appear more like that for a single well. Another major difficulty is that the economic limit is not clearly defined for aggregated production. This difficulty also makes forecasts hazardous because some of the wells will be abandoned during the forecast and will no longer be contributing to the aggregate. Clearly, the theoretical base for aggregated production is not as solid as that for single wells. Despite the foregoing problems, decline curve analysis can be successfully applied to aggregated production. For example, if there is a wide variation in rates, the
analysis can be improved by splitting the wells into a few groups having similar characteristics (e.g., rates, water c~ts, remaining life, well spacing, gas-oil ratios). By making sub-groups of wells with similar remaining lives, an economic limit can be applied with more confidence. To establish a reliable decline for tight gas pools, the wells should be grouped on the basis of their onproduction date and initial production rate. In general, aggregate rate vs. cumulative curves exhibit decline trends which are easier to interpret (i.e., better defined) than those for aggregate rate vs. time curves. Figure 18.6-1 is a production-performance graph for the Leduc D-3 A Main Pool showing its final 19 years of oil production. The pool had an oil zone of 11.6 m sandwiched between a large gas cap and an aquifer. By 1974 the oil zone's thickness had decreased to about 2.5 m and, by the start of gas cap blowdown in November 1989, the oil zone was less than 1 m thick. Figure 18.6-1 shows that three correlation segments are required to account for the distinct change in the slopes of the curves related to pool oil operations. The gas production from oil wells (solution plus coned gas) was constrained by the portion of the Devon gas conservation plant capacity available to the pool. Thus depending upon the available capacity, many ofthe highest-GOR wells would be shut in. Also in September 1978, the pool water handling (artificial lift and treating capacity) became another constraint and contributed to the change in slope of the '10, qo + qw' and oil cut curves. From 1984 to 1988, several wells were worked over and every effort was made to maximize oil recover; prior to the impending gas cap blowdown. After blowdown commenced, wells were shut in as the aquifer displaced the extremely thin oil zone past the completion intervals of the oil wells. It is apparent from the linearity of the dash lines that valid rate forecasts could have been made after a short bit of history in each correlation segment. Figure 18.6-2 is the corresponding rate vs. cumulative graph. A good estimate of ultimate recovery could have been made from this graph as early as 1975. Note also that the straight-line segments on Figures 18.6-1 and 18.6-2 are characteristic of exponential decline.
18.6.1 Statistical Method Purvis (1990) showed that many of the deficiencies of decline analysis for a group of wells can be overcome by using a log-normal distribution to quantify the changes in well rates over time. The method provides a means of forecasting future well counts through proper application of the well economic limit. The method 231
DETERMINATION OF OIL AND GAS RESERVES
.,---
---,1
- ---
Oil Cut
0.1
10'
Producing Wells
10
q, (m'/d/well)
Figure 18.6-1
Production Performance Graph TIme Line (years)
l!.
E 2
"0
27
29
31
33
Cumulative 011 Production (106 m3)
Figure 18.6-2
Rate-Cumulative Production Graph
was applied in Alberta to the Redwater 0-3, Swan Hills BHL C, Twining Rundle A, and Viking Kinsella Wainwright B Pools. Figure 18.6-3 shows some historical and forecast distributions ofwell rates for the Pembina Cardium Pool. The pool has been penetrated by over 5000 wells. In June 1990, there were 3411 producers and 1426 injectors, but only 2716 producers and 1037 injectors were operated. The number of inactive producers and
injectors indicates the maturity of the waterflood and the fact that, at current prices, it is uneconomic to operate over 1000 wells. In December 1970, the median well rate was 4.25 m3/d, and by October 1990, it had decreased to 1.53 mvd. The dash lines show that the median well rate is forecast to decline to 0.45 m3/d by the year 2030. The variance of the distribution has continuously decreased, as shown by the decreased slope of the distributions. A significant feature of the plot is that the lines for years 1982, 1986 and 1990 tend to focus and pivot at about 0.3 m 3/d/well and 2 percent of wells. This focus is taken to mean that the economic limit is about 0.3 m3/d/well and that typically 2 percent of the wells are at, or below, the economic limit at any time. The coordinates of Figure 18.6-3 are representative of the log-normal distribution, the properties of which are the subject of a classic textbook (Aitchison and Brown, 1966). The linearity of the well rate distributions indicates that they are approximately log-normal. The median for the log-normal distribution is the arithmetic average of the logarithms of the population (i.e., numerically the geometric average). The small circles at 50 percent of the wells in Figure 18.6-3 are the geometric averages of the numeric values used to plot the distributions. These values are in good agreement with the values that would be read from the graph. Indirectly the circles show that the distribution is approximately log-normal. Table 18.6-1 summarizes other numerical averages and also summarizes the results of chi-square goodness-of-fit tests of the raw data when divided into 13 equally spaced class intervals. The log-normal is characteristic of phenomena or processes defined by multiplication (or division). Examples oflog-normal distributions range from sedimentary petrology (Podruski et aI., 1988) to the probit methods used for biological assays. There are good reasons to expect well production rates to have a log-normal distribution. The radial flow equation that defines steadystate production rates simply multiplies and divides parameters that are constant or that change very slowly. The two most important terms are the pay thickness and permeability which are often log-normally distributed. These actually help to ensure a log-normal rate distribution because the product of two log-normal distributions is another log-normal distribution. The rate-ratio-cumulative graphs in Figure 18.6-4 show that the pool has been on continuous decline for 20 years. The linearity ofany rate graph on the linear coordinates is characteristic of exponential decline. The most
232
---------------
..sa
DECLINE CURVE METHODS
notable change in Figure 18.6-4 is the levelling ofpool water production at about 20 000 m 3/d in 1975. The levelling occurred after the pool was put on good production practice and voidage replacement was relaxed. The pool oil rate exhibits different decline rates before and after the water production rate levelled. The median well decline rate was less affected by the change. It is not clear from Figure 18.6-4 if the decline is several segments of exponential decline or if the levelling ofoil rates is due to harmonic decline. Some ofthe mitigation ofpool oil-rate decline from 1980 through 1986 was due to increased well count (re-activations and new drilling) and through judicious selection of the wells to be operated. Since 1986 the pool performance has deteriorated significantly.
Standard Deviation -1
12 12 12 12 12 10
10
0
+1
+2
1970 1974 1978 1982 1986 1990
a;
roi (])
1ii
II:
The forecasts shown by the dash line were calculated on the basis of exponential decline rates of 3 and 4 per cent per year for the median well rate and the pivot point shown in Figure 18.6-3. The same forecasts on logarithmic coordinates are shown in Figure 18.6-5 which illustrates that it would not be realistic to assume a linear (i.e., harmonic) extrapolation of either the pool rate or median well rate. The median well rate only changes from 4.25 to 1.53 m3/d so it appears linear on both linear and logarithmic coordinates. The calculations for Pembina support the pragmatic approach of limiting a harmonic decline to some time period followed by an exponential decline. Forecasts based on harmonic declines of 3 and 4 per cent for the median well rate resulted in ultimate recoveries of 212 and 207 million m3, respectively. These forecasts were not
(5
1
10.1 L--J..._'----'-_'--.L-.L-.L--'---'-_-'-_----'
2
5
10
20 30 40 50 60 70 80
90
95
98
Percentage of Wells
Figure 18.6-3
Distribution of Well Rates, Pembina Cardium Pool
Table 18.6-1
Statistical Parameters for Pembina Cardium Pool Date (month-year)
Statistics for Raw Data Numberof wells Arithmetic average Geometric average Harmonic average
Standard deviation Coefficient of variation Variation Lorenz measure
12-1970
12-1974
12-1978
10-1982
12-1986
10-1990
2681 8.61 4.25 2.43 13.35 1.55 0.68 0.59
2490 5.89 3.33 1.88 7.79 1.32 0.65 0.55
2481 4.01 2.57 1.58 4.84 1.21 0.60 0.48
2514 3.20 2.20 1.45 3.33 1.04 0.57 0.45
2658 2.77 1.90 1.30 2.90 1.05 0.58 0.46
2615 2.07 1.53 1.13 1.84 0.89 0.54 0.42
4.15
17.15
14.53
16.79
6.48
7.87
X' Goodness-of-fit test for log-normal (13 class intervals)
233
7
--.r
DETERMINATION OF OIL AND GAS RESERVES
24 ,---,-"7"<--,-----.----r--..,.--.,.---,
105r--------Oil and Water Rate (m
Pool Oil Rate (10' m'/d)
16
~,
,,
8
Oil Rate (m3/d)
--. g ~
"
24
~
3/d)
." ~,
"
"
4% ~~ -, 3%
"
'" ~
103
""'''''11'11'
"
\ \ I I
\ 13%
\\4% \I II
Well Count (x 10')
II
.........~%113%
""\\
Pool Water Rate (10' m'/d) 10'
1\ I I 113%
Median Well Rate (m'/d)
4
4%
---
2
10
-~-4%" '3%-
a 4
:=><--
Median Well Rate (m3Id) ~" - , -_ ,
WaR (m'/m')
o co
2 ,
a I---.--=r--,--,.-....,...---r----r--i 60
\ \ \ \
"
_ _ _ _ _ _ _-
16
"
100
140
180
220
Cumulative Oil Production (10' m')
,
100
I
!
,
"
'"
,',
140
,
\ \
o
'" '" ~
\ \3% 4%
I" '.",1,
180
220
Cumulative Oil Production (10' m')
Figure 18.6-4 Rate-Ratio-Cumulative Graph, Pembina Cardium Pool
Figure 18.6-5 Production Performance Graphs, Pembina Cardium Pool
believed because of the extremely long producing life for the pool.
Lohec (1984a) demonstrated the effect of reservoir geometry on production rate in reservoirs involving frontal displacement mechanisms. He noted that reservoir geometry is one of the first characteristics ofa reservoir to be defined and understood (e.g., seismic structure definition, well control, gas-oil and water-oil contacts). If the frontal displacement is gravity-dominated, the remaining hydrocarbon volume often approximates a simple geometric shape (e.g., a cone, wedge, or cylinder) and simple expressions may be developed for the change in hydrocarbon volume with hydrocarbon recovery. Next, the rate ofproduction is assumed to have a simple power law relationship to the remaining hydrocarbon volume. These simple expressions provide the theoretical basis for calculating rate-cumulative and rate-time performance (i.e., the same role that material balance and inflow performance relationships play in developing type curves for wells). Lohec (1984b) applied the method to the East Texas, Friendswood, Conroe, and Hawkins fields.
18.6.2 Theoretical Methods Simple theoretical models are sometimes used to make forecasts of pool production. These models can often be rearranged into rate-time or rate-cumulative equations to prepare a family of forecasts which have some key, but uncertain, reservoir property such as permeability as a parameter. The family of forecasts can then be used for matching aggregated pool production (i.e., similar to the type curve matching of individual well production). The performance of pools where oil is displaced by either natural water drive or by water injection can often be characterized by a semi-log plot of WOR, oil cut, or water cut vs. cumulative recovery. To provide a theoretical basis of these cut-cum curves, Ershagi and Abdassah (1984) proposed a co-ordinate transformation based on fractional flow and Welge's recovery formula.
234
7
DECLINE CURVE METHODS
Richardson and Blackwell (1971) showed that several reservoir flow mechanisms have an element of symmetry and a single dominant force such that a simple mathematical model could be developed to forecast reservoirperformance. Their models for gravity segregation and water under-running are simple enough to be used as the theoretical basis for some decline curve analysis.
18.7
SUMMARY
5. Decline curves for aggregated production from a group ofwells do not have a strong theoretical base, but with appropriate caution and understanding, the analysis can be reliable. 6. Experienced evaluators often use an exponential decline to extrapolate a hyperbolic decline to prevent unrealistically long lifetimes and reserve estimates.
Decline curves are widely used to convey information about past production performance and to forecast future performance and reserves. The following tips and precautions should be noted:
7. Dimensionless type curves are powerful tools for analyzing and forecasting individual well behaviour. These curves are particularly useful for tight-gas and other wells with extended transient behaviour.
I.
8. Well production rates for a group ofwells produced at capacity can be characterized by a log-normal distribution. Consequently, the decline rate for the median well is the statistically significant decline rate for aggregated production.
Production decline is caused by one factor or a combination of factors including reservoir depletion, equipment wear, operating practice, and regulatory environment. It is risky to extrapolate historical trends without understanding the factors contributing to the decline or anticipating new factors that can come into play. For example, the decline ofan oil well in an undersaturated pool will change as the pool pressure decreases below the bubble point. Failure to anticipate such a change can negate what would otherwise be reasonable extrapolation of past performance.
2. The product or quotient of two exponentials is another exponential. This recursive characteristic is useful for any linear functions on semi-log paper. For example, if the oil rate and gas rate are linear on semi-log paper, the gas-oil ratio must also be linear, with a slope related to the oil and gas rates. Similarly, iftotalliquid production is constant (typical of pumping wells) then both oil rate and oil cut must have the same slope. Another example ofslope interdependence is that a trend of increasing total fluid production will tend to offset or mitigate an oil rate decline. 3. The misallocation ofgroup production to individual wells can cause ratio curves to be more erratic than the corresponding rate curves. 4. Decline curves for single-well pools produced at capacity have the strongest theoretical base followed by single-well analysis in multi-well primary production pools. Well-by-well decline trends in multi-well pools subject to pattern floods can be difficult to recognize and forecast due to fluid migration.
9. All of the available data (e.g., reservoir pressure, gathering system pressure, injection volumes, etc.) should be plotted and considered when extrapolating a decline trend to make a production forecast.
10. The results of simple theoretical models and volumetric calculations may be used to constrain and enhance forecasts starting from an observed decline trend.
References Agbi, B., and Ng, M.e. 1987. "A Numerical Solution to Two-Parameter Representation of Production Decline Curve Analysis." Paper presented at Petroleum Industry Applications of Microcomputers, Montgomery, TX, Jun. 1987, SPE 16505. Aitchison, J., and Brown, J.A.C. 1966. The Lognormal Distribution. The University Press, Cambridge, U.K. Arps, U. 1945. "Analysis of Decline Curves." Trans., AIME, Vol. 160, pp. 228-247.
---,. 1956. "Estimation of Primary Oil Reserves." Trans., AIME, Vol. 207, pp. 182-191. Ershaghi, L, and Abdassah, D. 1984. "A Prediction Technique for Immiscible Processes Using Field Performance Data." JPT, Vol. 36, pp. 664-670. Fetkovich, MJ. 1980. "Decline Curve Analysis Using Type Curves." JPT, Vol. 32, pp. 1065-1077. Gentry, R.W. 1972. "Decline-Curve Analysis." JPT, Vol. 24, pp. 38--41.
235
s
DETERMINATION OFOILANDGASRESERVES
Lohec, R.E. 1984a. "Analytic Approach Evaluates Frontal Displacement Mechanism." O&GJ, Vol. 82, No. 38, pp. 83-89.
Purvis, R.A. 1987. "Further Analysis of Production_ Performance Graphs." JCPT, Vol. 26, No.4, pp. 74-79.
- - - . 1984b. "Analytic Approach Applied to Known Reservoirs." O&GJ, Vol. 82, No. 39, pp. 92-97.
---.1990. "Pool-Production and Well-Count Forecasts." JCPT, Vol. 29, No.6, pp, 80-87.
Long, D.R., and Davis, M.J. 1988. "A New Approach to the Hyperbolic Curve." JPT, Vol. 40, pp. 909912. Mead, H.N. 1956. "Modifications to Decline Curve Analysis." Trans., AIME, Vol. 207, pp. 11-16. . Podruski, lA., Barclay, lE., Hamblin, A.P., Lee, PJ., Osadetz, K.G., Procter, R.M., and Taylor, G.C.
1988. Conventional Oil Resourcesof Western Canada. Geological Survey of Canada, Paper 87-26.
Richardson, lG., and Blackwell, RJ. 1971. "Use of Simple Mathematical Models for Predicting Reservoir Behavior." JPT, Vol. 23, pp. 11451154. Robertson, S. 1988. "Generalized Hyperbolic Equation." Unsolicited paper, Aug. 1988, SPE 18731. Slider, H.C. 1968. "A Simplified Method of Hyperbolic Decline Curve Analysis." JPT, Vol. 20, pp. 235-236. Schoemaker, R.P. 1967. "Graphical Method for Solving Production Decline Problems." World Oil, Vol. 165, No.5, pp. 122-125.
236
--------
..a
-I Chapter 19
RECOVERY FACTOR STATISTICS
19.1
INTRODUCTION
Proper management of a hydrocarbon reservoir requires . a reasonably accurate estimate of reserves early in the life of a pool when important decisions are made respecting the depletion strategy. The notion that a simple correlation exists between recovery factor and readily definable parameters has considerable appeal; however, attempts to find one have been largely unsuccessful (American Petroleum Institute, 1984). While there is no substitute for detailed geological and engineering evaluations, recovery factor statistics are useful for bracketing expected recoveries before such evaluations are possible. Average recoveries are generally reliable for estimating the aggregate reserves in a given geological play, but they can be very misleading if used to estimate the reserves ofan individual reservoir. For new discoveries, it is common practice to obtain a preliminary recovery factor from similar mature pools in the same geological play. Unfortunately, this method of analogy can be risky because the available pools may be immature or a poor match for the pool in question. When using analogous pools to estimate recovery, the evaluator is well-advised to monitor the early performance of the pool for deviations from expected behaviour, and to revise recovery estimates accordingly. This chapter focuses on natural or primary oil recovery, which results from the natural energy sources available in oil pools. These natural energy sources take the form of six drive mechanisms that can operate alone or in combination. The range of recoveries and relative importance of these drive mechanisms are discussed in Section 19.3 with reference to some Alberta pool examples. Unfortunately, a breakdown ofrecoveries by drive mechanism is not possible because many pools have combination drives, and this information is generally not captured in a reserve database. Recovery factor distributions and average recovery values are presented for various pool groupings to examine differences related to pool size, fluid density, lithology, and geological age. In addition, average recoveries by geological
play are included with a brief discussion of their use. Several plots of recovery factors vs. common reservoir parameters are also included to illustrate the problem of finding a simple correlation for recovery. Section 19.4 covers the drive mechanism for gas pool recovery.
19.2
DATA SOURCE AND RELIABILITY
The Alberta Energy Resources Conservation Board (ERCB) maintains several databases that store a wide variety of information useful in reserve studies. These databases are shown in Table 19.2-1. The recovery data presented in this chapter was taken from the ERCB's reserve database, which contained information for about 6800 oil pools and 23 800 gas pools at year-end 1990 (Energy Resources Conservation Board, 1991). Since most of the reserves in the western Canadian sedimentary basin are found in Alberta, this reserve database is relatively complete and, because of its size, should be representative of other major producing basins.
Table 19.2-1 Public Data Available for Reserve Studies Category
Types of Data
Geological
Core, well logs, regional maps
Basic well
Completions, treatments, drillstem tests
Performance
Production, pressures, deliverabitily tests
Analyses
Pressure-volume-temperature, conventional and special core; oil, water, and gas compositions
Reserves
In-place volumes of oil and gas, recovery factors, reserves, cumulative
production, pool area, net pay, porosity, water saturation, formation volume factor, fluid density, reservoirtemperature, initial pressure, datum depth Other
Progress reports for enhanced oil recoveryschemes, applications
Source: Energy Resources Conservation Board, 1993.
237
DETERMINATION OFOIL AND GAS RESERVES
The reliability ofan individual reserve estimate is largely a function of data quality and quantity, which are in tum related to available technology, and the quality, size, and stage ofdepletion of a reservoir. The reliability of a reserve also depends on the knowledge and experience of the evaluator. When setting a reserve, the ERCB often has the benefit of company geological and engineering estimates to compare with its own. Since recovery factors are obtained from the division of reserves by in-place volumes, they can be no more reliable than the least accurate of these two estimates. In general, recovery factors for large pools should be more reliable. Other things being equal, large pools will have more wellbores to help define the areal extent of a pool and other reservoir parameters. This added information should improve in-place volume estimates. Large pools are normally developed first in most producing basins; therefore, they will have accumulated the most performance data. Arps (1956) discusses how reserve estimates improve with the addition ofperformance data. Ofcourse, the rate and stage ofdepletion in a large pool will depend on reservoir quality, economics, and regulatory constraints. Another factor that affects recovery estimates is the cost of gathering and analyzing data. In pools with large reserve potential, it is much easier to justify these costs. On the other hand, advances in data acquisition technology will benefit new discoveries more than the large mature pools. There is one thing to keep in mind: a small change in recovery factor can translate into significant reserves for a large pool. As a producing basin matures, new discoveries become smaller. Many small pools have only a single wellbore penetrating them. In these pools, in-place volumes are based on an assumed area, usually some fraction of the drilling spacing unit of the well. In the future, 3-D seismic data may help to overcome this problem by providing a much improved understanding of pool geometry. Today, when a small pool is suspended or abandoned prematurely, it is seldom clear whether the disappointing recovery is due to an optimistic in-place volume or an optimistic reserve. When this situation occurs, the ERCB sets the pool's reserve equal to its cumulative production for administrative purposes. In many cases, the resulting recovery factor is less than 1.0 percent, but appears in the database as 1.0 percent due to rounding. The importance of reservoir quality in assessing recovery or reserves cannot be overstated. When reservoir quality is being characterized, the first items usually compared are average values of porosity,
238
permeability, and water saturation. Well-established methods are used to define these parameters. Some other important factors include layering or stratification, fractures, pool geometry, and rock wettability. These factors are not as easy to quantify using single numerical values. The inability to properly account for all these parameters and how they vary throughout a reservoir creates the largest errors in reserves.
19.3
CONVENTIONAL CRUDE OIL
19.3.1 Natural or Primary Drive Mechanisms The six primary recovery mechanisms are gravity segregation drive, solution gas drive, bottom-water drive, edge water drive, gas cap drive, and expansion drive. It could be argued that there are really only five if the water drive mechanisms are combined. In general, the different recovery mechanisms are additive with the proper combination of reservoir and operating conditions; however, they can also compete with one another, and they can be rate-sensitive. Guerrero (1961) provided recovery ranges for oil pools where each of these mechanisms dominate (Table 19.3-1). His ranges suggest that gravity segregation drive will give the highest recoveries. While this mechanism operates in all pools, it requires a pool with large vertical reliefand sufficient time for drainage to realize these high recoveries. In Alberta, Bonnie Glen D-3 has the vertical relief, buta strong bottom-water drive coupled with a gas-cap drive did not give the gravitysegregation mechanism enough time to fully develop. Table 19.3-1
Primary Oil Recovery by Drive Mechanism
Drive Mechanism
Expansion Solution-gas Gas-cap Edge-water Bottom-water Gravity-segregation
Recovery (% OOIP) Range
Average
2-5 12 - 25 20 - 40 20 -40 35 - 60 50 -70
3 18 30 30 45 60
Source: After Guerrero, 1961.
Solution gas drive (also known as dissolved-gas drive) is the most common primary recovery mechanism in Alberta oil pools. As it happens, this drive mechanism
RECOVERY FACTOR STATISTICS
operates in Alberta's largest oil pool, Pembina Cardium, which has a primary recovery factor of II percent. The efficiency of a solution-gas drive is largely a function ofpool production characteristics. As the pressure drops below the saturation pressure, dissolved gas comes out of solution and reduces oil flow in two ways: first, the oil becomes more viscous as it loses its lighter ends; second, the escaping gas soon reaches the point where it becomes mobile (critical gas saturation) and competes with the oil for access to the wellbore by lowering the oil relative permeability. Recovery seldom exceeds 20 percent, and values in the 5-15 percent range are more common. Guerrero's recovery range for solution-gas drives appears somewhat high for Alberta's light and medium density oil pools, but it is clearly too optimistic for heavy density oil pools, as will be shown later. Bottom-water drive is the most important recovery mechanism for many of Alberta's Devonian (carbonate) reef pools. A review of Leduc (D-3) pools along Alberta's Rimbey-Meadowbrook reef trend is instructive with respect to the relative efficiency of the bottom-water drive mechanism. Glen Park D-3 has a high primary recovery factor of 72 percent. The pool has a thick oil zone (39 metres), high permeability,little stratification, and no original gas cap. On the other hand, Homeglen-Rimbey D-3 has a recovery factor of only 9 percent with a thin oil zone (7.5 metres) and a large gas cap (53 metres). Leduc- Woodbend D-3 is similar to Homeglen-Rimbey with a thin oil zone (II metres), but it has less volatile oil, a less active bottom-water drive, and a smaller gas cap (18 metres). The primary recovery factor for the pool is 56 percent, but water injection was used to enhance the natural water drive, resulting in a total recovery factor of 66 percent. Further updip on the trend, Redwater has a recovery factor of 62 percent. The oil zone here is about three times as thick as at Leduc-Woodbend and there is no original gas cap. Of interest is the high salinity brine encountered at Redwater. The density contrast between the advancing water and the by-passed oil may be assisting recovery by a buoyancy effect. Edge water drives are less common in Alberta, and they are not as effective as bottom-water drives. Often where edge water is present, it is relatively inactive; where it is active, other factors tend to reduce its effectiveness. For example, Joarcam Viking has an edge-water drive, a gas-cap, a thin oil zone (3 metres), and small reservoir dip. These factors contribute to premature water and gas coning. The pool's primary recovery is 37 percent.
The gas cap drive mechanism can have a wide range of recovery efficiencies depending on the relative size and orientation of the gas cap to the oil zone. About 20 percent of Alberta's oil pools are discovered fully saturated with an original gas cap. Secondary gas caps may also form in pools with good vertical permeability under solution-gas drive. One of Alberta's highest primary recovery pools (75 percent), Westerose D-3, had a large original gas cap (64 metres), but it also had a thick oil zone (73 metres) and a strong water drive. Considering that most high recovery pools in Alberta have bottom-water drives or combination drives, Guerrero's recoveries for gas-cap drive also appear optimistic. Expansion drive results from fluid expansion with pressure depletion and is only significant for highly undersaturated pools. The majority of Alberta's oil pools are discovered at or near their saturation pressure; therefore, expansion drive normally contributes only a small fraction of the primary recovery. One of the most undersaturated pools in the province is Snipe Lake Beaverhill Lake. This pool had an initial pressure of26 MPa and a saturation pressure of 9 MPa. The pool's primary recovery is only 12 percent.
19.3.2 Oil Recovery Factor Distributions When oil recovery factors are plotted on a frequency histogram (Figure 19.3-1), they produce a skewed distribution with a long tail at higher recoveries. This shape is characteristic ofa log-normal distribution. McCrossan 50
40
~30
~ e..-
Pools = 5918 OOIP (1 Oem') = 8302 Mean = 11.67 Weighted-Mean = 18.68 Median = 10.00 Standard Deviation =10.76 Co-Variance = 0.92
(/)
"0
o
0..
20
10
o o
10 20 30 40 50 60 70 80 90 100 Recovery (% OOIP)
Figure 19.3-1 Oil Pools
239 t
~
DETERMINATION OFOIL AND GASRESERVES
(1969) showed that both oil in place and oil reserves in western Canada have log-normal distributions. A plot of oil recovery factors on probability paper (Figure 19.3-2) gives a reasonablystraightline,which also suggests a log-normaldistribution. The stair-step nature of
Standard Deviation
-1
0
+1
+2
1L
is
o
~
-
e
Primary 10
~
IJ!
12
5
Figure 19.3-2
10
20304050607080 Percent of Pools
90 95
98
Distribution of Primary Oil Recovery Factors
the plot results from rounding recoveries to the nearest 5 percent for new and immature pools.
19.3.3 Average Recovery Factors Several average values are used to suggest the central tendency of distributions: the weighted-mean, the arithmetic mean, the median, and the mode. Since oil recovery distributions are skewed, caution must be exercised with any use of these average values. The weighted-mean recovery is a value commonly reported (EnergyResources Conservation Board, 1990). The weighting parameter used is usually the in-place volume. This average recovery is particularly useful for aggregate reserves in a geological play. A quick method to calculate it is to divide the total reserves for a group of pools by the total in-place volume. This weighted-mean recovery is consistently higher than arithmetic-mean, median, or moderecoveries forskewed oil recovery distributions. One explanation is that large pools dominate the weighted-mean recovery and that they tend to have higher recoveriesthan smallpools. Of course,this generalization will not be true in everycase.
240
The difference between the arithmetic-mean and median recoveries is usually insignificant, but the arithmetic-mean recovery is consistentlyhigher. The mode,or most common recovery value, is only 1.0 percent if the completepool sample is considered. This value is artificially low because of the procedure mentioned earlier for handling small pools that become inactive prematurely. If these pools are excluded, the mode recoveryjumps up to 10 percent; this is a more reasonable valuefor a viablepool, andprobablythe best recovery to assumefor new poolsas a lastresort. It could be argued that this recovery level becomes more likely .as a producingbasin matures and pool size drops. This mode recovery of 10 percent will be used in later sections as a baselineto compare recoveriesfrom different pool groupings. Table 19.3-2 shows average primary and enhanced recoveries by crude oil type and recovery mechanism. Theseareweighted-mean values. The data suggests that two-thirds of Alberta's reserves come from natural or primary depletion mechanisms, and this represents 19 percent of the total oil in place. It is interestingto note that light and medium pools under waterflood have a lower primary recovery than those under straight primary depletion. This indicates that enhanced recovery is targeted for pools where the natural drives are less effective. Solvent floods appear to be very successful based on the enhanced recovery component. It should be remembered that waterfloodingis a viableoptionfor most solvent flood pools; therefore, the success of a solvent flood should be measured against waterflood recovery. If this is done, the averageincremental recovery for the solventmechanism drops by halfto about 14 percent. In general, vertical solvent floods, which are gravity stable, have the highest recoveries. Horizontal solvent floods generally sufferfromgravityoverride and rapid break-through of the solvent. In Alberta, most solvent floods use hydrocarbon-based solvents.
19.3.4 Pool Size Large pools (original oil in place over lOx 106m3) that have reached an advanced stage of depletion (over 50 percent) were used to generate Figure 19.3-3. When comparedwith the recovery distributionfor the total oil pool sample (Figure 19.3-1), the fraction of pools with recoveries less than 10 percent drops considerably and the average values are significantly higher; however, the mode (most common)recovery remainsat about 10 percent. While the distribution tapers off at higher recoveries, it is moregradual. This supports thesuggestion that large pools have better recoveries. However, it
RECOVERY FACTOR STATISTICS
Table 19.3-2
Average Oil Recoveries
Oil Type & Mechanism
Original Oil No. of In Place Pools (10' m3)
Light-Medium Primary Waterflood Solvent flood Gas flood Heavy Primary Waterflood Total
Average Recovery (% OOIP) Primary Enhanced Total
3374 2675 844 69
4485 213 53 8
22 16 27 41
na 14 31 5
22 30 58 46
I 184
270
1433 63
8 9
na 20
8 30
8416
6255
19
8
27
Source: Energy Resources Conservation Board, 1990.
Pools = 374 OOIP (10 6m3 ) = 5787 40
Mean =25.34 Weighted-Mean = 22.09 Median = 20.00 Standard Deviation = 16.86 Co-Variance = 0.67
~30
~
o'"o Cl..
Pools = 4485 OOIP (10'm 3 ) = 6884 40
Mean = 13.01 Weighted-Mean = 20.78 Median = 10.00 Standard Deviation = 11.42 Co-Variance = 0.88
20
10
10 20 30 40 50 60 70 80 90 100 Recovery (% OOIP)
Figure 19.3-3
Large Mature Oil Pools
should be noted that most of the higher recovery pools are Devonian reefpools with active natural water drives.
19.3.5 Fluid Type: Light and Medium vs. Heavy The density ofAlberta's conventional crude varies from 730 to 990 kg/m", Since oil density is closely related to oil viscosity, recovery would be expected to be sensitive to oil density. Figures 19.3-4 and 19.3-5 confirm that light and medium density pools will recover a higher percentage oftheir oil in place than heavy density pools.
10 20 30 40 50 60 70 80 90 100 Recovery (% OOIP)
Figure 19.3-4
Light and Medium Oil Pools
Using the mode recovery of 10 percent for comparison, 40 percent of light and medium pools will exceed this recovery level, but only IS percent of heavy pools do this well. Average recoveries for light and medium pools are about double those of heavy pools. Most of Alberta's oil pools are classed as light and medium density. A cutoff of 900 kg/m! generally distinguishes light and medium from heavy; however, the ERCB may also classify a pool by the type of market in which the crude oil is sold. Most of Alberta's heavy oil pools are found in shallow, Lower Cretaceous rock in 241
:",
~I DETERMINATION OF OIL AND GASRESERVES
60,---------------, Pools = 1433 OOIP (106 m' ) = 1418
50r----------::--:------., Pools = 3941 6 OOIP (10 m' ) = 4865
50
40
Mean = Weighted-Mean = Median = Standard Deviation = Co-Variance =
40
7.46 8.49 5.00 6.81 0.91
~30 ~
t.-
.!Il30 o
(;
a.
a. 20
.
Mean = 8.92 Weighted-Mean = 11.33 Median = 10.00 Standard Deviation = 7.41 Co-Variance = 0.83
Il)
o
o
20 10
10
10 20 30 40 50 60 70 80 90 100 Recovery (% OOIP)
Figure 19.3-5
Heavy Oil Pools
east-central Alberta. The initial gas in solution and the initial pressure in heavy pools are low relative to light and medium pools. Heavy oil pools assigned the higher recoveries (up to 45 percent) are usually associated with an active regional aquifer like the one found in the Provost Dina play. In many heavy oil pools, reduced well spacing is used to improve drainage and recovery. It is not uncommon to have a well every two to four hectares.
10 20 30 40 50 60 70 80 90 100 Recovery (% OOIP)
Figure 19.3-6 Clastic Oil Pools
50..----------------, Pools = 1977 OOIP (10 6 m' ) = 3437 40
Mean = 17.14 Weighted-Mean = 29.09 Median = 15.00 Standard Deviation = 13.85 Co-Variance = 0.81
19.3.6 Lithology: Clastics vs, Carbonates The type ofrock encountered in a reservoir can be very important in determining the level and range of recoveries expected. Figures 19.3-6 and 19.3-7 show recovery distributions for clastic and carbonate pools, respectively. There is a clear difference in the shape of these distributions; the carbonate group has more pools with higher recoveries. Two of every three carbonate pools exceed 10 percent recovery, but only one of every four clastic pools do this well. Average recovery values are consistently higher for carbonate pools. The weightedmean recovery for carbonates (29 percent) is more than double the value for clastics (11 percent). As previously mentioned, carbonate pools are very important in Alberta. Many have very efficient primary drives, mainly bottom-water. They represent just one-third of the pools and have one-third of the oil in place, but they contribute two-thirds of Alberta's conventional oil reserves. Alberta's largest pool, Pembina
10
10 20 30 40 50 60 70 80 90 100 Recovery (% OOIP)
Figure 19.3-7 Carbonate Oil Pools
Cardium, is an Upper Cretaceous clastic pool with a primary recovery of II percent. The pool's recovery dominates the calculation of the weighted-mean recovery for clastics, and this explains the small difference in average recovery values for clastic pools. Considering the different depositional environments of carbonate and clastic reservoir rock, it is not surprising
242
c
RECOVERY FACTOR STATISTICS
that many ofthe parameters affecting recovery are quite different. For example, carbonate reefs tend to be quite thick. This provides the opportunity to complete wells in a fashion that avoids excessive coning of water or gas. Many are connected to active regional aquifers. With the vertical relief, gravity stable displacement by water is common. The porosity of carbonate rocks is generally low, in the 5-15 percent range, but permeability is very high, especially in dolomitized rock that is fractured. Clastic deposits are generally much thinner, and oil recovery relies on less efficient horizontal displacement processes that are controlled by viscous forces. In clastic rock, the granular nature of the matrix gives higher porosities; however, permeabilities tend to be lower through the combined effects of compaction, cementing, grain size distribution, and fines migration.
50,--------------, Pools = 625 OOIP (10 6 m3 ) = 1982 40
~30 ~
~
'"
'0 o
a,
20
10
19.3.7 Geological Period Recovery distributions are shown for eight geological periods in Figures 19.3-8 to 19.3-15. The.most significant variation in the shape of these distributions occurs in the Devonian Period. Average recoveries for both the Upper and Lower Cretaceous pools are under 10 percent and vary only slightly. The median recovery for the Lower Cretaceous pools is only 5 percent because about halfthe pools are heavy. Average recoveries increase to the 10-15 percent range as one moves down the stratigraphic column to the Jurassic and Triassic pools. Ofcourse, the sample ofpools drops to one-tenth that of the Cretaceous Period. Although there are too few Permian pools to be of any statistical significance, the average recoveries are higher there as well. The Mississippian Period breaks from the trend, with average recoveries dropping back slightly to the 6-13 percent range. Mississippian pools are generally considered to be in the carbonate family, but their recovery distribution more closely resembles the one for clastic pools. Average recoveries continue to increase in the Devonian Period to the 15-34 percent range. The Upper Devonian group has the pools with the highest recoveries, in particular, the Leduc (0-3) zone. Table 19.3-3 provides a more detailed breakdown by zone for the Upper Devonian. In the Middle Devonian, recoveries do not get as high, but the number of pools with recoveries in the 20-40 percent range increases significantly.
19.3.8 Geological Play Perhaps the best way to group pools and recoveries is by geological play. This approach is more meaningful because it honours regional geology (e.g., pools have a
Mean = 8.31 Weighted-Mean = 9.74 Median = 10.00 Standard Deviation = 5.06 Co-Variance = 0.61
10 20 30 40 50 60 70 80 90 100 Recovery (% OOIP)
Figure 19.3-8
50
Upper Cretaceous Oil Pools
Poois = 2585 OOIP (10 6 m3 ) = 2111
40
Mean = Weighted-Mean = Median = Standard Deviation = Co-Variance =
~30 ~
~
7.62 9.53 5.00 6.67 0.87
'"
'0 o
a,
20
10
10 20 30 40 50 60 70 80 90 100 Recovery (% OOIP)
Figure 19.3-9
Lower Cretaceous Oil Pools
similar depositional environment). Podruski et aI., (1988) used this approach in a Geological Survey of Canada paper, Conventional Oil Resources of Western Canada. They considered only light and medium pools and divided them into 78 plays. Table 19.3-4, taken from this report, lists average recovery factors for each play. The data were taken from provincial databases and
243
DETERMINATION OF OIL AND GASRESERVES
50,-------------~
50,-------;::--:-;----.., Pools = 7 00IP(10'm 3 ) = 11
40
40
Pools = 220 OOIP (10'm 3 ) = 162
Mean = 11.68 Weighted-Mean = 16.34 Median = 10.00 Standard Deviation = 8.57 Co-Variance = 0.73
~30
~
o'"o
0..
•
Mean = 11.00 Weighted-Mean = 21.61 Median = 10.00 Standard Deviation = 6.99 Co-Variance = 0.64
20
10
10
o
o o
o 10 20 30 40 50 60 70 80 90 100 Recovery (% OOIP)
Figure 19.3-10 Jurassic Oil Pools
Figure 19.3-12 Permian Oil Pools
50,---------------, Pools = 214 OOIP (10'm 3 ) = 228 Mean = 10.91 Weighted-Mean = 14.48 Median = 10.00 Standard Deviation = 5.98 Co-Variance = 0.55
50
40
-30 ~
o'"o
0..
10
Pools = OOIP (1 0'm 3 ) =
321 617
Mean = 7.96 Weighted-Mean = 13.01 Median = 6.00 Standard Deviation = 7.16 Co-Variance = 0.90
20
10
o
o 10 20 30 40 50 60 70 80 90 100 Recovery (% OOIP)
Figure 19.3-11
10 20 30 40 50 60 70 80 90 100 Recovery (% OOIP)
Triassic Oil Pools
included both primary and secondary recovery; therefore, the average recoveries will be slightly optimistic for primary recovery mechanisms. An acceptable way to make a preliminary recovery estimate ina new pool is by analogy. The success ofthe method depends on whether or not there are sufficient pools with reliable recovery data in the same geological play. Once the correct play has been identified,
10 20 30 40 50 60 70 80 90 100 Recovery (% 001 P)
Figure 19.3-13 Mississippian Oil Pools analogous pools may be found by comparing rock and fluid properties, pool geometry, and fluid contacts. Pools should also have a reasonable amount of performance history before the analogy is accepted. It is important to remember that this is only a preliminary estimate, and it must be confirmed with performance. If the early performance is inconsistent with the analogous pool, the evaluator should suspect that a different drive
244
c
RECOVERY FACTOR STATISTICS
50 r---------::-----:---=-:-.,.--, Pools = 706 OOIP (10'm') = 2292
50r------------------, Pools = 1239 OOIP (1O'm') = 897
40
40
Mean = 19.21 Weighted-Mean = 33.93 Median = 15.00 Standard Deviation = 16.47 Co-Variance = 0.86
Mean = 18.59 Weighted-Mean = 26.38 Median = 20.00 Standard Deviation = 11.65 Co-Variance = 0.63
~30
C en
(;
o
a.. 20
10
o
o
10 20 30 40 50 60 70 80 90 100 Recovery (% OOIP)
Figure 19.3-14 Upper Devonian Oil Pools
Table 19.3-3 Upper Devonian
10 20 30 40 50 60 70 80 90 100 Recovery (% 001 P)
Figure 19.3-15 Middle Devonian Oil Pools
Recovery Factors for Upper Devonian Zones OOIP
No. of Pools
(10' m3)
I
Pay (m)
Recovery Factor (fraction) Avg.
Min.
Max. W-Avg.
0.429
0.040
0.040
0.040
0.040
4.82
4.82
4.82
4.82
112
44.183
0.157
0.010
0.350
0.162
33.34
0.91
108.30
40.20
Stettler
I
0.053
0.200
0.200
0.200
0.200
2.80
2.80
2.80
2.80
Blueridge
5
2.432
0.096
0.010
0.200
0.125
17.12
6.43
30.20
15.27
24
15.233
0.125
0.010
0.360
0.167
6.61
2.10
16.20
9.18
178
342.719
0.243
0.010
0.650
0.405
16.17
1.00
90.35
21.55
11
1.992
0.134
0.010
0.250
0.145
7.48
3.40
10.36
7.95
2
0.742
0.110
0.070
0.150
0.105
9.29
3.05
15.54
10.05
148
812.328
0.291
0.010
0.750
0.568
16.17
0.90
135.64
42.40
2
0.541
0.125
0.100
0.150
0.106
4.66
3.70
5.63
5.39
53
944.084
0.142
0.010
0.350
0.164
9.89
1.62
37.00
18.79
Sulphur Point
133
108.224
0.119
0.010
0.450
0.132
7.14
0.88
27.90
7.99
Totals
670
2272.960
0.198
0.010
0.750
0.343
16.18
0.88
135.64
27.46
Devonian System Wabamun (D-1)
Arcs Nisku (D-2) Camrose Ireton Leduc (D-3) Cooking Lake Beaverhill Lake
Avg.
Min.
Max.
W-Avg.
245
. T
DETERMINATION OFOIL AND GAS RESERVES""'"
Table 19.3-4
Recovery Factors for Geological Plays in Western Canada
Geological Period
Play
and Plays
Depth (m)
Recovery Factors Average Small Pool
Cretaceous
Geological Period
Play
and Plays
Depth (m)
Recovery Factors Average Small Pool
Carboniferous
Cardium Sheet
1000
0.20
0.10
Midale
1400
0.36
Viking-Alberta
1800 2100
0.19
0.10 0.10
Frobisher-Alida Pekisko Edge
1100
0.24
0.20
0.10 0.10
Elkton Edge
1650 1700
0.10 0.10
Lodgepole Souris Valley-Tilston Banff Edge-C. Alberta
1100 1100 1500
0.13 0.28 0.16 0.16
0.07 0.10 0.07 0.11 0.12
0.06 0.09
Ratcliffe Stratigraphic Ratcliffe Structure
1800 1100
0.10
Desan Carb.-Sweetgrass Arch BanffEdge-S. Alberta
Lower Mannville Viking-Saskatchewan
500
0.15 0.14
Upper Mannville Belly River Shoreline Cardium Scour
1000 1000 2000
0:15 0.21 0.19
Cantuar Dunvegan-Doe Creek
1000 750
0.15
Belly River Fluvial Ostracod
1500 2200 2000
nd
1" & 2 White Specks
0.09 0.21 0.15 0.10
0.10 0.10
Debolt-Peace River Jurassic
0.20 0.20
700
0.20 0.08
0.10 0.10 0.Q3
2000 1300 1500
0.10 0.10 0.08
0.10 0.10 0.08
0.10 0.30 0.30 0.25
Devonian
Shaunavon
1200
0.23
Roseray-Success Gilby-Medicine River
900 2200
0.30 0.25
0.10 0.10 0.10
2950 Beaverhill Lake Leduc-Rimbey-Meadowbrook 1700 Keg River 1800
0.42 0.61 0.41
Sawtooth Rock Creek
900 2200
0.22 0.16
0.10 0.10
Nisku-Shelf
0.55 0.61 0.61
Triassic Boundary Lake Montney Peejay-Milligan Halfway Stratigraphic Inga Structure Charlie Lake Sandstone Halfway Drape Charlie Lake Algal Doig Structure Permian Belloy-Peace River Belloy-Erosional Edge
1300 1596 1130 2150 1600 1800 1900 1800 1900
1850 2000
0.27 0.17 0.32 0.28 0.15 0.18 0.19 0.15 0.08
0.28 0.37
0.10 0.12 0.20 0.12 0.10 0.18 0.10 0.15 0.05
0.12 0.10
1700 2000
Leduc-Bashaw Leduc-Deep Basin
2000
Nisku-West Pembina Middle Devonian Clastics Slave Point-Sawn Zama Leduc-Nisku-S. Alta Wabamun-Peace River
2800 1800 1600 1600 1700 1250
Slave Point-Golden Nisku-Meekwap Keg River-Senex
0.20 0.15
Wabamun-Crossfield
2000 2000 1300 2500
0.40 0.25 0.20 0.17 0.15 0.16 0.35 0.40 0.20 0.15
0.30 0.20 0.10 0.17 0.15 0.13 0.30 0.15 0.10 0.10
Bistcho Muskeg Wabamun-Eroded Edge Leduc-Peace River
1600 1600 2050 2800
0.15 0.20 0.14 0.20
0.07 0.20 0.14 0.15
Source: Conn and Christie, 1988. This table is reproduced with the permission of the Minister of Supply and Services Canada, 1993.
246
0.20
I
'
RECOVERY FACTOR STATISTICS
mechanism is operating and be prepared to gather or re-examine data to revise the recovery estimate.
100
Ideally, recovery estimates should fall out of a simple relationship between several reservoir parameters that are readily quantifiable. Unfortunately, this relationship continues to elude everyone. One possible explanation is that the methods used for quantifying these parameters over-simplify reservoir heterogeneity. Another is that the interplay of rock properties, fluid properties, drive mechanisms, and production strategies is too variable and dynamic for simple solution's to work consistently. Whatever the reason, simple correlations have yet to be found. To help illustrate the problem, recovery factors were plotted against three familiar reservoir parameters: porosity, net pay, and water saturation (Figures 19.3-16 to 19.3-18, respectively). The parameters used were pool average values. In each case, there is considerable scatter of the data, but some general trends are evident. For example, maximum recovery increases as porosity and water saturation decrease. Intuition might suggest the relationship between water saturation and recovery, but the one with porosity is less obvious. If the data is reexamined by rock type, clastic and carbonate, the reason is clear. Many of the higher recovery carbonate pools have lower average porosities and water saturations.
80
CONVENTIONAL GAS
The dominant natural drive mechanism that operates in most gas pools is fluid expansion. Recovery from a volumetric gas reservoir largely depends on how low a reservoir abandonment pressure can be achieved with the production facilities available. A simple correlation is used to estimate the abandonment pressure as a function of well depth. Normally a value around 1500 kPa per 1000 m ofdepth is used (Stoian and Telford, 1966). Two additional corrections are sometimes required to adjust for well deliverability, and fluid invasion at the wellbore. This fluid invasion can be oil, but can also be water either from an active aquifer or coning. Gas flows through reservoir rock far more readily than oil because gas viscosities are several orders of magnitude lower. In general, recoveries for gas pools without an active water drive are expected to exceed 75 percent and can reach 90 percent or more. On the other hand, if water invades a gas pool too quickly (e.g., with little or no pressure depletion), the residual gas trapped behind
..
;
ii:' 70
5 0
...
. . . . . oo •
60
.. . a: > ac
~
... .:::",: . .r., __ ·:"0 ...
~ 50 ~
\,'
..
. ..~,:-... .. ... ... .• . ''''~~''''-'.'' . _. ....i............
40
'.
30
..
'0 .....
_ ~
•
00
:
.. ~ •• _
. ._
. . . . . . . . . _00 • • • •
: . . .: . . : . ,. .: : .• •
,,+::··.;;f~-::,i:.:·;-,:,::- r : '. '''~'''''l;l1'''''''''''''' .~ ~~~·;r;;-H1~"''::': :.:.:.,..
20 10
"[,"1
•••.1.; -l-.......·~i W' • •....- ~j, {.
0
l.
'.-1,. ••
..". . .....!-.jjijjil:.:i i ...l.rj!~ r..
.
0
4
II
8 12 16 20 24 28 32 36 40 Pool Average Porosity (%)
Figure 19.3-16(a) Oil Recovery vs. Porosity
25 Pools = 5915 20
;g
15
'a.8'"
10
~
19.4
Pools = 5915
90
19.3.9 Recovery vs. Common Reservoir Parameters
Mean = Weighted-Mean = Median = Standard Deviation = Co-Variance =
14.42 14.16 13.00 7.44 0.52
5
o
4
8 12 16 20 24 28 32 36 40 Pool Average Porosity (%)
Figure 19.3-16(b) Porosity Distribution
the water-invaded zone can lower recoveries to 50 percent or less. The recovery distribution for some 9000 gas pools that have produced in Alberta is shown in Figure 19.3-19. As expected, recovery factors are in the 50 to 90 percent range. It is interesting to note that, unlike oil, the gas recovery distribution appears to be normal (e.g., symmetric about a mean value of 75 percent). About
247 d
DETERMINATION OF OIL AND GASRESERVES
100
100 Pools ~ 5915
90
a:-
(5
80
..
70
... .. , ..
" c: "
80
a:- 70
..
0 60 ~ e, ~ 50 > a 40
.. .. ._. '. .. °
0
.... 0
... 0
':~:""
0
(5
.. ._...... . ,
30.
0
••
0.
-oo
.'
'0,
'.,..r_•••__.:..:.... ':" ,0
.t: _\":..
.j:-=_ ,:........:..: :
'!!rfi-':' i '':.J~ .... :.
10
.-
. - - - - - -..- ..... -
0. . . . '
20
;0
.'
..
.. . -.. i" :" . ., . ...
; _ _.'!'.\O:.. ........ '_0",,",: •
0" ,0
•
" 1 '0'
• '0'
60
~
50
~
"1o; 40 c": 30
... . .. ..
-..
o
;ft.
.• 0. :
.0 ••
',1° •• • ::.. to ' . . . . ':.
~::•...;~....:; l~:' ~~:!:'.:::~..;:. :.~:: :.
20
•
~~':O .:7::vl.r...;:'>::....... 0:: 0: •
0
10
:
s
:
·:: ...•.. •.. ····,····· .. 1.. ••••.. ···1·....
Bllfu,;'u
25
Mean ~ Weighted-Mean = Median = Standard Deviation = Co-Variance =
10.20 15.15 4.40 15.86 1.55
Pools = 5915 20
~
C
15
Mean = Weighted-Mean = Median = Standard Deviation = Co-Variance =
29.08 21.74 30.00 11.78 0.41
en
~
2!
....:;:,.:;1.: ...... .1,::,' . .
Figure 19.3-18(a) Oil Recovery vs. Water Saturation
Pools ~ 5915
C
... i.. d,
..·f;:I·......•···.•.•. ·•.
Pool Average Water Saturation ('Yo)
60
40
,O,::L:.::.
oL...J=iilillli!illllil==iLhlli!.li....;~~~--' o 10 20 30 40 50 60 70 80 90 100
10 20 30 40 50 60 70 80 90 100 Pool Average Net Pay (m)
Figure 19.3-17(a) Oil Recovery vs, Net Pay
50
Pools = 5915
90
'0 o
30
a. 10
20
5
10
o o
a.
10 20 30 40 50 60 70 80 90 100 Pool Average Net Pay (m)
10 20 30 40 50 60 70 80 90 100 Pool Average Water Saturation ('Yo)
Figure 19.3-18(b) Water Saturation Distribution
Figure 19.3-17lb) Net Pay Distribution
(Figure 19.3-20), unlike oil recovery distributions, which are skewed to the right.
7600 of the pools used have an original gas in place under 300 X 106 m', A separate distribution was not included for this group, but it looks very similar to Figure 19.3-19. The remaining 1400 larger pools have a recovery distribution that is slightly skewed to the left
The same problem encountered with the reliability of reserves for small oil pools also applies to small gas pools. Warren (1990) discusses the impact of overly optimistic area assignments for Alberta's growing inventory ofsmall gas pools, and provides area defaults for use in setting the initial reserves of single-well pools.
248
____________________..iII
RECOVERY FACTOR STATISTICS
These area defaults are expected to improve aggregate reserves in a geological zone, but may not significantly improve the reliability of an individual pool's reserve.
19.5
USING RECOVERY FACTOR STATISTICS
It should be clear from the data presented that caution must be exercised in the use of recovery factor statistics. Nevertheless, for many new discoveries, there is 50 ,-----------"---,--,,..,...,-::--1 Pools = 9016 OGIP (10 9m' ) = 4143
Mean = 71.08 Weighted-Mean = 74.74 Median = 70.00 Standard Deviation = 9.26 Co-Variance = 0.13
40
~30 ~
~
so '0
o
a. 20
10
often no better alternative until performance data becomes available. The best approach is to find an analogous pool in the same geological play, preferably nearby the pool in question. If an analogous pool cannot be found, it may be necessary to look at an expanded data set; however, it should be understood that the reliability of recovery statistics decreases each time this is done. Recovery factor distributions will give an idea of the range of recoveries that are possible, and the probability ofdifferent recovery levels within the range. In cases where the range of recovery is small, the use of average recovery factors may be satisfactory. For new, light and medium pools that are small, a recovery of 10 percent, the mode, is a reasonable assumption for a solution-gas drive mechanism. It should be less if the pool is heavy. For a carbonate pool, there is more risk with the use of average recovery factors because they often have a combination drive at work, which can have a wide range of efficiencies. If aggregate reserves are being assessed, weighted-mean recovery factors should be quite reliable. The most important thing to remember is that recovery statistics should only be used for a preliminary estimate until more detailed analyses are possible.
References OL~-~~---"--
o
10 20 30 40 50 60 70 80 90 100 Recovery (% OGIP)
Figure 19.3-19
Gas Pools (Producing)
50,------------Pools = 1411 OGIP (10 9 m' ) = 3655 40
Mean = 74.89 Weighted-Mean = 75.24 Median = 75.00 Standard Deviation = 12.61 Co-Variance = 0.17
American Petroleum Institute. 1984. "Statistical Analysis of Crude Oil Recovery and Recovery Efficiency." API Bulletin, 014, 2nd ed., Apr. 30, 1984. Arps, J,J. 1956. "Estimation of Primary Oil Reserves." Trans., AIME, Vol. 207, pp. 182-191. Conn, R.F., and Christie, J.A. 1988. Conventional Oil Resources of Western Canada (Part II). Geological Survey of Canada, Paper 87-26, p. 131. Energy Resources Conservation Board. 1990. Alberta's Reserves of Crude Oil, Oil Sands, Gas, Natural Gas Liquids and Sulphur. Report ST 91-18, Dec. 1990, Calgary, AB. - - - . 1993. Catalogue - Publications, Maps and Services. Guide G-I, Calgary, AB.
10
oL_~_dJi!ib" o 10 20 30 40
50 60 70 80 90 100 Recovery (% OGIP)
Figure 19.3-20
Guerrero, E.T. 1961. "How to Find Ultimate Recovery and Performance of Oil Reservoirs." O&GJ, Vol. 59, No. 35, p. 101. McCrossan, R.G. 1969. "An Analysis of Size Frequency Distribution of Oil and Gas Reserves of Western Canada." Canadian Journal ofEarth Sciences, Vol. 6, pp. 201-211.
Large Gas Pools (Producing)
249
-
-.
DETERMINATION OFOILAND GASRESERVES
Podruski, lA., Barclay, J.E., Hamblin, A.P., Lee, PoI., Osadetz, K.G., Procter, R.M., and Taylor, G.C. 1988. Conventional Oil Resources of Western Canada (Part I). Geological Survey of Canada, Paper 87-26.
250
Stoian, E., and Telford, A.S. 1966. "Determination of Natural Gas Recovery Factors." JCPT, Jul.-Sep., 1966, pp. 115-129. Warren, A. 1990. "Alberta's Small Gas Pool Reserves." JCPT, Vol. 29, No.4, pp. 34-40.
PART FOUR PRICES, ECONOMICS, AND MARKETS
a
Chapter 20
OVERVIEW OF PART FOUR
Technical principles, supplemented by empirical data, form the basis for estimates ofpetroleum resources. An estimate of.reserves, on the other hand, is based on the principle that only a portion of these resources is economically recoverable. Assessment of economic viability requires information from areas such as financial analysis, regulatory guidelines, and market conditions. Part Four provides the basis for including these areas in the process of estimating economically recoverable reserves. The impact of these parameters on the reserves estimate is illustrated by an example from the Alberta oil sands. While the resource is estimated at 3800 x 106m3 , the reserves are only 280 x 106m3 (Energy Resources Conservation Board, 1991). Until the economics become more favourable through either reduced cost or increased revenue, the recognized reserves will not increase.
regulatory environment for the petroleum industry in Canada as an example. Reference is made to the Alberta Energy Resources Conservation Board and its role in defining production practices. Government policy initiatives, including tax and royalty regulations, have a major influence on reserves evaluations. One of the most important factors in any economic evaluation is commodity pricing. The two major commodities involved in reserves evaluations are crude oil and natural gas. Chapter 24 provides an overview of Canadian crude oil markets in the context of a globally traded commodity. There is particular focus on Alberta and other western Canadian provinces, including a description of transportation networks and major markets. The basics of price forecasting are examined, as well as price risk management and some of the products that are available to mitigate price volatility.
Reserves evaluation is based on analysis ofthe cash flow. Chapter 21 summarizes the major components ofa cash flow analysis using the Alberta (Canada) fiscal regime as a reference. The principal sources and uses of cash are addressed, including some ofthe details ofAlberta's provincial royalty regulations and federal corporate income tax. Some aspects of accounting and business finance that illustrate how these areas correlate with cash flow analysis are also discussed.
Natural gas markets in North America are addressed in Chapter 25 with particular emphasis on the dynamics imposed through changes in government regulations. The chapter focuses primarily on the Canadian market, but its virtual integration into the United States market has effectively blurred any boundaries between the two. Important demand forces are examined as well as various contract options that have evolved.
In all evaluations there is a degree of uncertainty. Indeed, it could be argued that every parameter involved in an evaluation is more accurately defined as a range of possible values. These uncertainties exist as a result of everything from physical measurement to changes in government regulation. Chapter 22 discusses the concepts of risk and uncertainty and presents criteria for identifying situations where risk analyses are warranted. Chapter 6 in Part Two discusses stochastic modelling of resource estimates. The impact of government regulations on reserves evaluations is discussed in Chapter 23, using the
References
The contents of reserves evaluations are available for a variety ofuses and users. These range from governments developing resource planning policy to companies contemplating production development or an acquisition. Chapter 26 decribes some of the uses of reserves evaluations.
Energy Resources Conservation Board. 1991. Alberta's Reserves ojCrude Oil, Oil Sands, Natural Gas Liquids and Sulphur. ERCB ST 91-18, Dec. 1990, Calgary, AB.
253
Chapter 21
CASH FLOW ANALYSIS
21.1
INTRODUCTION
In any industry, accurate analysis of cash flow is an essential part of investment decision-making and optimum capital budgeting. Like all resource-based industries, the oil and gas industry depends on such analysis to quantify its resource base; remaining reserves are actually defined in the context of economics. Methodology for cash flow analysis in the energy industry is consistent with the general principles of business finance. The purpose of this chapter is to discuss the cash flow analysis in the context of individual property analysis as well as in the corporate context, with particular reference to western Canada, and specifically Alberta.
It should be noted that the chapter is an overview. Parties utilizing cash flow analysis in contemplation of making business decisions are advised to retain professional advice.
21.2
MINERAL RIGHTS OWNERSHIP
Participants in the oil and gas industry lease the right to develop the minerals from the holders of the mineral rights. A number of different forms of property interest evolve from the leasing of these rights, with differences that are typically based on the sharing of risk and the requirement to provide development capital. Mineral Interest. Subsurface mineral rights are usually separate from surface rights. Ownership may be held privately, in which case it is known as a freehold interest, or by the government, in which case it is known as a Crown interest. Most of the mineral rights in Canada reside with the Crown; however, some do reside with individuals under freehold rights. These rights originate from two sources: (I) in 1869 the Hudson's Bay Company transferred to Canada what was to become the provinces of Alberta, British Columbia, and Saskatchewan, and the company retained some ofthe mineral rights; (2) in the nineteenth century, the federal government granted
certain mineral rights to railway companies as construction incentives. Royalty Interest. When a property is leased, the interest retained by the mineral rights holder is known as royalty interest. Freehold royalty interest would be contractually defined. Crown royalty interest is defined in legislation. The interest participant shares in the revenues, but has no obligation to fund development. Overriding Royalty Interest. This is an economic interest that is retained by a lessee when property rights are "farmed out" to another party. If the original lessee retains the economic interest without obligation to contribute to development costs, it is called a "gross overriding royalty" (GORR). If the original lessee allows the lessor to deduct certain defined expenses before paying royalty, it is called a "net overriding royalty" (NORR). Production Payment Interest. This is similar to an overriding royalty but would typically be limited to either a production amount or quantity, or to a certain time. Working Interest. This is the most commonly held interest. The holder receives the net benefits after the previously mentioned interests are realized and is responsible for development costs. Typically, there is more than one working interest owner. These owners would normally develop a property under the terms of a joint venture agreement which designates one ofthe owners as Operator. This arrangement is not usually considered a partnership as all the working interest owners are free to take their own production and dispose ofit on their own terms. Whether the business arrangement is a joint venture or a partnership can be significant from the perspective of calculating income tax. Carried Interest. This results when one party "carries" the development costs of another party. Working interests of the carried party would differ before and after payout of the carry amount. When the party has opted
254
c
CASH FLOW ANALYSIS
out of a development under an existing agreement, it will likely have to wait until an additional "penalty" amount has been earned by the other parties before reverting to a working interest position. Net Profits Interest. The net profits referred to will be calculated according to a contracted accounting procedure. In this case, the holder has no obligation to share development costs. Pooling and Unitization. These terms refer to arrangements made, voluntarily or in accordance with government regulation, to jointly develop a resource property. Pooling generally refers to agreements within a drilling or production spacing unit, these typically being one quarter section for oil and one section for gas. Unitization normally has a broader context and would address a development that extends beyond the standard spacing units. Individual participants can also undertake different conventional business arrangements such as joint ventures and partnerships. These structures also have tax implications that must be addressed.
21.3
PRINCIPAL SOURCES AND USES OF CASH
Ultimately, the revenue generated by the industry as a whole is through the sale of crude oil, natural gas, natural gas liquids (NGL) and sulphur. Estimated total revenue for the Canadian oil and gas industry for 1991 was approximately 18.4 billion dollars (Curran, 1992), ofwhich 61 percent was from crude oil, 27 percent from natural gas, 10 percent from NGL, and 2 percent from sulphur. These percentages are, of course, subject to change as product prices and relative volumes change. A lower average gas price in recent years has been partly responsible for the decease in the ratio of revenue from natural gas to crude oil from 0.54 in 1988 to an estimated 0.43 in 1991. Within this broader context of industry gross revenues, cash flows originate from a variety of other sources, including such things as overriding royalties and processing fees. Most cash flows are determined with the wellhead or lease as a reference point; however, there are other uses ofcash that have to be considered at the corporate level. Production Revenue. Custody transfer of product usually takes place at the lease boundary where the product is metered. Gross revenues are then the product of quantities sold and prices received.
It is important to differentiate between production revenue and processing revenue as the income tax treatment for these two items is different. In broad terms, production of crude oil stops at the exit of the battery while production of gas stops at the downstream end of the inlet separator of a gas plant. Revenue Canada uses the gas cost allowance calculation, from provincial Crown royalty calculations, as the basis for determining production vs. processing revenue for tax purposes.
Crude oil prices are typically posted by the buyer ofthe oil on a volume basis. The price is at a specific location, either in the field or at the refinery gate, and transportation tariffs must be deducted to bring the price to the wellhead. Future projections of price can be based on any number of approaches, ranging from a function of inflation to models that incorporate anticipated international supply and demand. Gas prices are, for the most part, set by buyer and seller at a price per energy unit. These prices will be set in reference to a delivery point such that a transportation tariff correction may be required to bring the price to the wellhead. Forecast future pricing will vary from fixed escalation to an index reflecting current market conditions and should include an awareness of market dynamics, regulatory issues, and the amount of contracted gas a buyer is likely to be able to purchase. Natural gas liquids is a term that includes ethane, propane, butanes and pentanes plus. Ethane, ifextracted at source, will usually be priced on the basis of a contract with the buyer. Propane and butanes (normal and iso), known as liquefied petroleum gas (LPG), are normally sold as a mixture with pentanes plus. The wellhead price of this mixture will be based on individual product prices with corrections applied for transportation and fractionation. Product price forecasts would incorporate the basics employed in crude oil and natural gas forecasting. Processing Fees. In most cases, production will require some treatment prior to sale (in a few cases, production is sold as is at the wellhead). Oil typically requires removal ofdissolved gas and produced water. Natural gas may require removal of water, heavier hydrocarbons, hydrogen sulphide, and carbon dioxide. If the producing company has its own treatment facilities, associated capital costs and ongoing operating costs will have been recognized in the economic evaluation of the property. If the producing company does not have the necessary facilities, it will have to incur an expense by paying another party to provide them. 255
DETERMINATION OFOIL AND GASRESERVES
Methods of determining fees range from "what the market will bear" to a contractually defined fee and may include minimum charges that must be paid irrespective of quantities processed. Some fees consist of two parts: a capital recovery component, and a separate operating component, which is a pro rata share of facility operating costs. A facility owner with spare capacity can make it available to other producers. In this case, processing fees become a source ofcash rather than a use ofcash. These are processing revenues and, as discussed later, should be included in taxable income after the resource allowance and earneddepletion deductions. (These two terms are discussed in Section 21.5). Another source of cash for one of the parties is the overhead fee paid by the owners of a facility to the operator of the facility. This fee will be specified in the operating agreement among the owners as a percentage of direct operating costs. Operating Costs. Field costs are typically some combination of a fixed cost, a unit cost per volume produced, and a monthly well cost. Disposal of produced water will sometimes involve off-lease disposal and an associated cost which can be based On the volume of water handled or can be a fixed fee per haul. Facility operating costs are generally considered to be a combination of fixed and variable costs. When facility ownership is shared, operating costs are usually apportioned on the basis of throughput or contracted commitments. Capital Costs. One of the more significant things to consider about a capital expenditure is the tax treatment of its components. This is addressed in greater detail in Section 21.5. Another consideration is funding. If a company can not finance its capital requirements from cash flow, it has to consider other sources, and the risk associated with the planned development can affect both the source and cost of this funding. When debt financing is contemplated, the development may not on its own qualify for the balance of needed funds, and additional security or additional equity may be required. Either alternative could restrict the company's options On further development. Other avenues available include the various leasing arrangements already mentioned. Site Restoration and Reclamation. This aspect of oil and gas operations has recently been receiving much more attention, and costs for decommissioning of
facilities and reclamation and restoration ofthe wellsite should be included in cash flow analysis. While thes~ costs may be accrued Onan annual basis for accounting purposes, they are not actually incurred until the end of the economic life ofthe producing property. (Quite apart from facility decommissioning and wellsite reclamation and restoration costs, there may be ongoing operating costs to satisfy regulatory requirements.) Well abandonment costs can be estimated on the basis ofpublished regulations. Decommissioning and site restoration costs depend in part on the final condition of the site and the regulations in effect at the time. As such they are more difficult to estimate. ' General and Administrative. Commonly referred to as G&A, these are the costs a company incurs in other than the direct operation of its properties, such as at a district or head office, and they must be recognized at some level in cash flow analysis. If any G&A costs are allocated to a property that is less than 100 percent owned by the company, they will then be shared by the other participants in that property. The effect of charging this expense to the field will be to marginally reduce field economics, with a corresponding impact on company economics. While the impact on the economics ofa particular field will likely be minor, if enough costs are allocated to enough fields, the impact on company economics can be significant. Interest Expense. Financial charges incurred in funding investments are often not included in cash-flow forecasts. From the perspective ofdiscounted cash flows, these charges, which include interest on debt or shareholder returns, have already been considered in the determination of the discount rate. If they were to be incorporated into discounted cash flow calculations, they would effectively be counted twice. From the perspective of annual cash flow, however, the amount of interest to be paid and, for that matter, principal, should be taken into account. This is particularly important at the corporate level because of the implications with respect to corporate liquidity. One area where interest must be considered is in the determination of income tax. Interest is tax deductible and, as discussed in Section 21.5, it is deducted after resource allowance and before earned depletion. Topgas. Derived from take-or-pay (TOP) gas, this refers to two agreements that were developed to alleviate problems associated with TransCanada Pipelines' contractual obligations to pay for natural gas which, because of the absence of markets, the company could
256
_______________________11
CASH FLOWANALYSIS
not take delivery of. In exchange for reduced TOP commitments, producers were pre-paid in excess of $2 billion. These amounts are being repaid over a lfl-year period, scheduled to end in 1994. A minimum of 10 percent of the gas recovery is made annually in the period from November to February, inclusive, that being the traditional period of highest gas deliveries. A producer's annual payments consist of what is in effect a combination of principal and interest, with the interest netted out of its gross revenue for gas.
It should be noted that the producer does not have to pay royalty on this gas until it is produced, even though the producer will have already been paid for the gas. Appropriate cash flow and accounting consideration must therefore be given this issue .. Working Capital. Changes in working capital are not generally considered in cash flow analysis at the field level; however, it must be recognized that a company will have to fund any increase in its working capital position. This is a particularly relevant item in a startup situation when funds for working capital have to be provided.
21.4
ROYALTIES AND MINERAL TAX
Generally speaking, Canadian provincial governments own and administerCrown lands within the provinces. The territories are the domain of the federal government, and the offshore is the domain of negotiation. Responsibility for mineral rights on aboriginal lands may lie with the aboriginal peoples. Royalties will be taken in cash or in kind and according to the regulations and formulae determined by the administrative authority. O'Dell et al. (1991) provides a summary of the fiscal regimes, addressing both tax and royalty, in the Canadian petroleum industry. The British Columbia royalty system is described in more detail in the British Columbia Oil and Gas Royalty Handbook (British Columbia Ministry of Energy, Mines and Petroleum Resources, 1992). The Saskatchewan royalty system is outlined in Statutes and Regulations (Saskatchewan Energy and Mines, 1990) and in information circulars published by the Economic and Fiscal Analysis Branch of the Department of Energy and Mines. Royalty regimes are administered by provincial governments and thus are subject to change at their discretion. This is best exemplified by the present situation in Alberta where recently announced changes have significantly altered royalty calculations.
As an example of a provincial royalty structure, a brief description of basic royalty calculations for the province of Alberta follows. This description encompasses the changes being implemented starting in October, 1992. Natural Gas Royalty. Royalty is basically a function ofage, price and production rate. Age refers to the classification ofthe gas as "old" or "new," new gas effectively being gas which is discovered or brought on stream after January I, 1974. Price refers to the "average Alberta market price" (AMP), as prescribed by the Minister and published in Department ofEnergy information letters. Price also refers now to "select price," for both old and new gas, again as published in information letters. Production rate refers to the average daily production for the month, with 16 900 m 3/das an amount below which ro yalty is reduced. Prior to October of 1992, Alberta Regulation 246/90 (Province of Alberta, 199Ia), as amended, was the basis for natural gas royalty calculations in Alberta, and details of the application of the regulation were provided in Gas Royalty Guidelines (Alberta Energy, 1990). With the introduction of changes to gas royalty calculations, these documents will require some changes as well. Schedule I of the regulation formerly described the calculation of royalty for natural gas and residue gas, and is presently being updated to reflect the royalty changes. Minimum royalty on natural gas, both old and new, is now 15 percent, the rate charged when the AMP is less than or equal to the select price. When the AMP is greater than the select price, the royalty rate, R%, is calculated according to Equation (I): R = [(.15)(GSP) + (.4) (AMP.GSP)] x where
~
AMP
(I)
R = the Crown royalty share (%) GSP = the old or new gas select price AMP = the average Alberta market price
Maximum royalty on natural gas is now 35 percent for old gas and 30 percent for new gas. Table 21.4·1 summarizes the changes to calculation of Alberta natural gas royalty as announced by the provincial government in October of 1992. A low productivity allowance is also available. If the average daily production during a month is less than 16900 m 3/d, the basic royalty is calculated according to Equation (2): 257
DETERMINATION OF OIL ANDGASRESERVES
Table 21.4-' Summary of Alberta Natural Gas Royalty Changes
Base Rate Marginal Rate Rate Cap (%) (%) (%) Current New Current New Current New New Gas Old Gas
22 22
15 15
30 40
40 40
30 40
30 35
Source: News release, Alberta Energy, October 21,1992.
R=R _[(R,-5)(16.9-P)'] c (16.9)'
(2)
where R., = the normal royalty payable P = the average daily production Effective January 1, 1994, calculations of Crown natural gas royalty share will be based on either the AMP or a corporate average price, the choice having beenmadeby the producer. Forpurposesofthis calculation, the corporateaverageprice cannot be less than 90 percent of the AMP. From Schedule 2 of the regulation, pentanes plus royalties are determined accordingto Equation (3): R = ::.:22'-'.(B=c),---+---,c,-,,(F,---=."B) F
where R B c F
= = = =
(3)
the Crown royalty share (%) the select price for the month the royalty factor for the month the producer's average selling price for the month
Royalties on pentanes plus are presently under review and changes are potentially forthcoming in the near future. From Schedule3 of the regulation, royalty payable on sulphur obtained by processing natural gas is 16'/3 percent of the sulphur. Sulphur royalties, unlike other royalty payments, are deductible against income in calculating federal income tax; however, sulphur revenues also do not qualify for resource allowance, as discussed in Section21.5. From Schedule 4, the percentage rate of royalty payable on any product obtained by processing natural gas and to which Schedules 1,2 and 3 do not apply is 30 percent of the product. This rate is presentlyapplied to propane and butane; however, with the royalty review, this rate is also subject to change.
In Alberta, gas royalties are taken in cash with the producer paying from the proceeds of sale. Gas cost allowance (GCA) is an amount deducted from the royalty obligationto account for the fact that Crown gas is being processed. When the gas is processed by a third party, the processingfee can be consideredthe GCA. If the producer is processing its own gas, the GCA is calculated according to an accepted formula which includesoperatingcosts, depreciationover remaining life and a 15percent return on averagecapital employed. A good descriptionof GCA and customprocessingfees is provided in Chapter 6 of the Gas Royalty Gn/defines (Alberta Energy, 1990). Crude Oil Royalty. Alberta Regulation 248/90 (Province of Alberta, 1991 b), as amended, was the basis for crude oil royalty calculations in Alberta, but it will have to be updated after the changes announced in late 1992. Crude oil royalties are a function of age, gravity, price and production rate. "Age" refers to the classification of oil as old, new, or third tier. "Old oil" is basicallyoil discovered prior to April, 1974; "new oil" dates from after March, 1974. "Third tier oil" was introducedas of October, 1992 and initiallydescribed as oil from"newly discovered pools," with any further definition to be containedin revisions to regulations. "Gravity"refersto the classification of oil as eitherlight, medium or heavy with the intention that heavy oil will be subjectto a lower royalty. Price adjustmentis based on a "par price" and a "select price," both as published in Alberta Department of Energy information letters. "Production rate" refers to monthly production rate. Crude oil royalty is calculated according to Equation (4):
R = [S + fS«A-B))] x 100
A
(4)
Q
where R S f A B
= the royalty rate (%) = the basic royalty = the royalty factor = the par price = the select price Q = the monthly production rate When the par price is less than or equal to the select price, the equation simplifies to Equation (5): S
R=-xIOO Q
(5)
258
_______________________.-sft
CASH FLOW ANALYSIS
There are two kinds of basic royalty: one defined for new and old oil and the other for third tier oil. In turn, the basic royalty for new and old oil is calculated by using one of two equations: one for production rates less than or equal to 190.7 m3/month and one for production rates over that value. The basic royalty for third tier oil has three values. It is equal to zero for production less than or equal to 20 m3/month and is calculated using one equation for production between 20 and 190.7 m3/month and another for production over 190.7 m3/month. Table 21.4-2 summarizes the equations used to calculate basic royalty. Table 21.4-2
Summary of Equations for Basic Royalty
Rateimvmon) 0-20
New Oil
TOld Oil
ThirdTier Oil 0
Q2 2755.04
(Q-20)'
20 - 190.7
[(Q-190.7) x 0.115385] + 13.2
Source: Newsrelease,AlbertaEnergy,October21,1992. The par price, A, is a representative wellhead price. There are now separate par prices for light and heavy oils for purposes of determining royalty rates. There is one select price, B, for old oil which applies to both light and heavy oil. There are two select prices for new oil with one set for new heavy and another for new light oil. Still another select price will be set for third tier oil. The royalty factor, f, is further identified as k for old oil, y for new oil, and z for third tier oil. These three factors are all calculated using Equation (6): k where r
,y, =
z = [(r%)(572.1) -I] / (A-B)] 57.2
[A
(7)
It should be noted that while the same formula is used for all types ofoil, there is a different maximum royalty intent set for each. This maximum is 35 percent for old oil, 30 percent for new oil, and 25 percent for third tier oil. The royalty rate, R, is price sensitive only up to the par price that causes "r" to reach its cap. Above this par price the royalty factor is reduced to maintain R at its maximum. Table 21.4-3 summarizes the changes to calculation of Alberta crude oil royalty as announced by the provincial government in October of 1992. Table 21.4-3
Summary of Alberta Crude Oil Royalty Rate Changes
BaseRate
Marginal Rate
Rate Cap
('Yo)
('Yo)
('Yo)
Current New Current New Current New
2207.46 > 190.7
r = (O.IB+O.4 (A-B)) x 100 A
(6)
the royalty intent (%)
The royalty intent is based on a well reference rate of 572.1 mvmonth and is calculated according to Equation (7):
Third Tier n.a. New Oil 212/ J Old Oil 212/ J
10 10 10
n.a.
30 40
40 40 40
n.a.
30 40
25 30 35
Source: Newsrelease, AlbertaEnergy, October21,1992. Royalty factors, par prices and select prices are to be listed in the Alberta Energy information letters. The Operator, as Agent for the Crown, is responsible for delivering Crown royalty crude oil volumes to the Alberta Petroleum Marketing Commission (APMC), a government-sponsored agency. The APMC markets the crude oil and, since the buyer of the APMC crude may not be the same as the buyer of the producers' crude, the price received for the Crown royalty volumes will not necessarily equal the producers' sale price. Royalty Tax Deduction. As discussed in more detail in Section 21.5, certain Crown charges, principally royalties, are not deductible in calculating federal taxable income. Instead, the provinces of British Columbia, Alberta and Saskatchewan make available a royalty tax rebate that is based on the difference between these Crown charges and resource allowance. Alberta taxpayers can deduct from Alberta tax payable an amount which is essentially the product of the provincial tax rate and the "attributed Canadian royalty income" (ACRI), the amount by which provincial levies exceed resource allowance. Any unclaimed
259
deduction can be carried forward indefinitely. Ifresource allowance exceeds the provincial charges, no royalty tax deduction is available; however, neither is there tax on the excess. Saskatchewan has a royalty tax deduction similar to Alberta's; the rebate is the lesser of Saskatchewan tax otherwise payable or the royalty tax credit. The tax credit for the year is a function of the Saskatchewan tax rate and the "adjusted attributed Canadian royalties and taxes" (AACRT). Unclaimed credits can be carried forward, and any excess ofresource allowance over Crown charges is not taxed. In British Columbia, a taxpayer first computes a basic tax, using resource allowance and nondeductible Crown charges. A notional tax is then calculated based on no resource allowance and deductible Crown charges. The difference between the two is the rebate, which is added or subtracted as an adjustment to the total tax payable. Alberta Royalty Tax Credit. As royalty is a government program, there is opportunity for government to make incentives available. One such incentive program is the "Alberta royalty tax credit" (ARTC). First implemented in 1974, it was updated as of January I, 1990 to a 5-year program providing a variable percentage tax credit. While there is now no specified limit on the refund itself, there is a limit of $2,500,000 on the amount ofroyalty base that is eligible for the credit in each year. The credit is a function of the "par price" of oil and is set quarterly by reference to average par prices in the preceding quarter. It varies from a high of 85 percent when the average par price falls below $100 per cubic metre, to a low of25 percent when the average par price rises above $210 per cubic metre. A number of amendments were made to the original program to limit the multiplication of royalty tax credits that could otherwise occur if a corporation that was claiming the maximum credit disposed of producing properties to a party claiming less than the maximum. Briefly, an above-limit, or restricted, corporation is one that has a royalty obligation in excess ofthe amount on which it can eam a credit. A restricted resource property is an interest in a producing property that was completed before 1989 and disposed of after 1989 by a restricted corporation. Royalties on production attributed to that interest cannot be included in the Alberta Crown royalty base for any of the holders of the interest. Also, as a general rule, the maximum allowable credit under the existing program must be allocated among corporations that are associated in a taxation year.
260
Further details on the present ARTC program are available from the Corporate Tax Administration groUp of Alberta Treasury. Production Royalty. Production royalty is defined with reference to the recipient. If the recipient is subject to Crown charges, such as Crown royalties, provincial minerai taxes and road allowance levies, i.e., nondeductible Crown charges for income tax purposes, the royalty is termed a production royalty. This definition is important for tax purposes because production royalty income is eligible for resource allowance (Section 21.5). Resource Royalty. Resource royalty is royalty received by a recipient; it is not subject to Crown royalty charges and is ineligible for resource allowance.
Oil Sauds Royalty. Royalty for oil sands development, such as the Syncrude operation, is usually determined according to contract terms negotiated between the developer and the provincial, and sometimes the federal, government. Freehold Royalty. Where mineral rights are held by a party other than a government, they are classified as freehold mineral rights, and the lands are generally referred to as freehold lands. Royalty obligations associated with production by other than the owner of the rights are negotiated between the lessor and the lessee. Mineral Tax. In the absence of ownership rights on oil and gas produced from freehold lands, and the concomitant right to impose a Crown royalty, governments impose a mineral tax, typically calculated on an annual basis. To illustrate, the following is a discussion of the Freehold Mineral Rights Tax as levied on production from nonCrown lands in Alberta. However, the mineral tax structure is presently being reviewed with a view to possible updating. This tax is a function of both price and rate, and is substantially lower than Crown royalty, to account for the fact that the producer is paying royalty to the owner of the freehold mineral rights. For crude oil, solution gas and condensate, the tax formula is: tax=RxVxM where R
(8)
tax rate: 0.269 for liquids, 0.069 for solution gas V = price per m3 for liquids, or 103 scm for solution gas M = annual production
CASH FLOW ANALYSIS
For solution gas, M is the production in thousands of standard cubic metres. For crude oil and condensate: M ~ (0.0833Q)' 105.94
Amounts in the accumulated CDE account may be deducted from taxable income at rates of up to 30 percent of the remaining balance.
For natural gas, ifaverage daily production for a year is less than 16.9 thousand standard cubic metres:
Canadian Exploration Expense (CEE). Defined in paragraph 66.1(6)(a) of the Income Tax Act, CEE is exploration cost incurred after May 6, 1974 and includes such things as geological, geophysical and geochemical expense, the drilling of exploration wells, and the cost of dry holes. A principal business corporation, as defined in paragraph 66(15)(h) of the Tax Act, must deduct the lesser of the amount in the account and the company's income for the year (exclusive of dividends from foreign and Canadian corporations'that are exempt from tax, and before any amount is deducted for depletion.)
tax s Ax Vx M
For all other taxpayers, deducting the full value of their CEE pool against their income is an option.
(9)
for annual production, Q, less than 2288.4 cubic metres, and: M
~(Q
) -228.04
(10)
4
for Q greater than or equal to 2288.4 cubic metres.
A ~ R _[ (R-.Ol)
where
(16.9-ADP)' (16.9)' ] X
(II)
(I 2)
R ~ tax rate; currently 0.069 ADP ~ average daily production per well
If the average daily production for a year is greater than or equal to 16.9 thousand standard cubic metres, the formula is the same as that for solution gas.
21.5
FEDERAL CORPORATE INCOME TAX
Tax rules are constantly being updated, either through legislative change or court interpretation, and tax planning is, at least in part, a function of corporate objectives. Consequently, planning and calculation of income taxes should be done with professional advice. With the rules of the game constantly changing, it is difficult to find an up-to-date reference for the Canadian tax system. Nevertheless, Krukowski (1987) provides not only a good overview, but also some useful background on the oil and gas industry. Canadian Development Expense (CDE). Defined in paragraph 66.2(5)(a) of the Income Tax Act, CDE is development-related cost incurred by the taxpayer after May 6, 1974. The cost is an intangible cost which, generally speaking, is expended in the drilling ofwells. It includes the drilling, completing or converting of any well that does not qualify as a Canadian exploration expense, the cost of recompleting a well after November 16, 1978, and the cost of any Canadian oil and gas resource property acquired before December 12, 1979.
Canadian Oil and Gas Property Expense (COGPE). Defined in 66.4(5)(a) of the Income Tax Act, COGPE is basically the cost incurred in acquiring a Canadian resource property after December 11, 1979. This is defined in paragraph 66(15)(c) of the legislation and can include drilling and production rights and royalty interests. Cumulative COGPE, the amount in the tax account balance, may be deducted at an annual rate of up to 10 percent ofthe balance in the account. Nontangible portions of property sales are charged directly to this account. If a negative balance results at year end, this balance is transferred to the CDE account. Any resulting negative balances created in the CDE account must go into income. Resource Allowance. With the exception of sulphur royalties, provincial Crown royalties are not deductible against federal income tax. While not explicitly identified as such in tax legislation, resource allowance exists as a means of recognizing this inequity. "Resource allowance" is a deduction against income and is calculated as 25 percent of adjusted resource profits (Table 21.5-1), using only production-related income and deductions. In an ongoing debate between tax authorities and taxpayers as to what constitutes "production-related income and deductions," it has been Revenue Canada's position that, for principal business corporations, all G&A expense is to be deducted in calculating resource allowance. This interpretation has now been successfully challenged, and the courts do not agree with Revenue Canada's interpretation [see The Queen v . Gulf Canada Ltd., 92 DTC 6123 affirming 90 DTC 6622(FCTD)].
261
Table 21.5-1
Cash Flow and Income Tax Summary
Gross Revenue Working Interest Production Royalty Deemed Income Expenses Crown Royalty Mineral Tax Production Royalty Lease Operating CrownLease Rentals CCA - Production G&A - Production CEDOE Adjusted Resource Profits (ARP) Resource Allowance (25% of ARP) Resource RoyaltyIncome Resource Royalty Expense CCA-Other-Resource Profit G&A-Other Interest COGPE CDE CEE Resource Profits (RP) Earned Depletion (25% of RP) Other ForeignIncome ForeignExpense Nonproduction Income Nonproduction Expense
Income Tax
Earned Depletion
Resource Allowance
Cash Flow
XX XX XX
XX XX XX
xx
XX XX
-
-
-
-
XX XX
-
yy
yy
yy
yy
yy
yy
yy
yy
-
yy
yy yy
-
-
-
yy yy yy
yy yy yy yy yy
-
yy yy
-ZZ
yy XX yy yy yy yy yy yy yy
yy XX yy
-
XX yy
-
-
yy
yy
yy yy yy
yy
-
-
yy
ZZ
yy XX yy XX yy
,
Net Income For Tax Purposes
-
ZZ
XX yy XX yy
Cash Flow before IncomeTax
ZZ
Source: University of Calgary and Canadian Petroleum Tax Society, 1991. Notes: XX represents an added amount. yy represents a subtracted amount. ZZ represents a sum.
262
,» , _____________________Fi1
CASH FLOW ANALYSIS
Table 21.5-1 is a simplified summary of the federal income tax calculation and the cash flow calculation for an oil and gas company. The four columns illustrate the calculation of income tax, eamed depletion, resource allowance and cash flow by identifying the parameters which are employed in the determination of each. Earned Depletion. While this item has been effectively eliminated for oil and gas producers, some companies still have an unclaimed earned depletion base that may be utilized as a deduction against income. A taxpayer is permitted to deduct the lower of the earned depletion, which would have existed under prior legislation, or the remaining base. The calculation of earned depletion is illustrated in Table 21.5-1. Capital Cost Allowance (CCA). This is the tax equivalent of accounting depreciation and in theory allows a business to recover its original tangible asset investment without having to pay tax on it. CCA accumulates in pools of prescribed classes which are deducted, at the option of the taxpayer, on the basis ofa fixed percentage of the declining balance. Tangible costs, which are grouped into CCA, should be differentiated from intangible costs, which are grouped into CDE and CEE. As a first approximation, tangible assets are located above ground, although they would also include production tubing and sucker rods. Production-related assets, which reduce resource allowance, must be differentiated from nonproduction assets, which do not. Again, as a first approximation, equipment which is upstream of an inlet separator is production-related. Production-related CCA is a deductible expense when calculating resource allowance, thereby reducing its effectiveness as a tax shelter for resource income by 25 percent. Accordingly, a taxpayer would be motivated to maximize not only the amount ofCCA which is deducted against nonresource income, but also amounts of COGPE, CDE and CEE. Disposal of a tangible asset yields a credit (not to exceed the original cost of the asset) for the pool into which the assets were originally grouped. If a negative balance in the pool results, this balance must be included in income. If the assets in question are productionrelated, this income will qualify as resource profits. Canadian Exploration and Development Overhead Expense (CEDOE). This G&A expense is not substantially directed toward exploration and development and may be completely written off in the current year, or capitalized. If capitalized, it would be deducted in calculating resource allowance in the year it was
incurred, added back in the income calculation, and then included in either CDE or CEE. Successor Rules. Alterations in a corporation's status, brought about by such things as mergers, acquisitions and changes in control, receive particular treatment within the Income Tax Act. A proper understanding of the associated rules and regulations is best left to professional advisors.
21.6
FINANCIAL STATEMENTS
Companies produce annual financial statements as an accounting of their performance during the year and their status at the end ofthe year. The information contained in these statements can yield historical annual cash flow numbers. Balance Sheet. If a company follows the full cost method ofaccounting, whereby all costs ofacquisition, exploration for, and development ofoil and gas reserves are capitalized, the value ofthe asset identified as "Property, Plant and Equipment" is limited by a "ceiling value." This ceiling value is effectively determined by performing a cash flow analysis on the company's reserves. It would include the value ofthe proved reserves plus the lower of cost and estimated value of undeveloped properties. "Depletion and Depreciation," listed on the asset side of a balance sheet, are accounting terms and are not equivalent to "Earned Depletion" and CCA, as used in the income tax calculation. On the liability side, anticipated future costs for "Site Restoration and Reclamation" are listed. These amounts are the company's estimate of future liabilities-at today's prices-and should be consistent with those used in the cash flow analysis, although they will have to be segmented into annual amounts and escalated to the appropriate year. Statement of Income. Revenue from petroleum and natural gas is usually net ofroyalties and includes ARTC. G&A, with the exception ofamounts capitalized, should be similar to that used in cash flow analysis, while, as mentioned previously, depletion and depreciation are not. Current income tax should correlate with that used in cash flow. Deferred income tax is a noncash item relating mainly to the timing difference between claims for tax purposes of CCA, exploration and development costs, and the amounts of depletion and depreciation listed in the financial statements. Statement of Changes in Cash Position. This statement can be derived from the balance sheet and the statement of income. Typically, the amount listed
263
'-1 "
DETERMINATION OF OIL AND GASRESERVES
as "cash flow from operations," when added to the interest expense listed in the statement ofincome, gives the cash flow being discussed. When interpreting these statements, the reader should also check to see how changes in working capital are addressed. Investments (such as capital expenditures and acquisitions) listed on this statement may have exceeded the company's cash flow for the year. In that case, the company will have had to either borrow money or get an injection of equity. These investments should have each been the subject of an investment decision process which would have included a cash flow analysis.
21.7
FINANCE AND ECONOMIC CONSIDERATIONS
Cash flow analysis and investment decision-making have a basis in theory, and to appreciate them some understanding of this theory is important. The following is a simplified discussion ofthe theory. For a more in-depth review, the reader is advised to consult a business finance text such as Lusztig and Schwab (1988). Net Present Value (NPV) and Internal Rate of Return (IRR). These are the two most widely used terms in investment decision-making. "Net present value" is the value obtained when all cash flow streams, including the investment, are discounted to the present and totalled. "Internal rate of return" is the discount rate which will give an NPV ofzero, meaning the discounted cash flow stream is equal to the cost of the investment. For investments involving initial expenditure and subsequent inflows of cash, a plot of NPV against discount rate yields a downward sloping curve which shows steadily decreasing NPV with increasing discount rate. This curve intersects the discount rate axis (NPV equal to zero) at the IRR. The apparent drawback of using the IRR is that it, by definition, assumes that the unrecovered investment can be re-invested at this rate. On the other hand, the NPV is expected to be positive, which normally implies that the IRR exceeds the cost of capital. When NPV is used, the investment is to provide a benefit beyond the cost of funding. When IRR is used, the yield is to exceed the cost of funding. In that respect, the two methods are complementary. Project Abandonment. Use of NPV as a decisionmaking tool should not be limited solely to the initial investment decision. Rather, a project should be checked throughout its life to ensure that it has a positive NPV. If at any point it does not, a sponsor should seriously consider abandoning the project since, from that point on, the investment will be incapable of generating its
funding costs. In this regard, the concept of "sunk costs" is introduced; monies that have been spent should no longer be incorporated into the investment decision. Weighted Average Cost of Capital (WACC). The "weighted average cost of capital" is the average aftertax cost to the company of all the components of its capital structure. These are not just loan interest costs but the cost of all forms of debt, including the cost of preferred shares and common shares. Lusztig includes internally generated funds, such as retained earnings and depreciation, when discussing a firm's WACC. All components should be included at their current cost since they will be used when making new investment decisions. The proportions of each can be based on the existing capital structure or a targeted new capital structure with total capitalization based on current market value. Discount Rate. By definition, an after-tax cash flow stream that is discounted at a firm's WACC and yields a positive NPV will pay for the project's funding costs and generate a residual gain for shareholders. In most situations, therefore, the appropriate discount rate to use is the WACC. The use of one discount rate for a firm's decisionmaking presumes that all ofthe firm's projected investments carry the same degree of risk. This may not be the case. Where a project is perceived to carry a higher risk, an investor would reasonably expect a higher yield. This would result in a higher WACC and a concomitant higher discount rate for the project. While theory suggests the derivation of a unique discount rate based on a project's WACC, other options are often employed. One common practice is to calculate the discount rate by adding a risk premium to the firm's normal WACC. This risk premium is usually based on intuition and is therefore, by its very nature, somewhat arbitrary. Nevertheless, it is often a practical way around the difficulties inherent in calculating a project WACC. Apart from the problems associated with determining a risk premium, there is normally some uncertainty attached to deriving any WACC, particularly the equity portion. This is one reason why, in actuality, the discount rate used is often the one that is in popular useage at the time, especially if two parties are negotiating a value.
264
___.-.a
!
CASH FLOW ANALYSIS
References Alberta Energy. 1990. Gas Royalty Guidelines. Alberta Energy Report, Dec. 1990, Pub. No. T/205·1990. British Columbia Ministry of Energy, Mines and Petroleum Resources. 1992. British Columbia Oil and Gas Royalty Handbook. Curran, R. 1992. "Slow Out of the Gate." Oilweek, Apr. 1992. Krukowski, J.V. 1987. Canadian Taxation ofOil and Gas Income (2nd ed.). CCH Canadian Limited, Don Mills, ON. Lusztig, P.A., and Schwab, B. 1988. Managerial Finance in a Canadian Setting (4th ed.). Butterworths, Toronto, ON.
0' Den, S., Pearse, J., Miller, c., and Tarvydas, R. 1991. Petroleum Fiscal Systems in Canada (rev. 3rd ed.). Energy, Mines and Resources Canada. Province of Alberta. 1991a. Mines and Minerals Act, Alberta Regulation 246/90. Office Consolidation, Queen's Printer for Alberta (amendments to 33/91). - - - . 1991b. Mines and Minerals Act, Alberta Regulation 248/90. Office Consolidation, Queen's Printer for Alberta (amendments to 31191). Saskatchewan Energy and Mines. 1990. Statutes and Regulations, Release No.9 (amended Ju\. 1991 and Sep. 1991). University of Calgary and Canadian Petroleum Tax Society. 1991. Taxation ofCanadian Oil and Gas Companies: An Introduction. Calgary, AB.
265
Chapter 22
UNCERTAINTY AND RISK IN RESERVES EVALUATION
22.1
INTRODUCTION
There is always uncertainty in an estimate of the volume or value of oil and gas reserves because few of the factors involved are known with certainty. The traditional deterministic approach does not make any allowance for uncertainty, and stochastic, or statistical, methods are required to assess it. Stochastic methods may be more time-consuming, but they make better use of available data and can yield important information that cannot be obtained from a deterministic evaluation. The degree of uncertainty can be of critical importance to investment and planning decisions, and an inadequate appreciation of it can lead to costly failures. For every evaluation, a decision has to be made as to whether the improved understanding resulting from a stochastic evaluation warrants the additional time that is required. The high cost of failure for most petroleum ventures suggests that stochastic methods should be used more than they are at present. This chapter examines concepts of uncertainty in the estimation of oil and gas reserves and discusses the aspects of statistics and decision theory that provide the methods for stochastic reserve assessments. Masters (1984) reviews the background of the approaches discussed in this chapter and emphasizes the need for common sense in their application.
22.2
CONCEPTS
22.2.1 Definition of Risk and Uncertainty The terms "risk" and "uncertainty" are used in many different ways, and caution is required when using them. In this chapter, risk is defined as the probability ofloss or failure and is relevant only in the context of decision-making; uncertainty is defined as the spectrum of possible outcomes of an evaluation. More complete definitions of various types of uncertainty are given in Section 22.2.3, and the relation of uncertainty and risk to probability distributions is illustrated in Figure 22.2-1.
266
22.2.2 Describing Uncertainty The uncertainty in a reserve estimate can be described in a number of ways, one of which is the use of the traditional terms, proved, probable and possible. However, there is no ready way of quantifying the level of differences expressed by such "point" estimates. Statistical measures such as ranges, standard deviations, confidence limits, and frequency, especially when shown graphically, convey a large amount of information that cannot be grasped readily in other ways and that is not given by point estimates. Expectation is the mean of all possible outcomes of an event and is a commonly used single-value summary measure that incorporates some of the effects ofuncertainty. It is often used as a decision criterion, but the following discussions on alternative approaches to decision-making are worth noting: Newendorp (1975a); McCray (I975a); and Tversky and Kahneman (1985).
22.2.3 Areas of Uncertainty Uncertainties arise in the following areas of reserves evaluation (Garb, 1988; Robinson, 1990): Technical Uncertainty, which can be further divided into the following: • Geological Uncertainty, which is concerned with the estimation ofhydrocarbon volumes in place. Once established, geological parameters are not usually changed significantly. • Engineering Uncertainty, which arises from the recovery process. Once engineering parameters have been established, significant changes usually occur only as a result of technical advances. Economic Uncertainty, which arises f;om market forces, and includes the major uncertainties in price, costs, taxes, and royalties. Economic uncertainty can be difficult to estimate because changes are usually less predictable than for the more stable technical areas. Political Uncertainty, which includes political aspects of local and national taxes, environmental regulations,
UNCERTAINTY ANDRISK INRESSRVES EVALUATION
(a) Probability Density Function (PDF) of Net Present Values (NPV)
0.4
Confidence Interval: There is about a 70% probability that the outcome will fall within this confidence interval.
Risk: Loss will occur in about 20% of the possible outcomes. Chance of Success = 80% 0.1
·100
a
-50 50 100 150 Net Present Value, NPV ($ x 10')
200
(b) Cumulative Distribution Function (CDF) or Expectation Curve of Net Present Values (i.e., the cumulative area below the frequency distribution curve in "greater than" form) 1 . 0 , - - - -__
~
c: Q)
::l
0.8
0-
~
u..
.~
1ij
0.6
:; E
0.4
o
0.2
::l
Loss ~--- - - - . Gain
a .L_ _~-~-_1_-_,_---,--__.;.::::::::O=r--100 -50 a 50 100 150 200 Net Present Value, NPV ($ x 10') Notes: 1. Uncertainty is represented by the fact that an outcome could fall anywhere on the NPV axis with differing probabilities. 2. Chance of success is 80%. 3. Risk of loss is 20%. 4. The mean outcome or expectation is $50 x 10'.
Figure 22.2-1
Risk and Uncertainty
267
•
DETERMINATION OFOIL AND GASRESERVES
market control, price control, and threats of nationalization, civil unrest, and war. Because political uncertainty operates ultimately through the same factors as economic uncertainty, political uncertainty may be regarded as an aspect ofeconomic uncertainty. However, the unpredictability and the potential for abrupt distortion of the market warrants the separate category. It is very difficult to quantify, and an assessment of several scenarios is often the best approach. . Uncertainty may also be classified as: Parameter Uncertainty, which is associated with the numbers used for an assessment, for example, porosity value taken as the average of core plug measurements. Model Uncertainty, which is a consequence of the degree to which a model used for the evaluation of reserves represents the real world. The effect is more likely to be one of "bias" rather than "error." This effect can be very significant, and may be difficult to assess. Examples of models used in reserve valuation are geological maps drawn assuming a particular depositional environment (e.g., beach sand or tidal channel sand?) and the algorithms used for log interpretation or for an economic evaluation (including the discounted cash flow model). Drew's (1990) account of the evolution of methods used for estimating undiscovered hydrocarbon volumes is a good illustration of the gradual reduction of uncertainty as progressively better models are adopted. The uncertainty in a reserve estimate decreases as production and knowledge increase until, at the time of abandonment, there is little or no uncertainty. Figure 22.2-2 is an idealized schematic representation of this. The range of reserves estimates is shown by the upper and lower limits of the estimates. As time passes and the well is produced, the range decreases and the limits converge until the range becomes zero, and they meet at the time of abandonment. In a real case, there would be: • A bias in the estimates • An asymmetry in the range of uncertainty • Changes in economics and technology over the life ofa project that would result in a curve not as smooth as this one
22.2.4 Causes of Uncertainty Reserve estimation is fundamentally a measurement procedure, and the relationships that exist between actual and estimated reserves and the associated
268
uncertainties can be summed up by the stochastic reserve relation: actual value of reserves = estimated value ± uncertainty where uncertainty
=
error ± bias
All of the factors in the relation will change with time and with time-dependent factors such as price and technology. The relation applies to all parameters involved in the assessment of reserves, and the individual uncertainties are combined according to established statistical procedures to give the uncertainty in a final result. Although it is usually not possible to separate "error" and "bias," an understanding of these concepts is essential to improving the quality ofreserve valuations. The effects of error and bias are shown diagrammatically in Figure 22.2-3. Actual Value is never known except, perhaps, at the time of abandonment of a property, as shown in Figure 22.2-3(g). Estimated Value, as shown in Figure 22.2-3(a) and (b), is determined by technical estimation procedures and economic evaluation and reported in reserves reports. Changes in technical and economic conditions result in changes to estimated and actual reserve volumes and values, even if the error and bias are zero. Error, as shown in Figures 22.2-3(e) and (t), results from the inherent uncertainty of measurement and analytical procedures, and can be positive or negative. The actual value lies at an unknown position within a confidence interval, the size of which is determined by the confidence level specified as shown in Figure 22.2-3(t). For example, "proved reserves are 250 ± 35 x 103 rn'" may mean that there is a 70 percent probability that they lie between 215 and 285 x 103 m3 . The probability of the actual value lying within a confidence interval of a particular size is represented by a frequency distribution (an envelope of all possible confidence intervals) as shown in Figure 22.2-3(e). Although errors cannot be eliminated, they can be minimized by careful technical work and quantified by statistical techniques. Errors also result from mistakes (e.g., arithmetic, clerical), but these are generally ofless importance and can be minimized by careful work and checking. Bias, as shown in Figures 22.2-3(c) and (d), is a systematic deviation from the actual value or distribution and is a combination of two effects. Campbell (1986) provides excellent examples of bias and other factors that can affect petroleum evaluations; the
•
UNCERTAINTY ANDRISK IN RESERVES EVALUATION
-Exploratlon - . - , - - - - - - Production
Actual Reserves
~
Q)
Q)
a:
<:
Ql
E <: o
"C <:
1l
Time
Analog - - - - - - .....~ - - - Volumetric - - - - - - - _ - Material Balance - - - - - - - ~ - - Decline Curves - - - - Source: Garb, 1988. Method of Determining Reserves
Figure 22.2-2
Level of Uncertainty in Reserves Estimates during the Life of a Producing Property
discussion and quotation that follow are from Spetzeler and Stael von Holstein (1975): • Displacement Bias is a shift ofthe whole frequency distribution curve to higher or lower values. This is shown in Figure 22.2-3(d). • Variability Bias is an alteration of the shape of a frequency distribution curve. This is shown in Figure 22.2-3(c). This is usually a central bias that makes a distribution narrower than is warranted (i.e., represents a greater degree of certainty than is justified by the state of knowledge). Capen (1976) convincingly demonstrates this tendency and suggests a method of minimizing the problem.
The following are the origins of bias: Motivational Bias, which is defined as "either conscious or subconscious adjustments in the subject's responses motivated by a perceived system of personal rewards for various responses. He may want to bias his response because he believes that his performance will be evaluated by the outcome. Finally, the subject may suppress the full range of uncertainty that he actually believes to be present because he believes that someone in his position is expected to know with a high degree of certainty what will happen in his area of expertise" (i.e., he wishes to appear more decisive than is really warranted).
269
,'"',"'''''''''0'''''''"'_ , Estimated Value (May not coincide with the peak of the frequency distribution)
(a)
Result of Deterministic Evaluation
-
Positive Displacement
(b)
(c)
+ Central Variability Bias
Result of Stochastic Evaluation
/Broadening"" " / (Less common) "
"
Variability Bias
Negative Displacement Bias , ... ,, , ,,
Positive Displacement Bias
~"'~
(d)
-
Displacement Bias
Error (Frequency Distribution)
(e)
Range of Uncertainty Introduced by Error
(f)
I
I
Actual Value
(g)
Figure 22.2-3
______t
Confidence Limits
_
Reserve Volume
The Effect of Error and Bias on a Reserve Estimate
270
c
UNCERTAINTY AND RISK IN RESERVES EVALUATION
• Cognitive Bias, which depends on an individual's mode ofjudgement. This arises from factors such as hislher knowledge base, method ofprocessing information (e.g., a judgement may be biased to a recent piece of information because it is the most easily recalled), or the representative nature of an analog used to make an assessment. Cognitive bias is probably an important source of model uncertainty.
I.
2. The comparison of projects in order to select the more appropriate one. For example, Figure 22.2-4 shows the expectation curves 1 for the evaluations of two projects with the same median NPVs of $100,000, but with very different risk profiles.
Specific, clear procedures, quality control, experience (i.e., a large knowledge base), competent technical work, the use of statistical techniques and third-party review, common sense, and a determined effort to maintain objectivity are all required to minimize the effect ofbias on reserves evaluation.
Project A is a low-risk venture that will not lose money, but has little chance ofmaking a great profit. Project B has a 25 percent risk that it will lose money, but it has a potential for a higher reward (e.g., a 20 percent probability of a net present value (NPV) greater than $250,000). Without the additional information provided by an analysis of the uncertainties, there is no objective way to choose between the two projects. Which of the projects is preferred will depend upon the risk acceptance level ofthe decision-maker and the budget available. This approach can also be used to analyze a portfolio of projects in order to avoid "Gambler's Ruin.'? With
22.2.5 Magnitude of Uncertainty The uncertainty in an evaluation ofhydrocarbon reserves depends on the particular property. However, for a single property in western Canada, at the start of production, the uncertainty in volume will typically be about ±25 percent. Uncertainty generally decreases as cumulative production increases and as more information becomes available (Figure 22.2-2).
100
A feeling for the magnitude of uncertainty in volume estimates can be gained from a study of the revisions in the reported proved reserves of 70 oil and gas companies over a period of7 years (Campbell, 1984, 1988) in which 86 percent of the companies displayed positive bias for oil (i.e., proved reserves were initially overestimated, and annual reductions were needed). The average annual reserve revision (mostly downwards) for oil was as follows:
o- 10%
for 72% of the companies
10 - 50% > 50%
for 26% ofthe companies for I % of the companies
\ _ _-
for 73% of the companies
10 - 34%
for 27% of the companies
Project A
~60 ~
.c 40
e
0..
Project B 20
-100 -50
0
50
100 150 200 250 300 3
Net Present Value, NPV ($ x 10 ) EMV'
Risk
3
Companies generally displayed neutral to negative bias for proved gas reserves. Average annual reserve revisions (almost equally up and down) for gas were as follows: 1-10%
Project evaluation, for which there would be an overall improvement ifthere were a better understanding of the risks involved.
Project A Project B
($ x 10 100 100
)
(%)
o 15
'Expected Monetary Value
Figure 22.2-4 Expectation Curves: Comparison of Results
22.2.6 Use of Uncertainty An appreciation of uncertainty and the associated risk of a reserve volume or value estimate is an important element in making decisions. Many ventures would benefit from a more thorough analysis that included estimates of uncertainty; for example, the elimination of one dry hole would justify a substantial amount of time spent on risk analysis. Other applications include:
An expectation curveis a cumulative distribution function showing the probability that a value on the x-axis will be exceeded. 2 Gambler's Ruin is the probability that in a series of ventures that will be profitable in the long run, a short run oflosses will exhaust the financial resources of tbe participants. 1
271
DETERMINATION OF OIL AND GASRESERVES
Sources of Funds
~--
Debt
.---~---------
Equity ---------------------
Low Risk Low Reward
High Risk High Reward
Banker
100
-«;
Development Engineer mell!
Development Geologist
80 ;? ~
>.
."" :0
60
Low-Risk Exploration (Western Canada Basin)
-" 0
~
40
High-Risk Exploration (Frontier)
0-
20
-Proven
Probable - - Possible - - Hydrocarbon Volume or Value
Figure 22.2-5
Expectation Curve: Reconciliation of Different Views of Hydrocarbon Volumes and Values
a strong enough budget, for example, the probability of Gambler's Ruin may be sufficiently low that a series ofhigher risk ventures like Project B can be attempted in the hope of a larger reward. 3. The reconciliation of different views of hydrocarbon volumes or values, arising from different levels of risk acceptance. This is illustrated schematically in Figure 22.2-5, which shows, for example, that the views of a banking organization, although different from that of a frontier explorer, are part of the same spectrum of possible results ofa venture. This figure is schematic and, in reality, there will be considerably more variation, but it shows the following:
4. The analysis of options for risk reduction. The strategies for this will be varied, for example:
• The probability ranges within which development, appraisal and exploration take place
5. The estimation of undiscovered hydrocarbon volumes on undeveloped lands. This is discussed further in Section 22.5.
• Typical levels ofactivity for various professional groups • Risk acceptance levels for different funding sources • Typical probability cutoffs for proved, probable, and possible reserves
272
• The acquisition of additional information (e.g., shooting more seismic before drilling a well) • Cost reduction as a result of spending more on project design • Forward contracts at a guaranteed price for productsales • Carrying out a project in partnership, rather than at full interest Some of these can be analyzed deterministically, but a stochastic analysis will yield a deeper level of understanding and consequent better decisions.
6. Classification of reserves. Although there is considerable debate on the definitions of various classes of hydrocarbon volumes, stochastic methods provide the only fully consistent approach. Without such an approach, there is only a limited
UNCERTAINTY AND RISK INRESERVES EVALUATION
understanding of the probability of recovering a quoted volume.
Institute), businesses whose purpose is the provision of this information and, ofcourse, internally generated data.
Further examples ofthe uses ofuncertainty can be found in several of the references cited.
Much of the data used in reserve valuation is obtained in quantitative form (e.g., well log data, production), and a wide variety of statistical techniques can be used for the assessment of the data. Although objective quantitative approaches should be used as much as possible, there will always be a major subjective component to any assessment. For data not available directly, and especially for geological parameters, analogy is particularly important. The selection of appropriate analogs is a critical element of the skill of a professional involved in a reserve valuation.
22.3
ESTIMATION OF UNCERTAINTY
22.3.1 Parameters to be Estimated Most of the parameters used to estimate reserve values are derived using a combination of subjective" and objective methods. All evaluations require ownership and fiscal information, but the technical parameters depend on the evaluation method being used. Volumetric evaluations require reservoir parameters (pay thickness, porosity, water saturation), drainage area, and recovery and formation volume factors. Produced volumes and pressures are needed for material balance and decline curve methods. More complex evaluations will require additional factors to be estimated, for example: • High, medium and low case maps or alternative interpretations, to estimate reservoir areas • A histogram of core porosities to represent reservoir porosities • A price forecast, with a spread of values at any particular time • Production decline curve parameters estimated by analogy with nearby wells • Market volumes that depend on predictions of weather and levels of economic activity • The availability of pipeline capacity • The probability of war or embargo • Tax levels The sources of data used to estimate uncertainty are the same as those for deterministic estimates although stochastic methods generally make better use ofthe data. The sources vary from proprietary to public. In Canada they include federal agencies (National Energy Board; Energy, Mines and Resources; Geological Survey of Canada), provincial agencies (Alberta Energy Resources Conservation Board; Alberta Petroleum Marketing Commission; and their equivalents in other provinces), business organizations (Canadian Association of Petroleum Producers; Canadian Energy Research " A subjective approach is essentially an opinion basedon previous experience, whereas an objective approach relies on the analysis of data (e.g., core data or previous well results).
22.3.2 Empirical Classification Time limitations mean that, despite the availability of more rigorous methods, most oil and gas volumes are classified as proved, probable and possible using a predominantly subjective empirical approach. Examples ofthis are the assignment of a one-section square drainage area for a gas well, or the classification as proved, of an undrilled spacing unit lying between two proved units. The major problem with this approach is consistency; what is reasonable to one person in one reservoir is not necessarily reasonable to another person, or even to the same person in another reservoir. An individual, or a group, may be consistent if clearly defined rules (i.e., in a "Reserves Manual") are prescribed and followed, but the results will almost invariably differ from other individuals or groups. Despite the advantages of the empirical approach, an inconsistent application of empirical rules is undoubtedly the source of many of the differences between reserve evaluations. When empirical methods are used, the probability associated with their recovery is, at best, poorly known. As an example, it is common to visually fit a straight line to the pressure decline in a gas reservoir, and extrapolate it to an abandonment pressure in order to determine the reserves. The value obtained in this way is usually called proved but, if quantified, it is often claimed to represent a 80 percent probability level (i.e., there is a 80 percent probability that a greater volume will be recovered). However, by definition, a best-fit, straight-line extrapolation will yield a value close to the mean (usually near a 50 percent probability level). This is a substantial inconsistency that is probably present in many, if not most, gas reserve estimates. Similar inconsistencies occur for other empirical approaches used for both oil and gas reserves evaluation.
273
-,
~I DETERMINATION OFOILAND GAS RESERVES~"
The criticism must be placed in perspective. A full-scale stochastic exercise can be time-consuming and is often neither practicable nor necessary. Sometimes, for instance, a reserve classification is not required, merely the assurance that a particular cutoff value or volume will be exceeded. The empirical approach is common because it is relatively easy to apply and, in many cases, will give an adequate answer. However, it should not be used carelessly or uncritically, and more sophisticated methods should be used when warranted.
22.3.3 Quantifying Subjective Estimates A subjective estimate is essentially the opinion of the person making the estimate. Although it depends ultimately on this person's expertise and objectivity, some measures can be taken to improve the quality of a subjective estimate. The Delphi Method uses the consensus of a team of "experts" to generate the required data. Estimates ofthe probability distributions of the parameters are made independently and perhaps anonymously by the experts and combined either by averaging or by consensus. The opinions ofthe experts can be weighted (that ofthe "expert in a related field" receiving the greatest weight in some schemes), and a number of iterations can be made. The Delphi method reconciles different opinions, including quantitative estimates, and the methods that the experts use can vary from entirely subjective to highly statistical. Familiarity with a problem will often allow the direct subjective estimation of a frequency distribution, and questions such as, "What are the maximum, minimum possible, or most likely values?" or "Is it likely to be log-normally or normally distributed?" are helpful. Several types of distribution can be used, although the normal, log-normal and triangular distributions, histograms, and some discrete distributions will cover most cases. These distributions are described in most statistical books and, more specifically in the context of reserves evaluation, in Newendorp (1975a) and McCray (l975a). A graphical sketch of a frequency distribution can sometimes be made; interactive graphical computer displays are particularly useful for this purpose. Subjective estimates can be "disciplined" to some extent. For instance, if the distribution is considered to be normal or log-normal, then an estimate of the probability confidence interval corresponding to a particular range (or vice versa) can be made and plotted on the appropriate probability paper (e.g., if data is
considered to be log-normally distributed, and the values estimated cover ±20 percent on either side ofthe median, then the high at 70 percent and the low at 30 percent are plotted on log-normal paper). From the line drawn through these points, the range at other levels of probability (e.g., at 90 percent to 10 percent) can be read and a decision made as to whether it is reasonable' ifnot, revisions can be made. Tests have shown that the• range at a particular probability level is usually underestimated (i.e., there is usually a central bias). This method is described by Capen (1976), whose paper should be consulted for details. Qualitative expressions such as "good chance of," "low risk," "very unlikely," or "probable" may be adequate for everyday use, but the lack of a common standard means that they are oflimited use for describing uncertainties in reserve estimation. Attempts have been made to interpret these terms quantitatively, and a useful summary is given by Mosteller and Youtz (1990). It is interesting to note that their study showed that different perceptions of the word "possible" are so varied that the word is virtually useless. A table presented by Kadane (1990) in a comment on the paper by Mosteller and Youtz is a useful codification of terms that can be used as.a guide to quantifying qualitative expressions. It is not ideal for oil and gas volume estimation, and questions would have to be framed appropriately, for example, "Will the porosity fall in the range of 10 to 12 percent?" This is an active area of statistical research, and improvements may be expected. Range of Probability (%)
oto 5 5 to 15 15 to 25 2'5 to 35 35 to 45 45 to 55 55 to 65 65 to 75 75 to 85 85 to 95 95 to 100
Verbal Description Almost never Seldom Infrequent Sometimes Less than an even chance Even chance More often than not Often High probability Very high probability (Virtually) certain
22.3.4 Quantitative Estimation The quantitative determination of uncertainty us.es the concepts of statistics and probability. Details on methods mentioned here and also on other methods
274
______________________..n 4~_~
UNCERTAINTY ANDRISK INRESERVES EVALUATION
(e.g., Bayesian statistics," time series, sampling, regression) may be found in statistical texts and in Megill (1984 and 1985), Newendorp (l975a), and McCray (1975a), who describe their use in the evaluation of petroleum projects, and more generally in Rock (1988) and Davis (1986). Methods of determining factors such as reservoir volume, petrophysical parameters, reservoir volume factors, and production forecasts are described in other chapters in Parts Two and Three. From the point ofview of estimating uncertainty, the traditional deterministic approach to these factors must be expanded to generate the required distributions, ranges, and high-mediumlow values, primarily using the methods of classical statistics. Alternative maps (e.g., high, medium and low case) can be drawn to derive some of the geological parameters needed. Geostatistical methods that incorporate spatial relations have recently become available. These methods generate a weighting function that is used to interpolate or extrapolate reservoir parameters and also to provide an estimate ofthe uncertainty. The resulting data is relatively objective and is particularly useful for applications such as unitization or building models for reservoir simulation. Details can be found in Clark (1979), Hohn (1988), and Isaaks and Srivastava (1989). Estimates of uncertainties in costs rely on subjective estimates, analogy, engineering analysis, and bid quotations. The type of estimate will depend on the evaluation scenario that is adopted and, in some cases, contingency costs and associated probabilities are required. The time that production starts or the phasing of expenditure in a major venture can affect the economic viability of a project. Improved estimates may require the use of techniques such as Critical Path Method (CPM) and Program Evaluation and Review Technique (PERT). CPM is a deterministic method for which high, medium- and low-case estimates can be generated, while PERT is a probabilistic technique that generates a frequency distribution (McCray, 1975a). Product pricing is usually the most important factor in the valuation of an oil and gas project. The forecasting of oil and gas prices is notoriously difficult, and methods range from purely subjective "guesses" to sophisticated, analytical probabilistic models that may include the effects of weather and levels of economic
• Bayesian statistics considers the idea of conditional probability in whichthe probability of an eventdepends on preceding events, as in decision tree analysis.
activity. Alternative scenarios can be used to assess the effects of different price forecasts, although a proliferation of scenarios can make the results meaningless.
22.4 METHODS OF ANALYSIS 22.4.1 Carrying Out a Stochastic Evaluation Stochastic evaluation methods use values that are expressed by probability distributions, not by single values. The approach taken for a particular evaluation depends on the magnitude of the expenditure, the data, the time and expertise available, and also the environment in which decisions are made. There are no hard and fast rules that prescribe the use of a particular method but, in general, the less familiar and the more complex, expensive or risky a venture is, the more sophisticated an evaluation will need to be. At least one scenario must be established for every project, and several scenarios may have to be constructed to examine sensitivities or to determine the most profitable course of action. Scenarios should be constructed with care as the same activities carried out under different scenarios can yield different results. For example, when several wells are to be drilled, the order of drilling and timing can make a difference to the probable outcome, as could a decision to reduce the risk by shooting seismic. Although scenarios can vary greatly, there are usually a number of common steps in an evaluation. The following steps (modified after Megill, 1984) assume that a decision has been made to assess uncertainties and carry out a stochastic evaluation: I. Collect data. The old adage of "garbage in, garbage out" is relevant, and time spent on ensuring that necessary data has been collected and is of good quality is usually well spent. 2. Isolate the key variables. Which parameters contribute most to uncertainty? Trial runs may have to be carried out. It is always better to spend more time on the assessment of a critical parameter than on a less important parameter.
3. Decide on the scenario and on the types and parameters of the distributions (high-medium-low, triangular, log-normal) and the method to be used (e.g., decision tree, Monte Carlo). Several scenarios may be evaluated and sensitivities investigated in order to optimize a project or reduce risk. 4. Carry out the evaluation. 5. Ask "Does the result make sense?" Ifnot try again. 275
DETERMINATION OF OILANDGASRESERVES
6. Express the answer in the fonn of a cumulative distribution (expectation curve) or probability density function although a single-value answer (e.g., an expectation or a cut-off value) may also be required. A sensitivity analysis, which shows the effect of variation in individual parameters, may also be appropriate.
The following matrix contains the calculation for the factor Soh x area: Area Net Oil (m)
Low
Value Probability
50
Not everyone has the capabilities for sophisticated simulation procedures, nor does every project warrant such an approach. In many cases, a simple manual or spreadsheet method of calculating an expectation can be used. A decision matrix is a simple method of combining probabilities that can be used when a computer program is not available. The simple example given here is the calculation of an oil-in-place expectation from three parameters: net oil column (Soq,h), area, and recovery factor). High, medium and low case estimates and their associated probabilities are as follows: Net Oil Column (Soq,h) in metres Estimate Probability Low 0.5 0.30 Medium 1.5 0.60 High 2.5 0.10
Medium
High
75
100
300
0.10 125
0.18 300
0.15 150
0.2
2.5 0.6
0.09
0.5 High
1.5
0.03 500
0030 450
0.06
0.05 750
0.12
0.02
Checksum: Sum of probabilities = I Nine values of Soh x area and associated probabilities result from the calculation. New high-medium-low case values are generated by combining the three lowest (25, 75, 100), the three medium (125, 150, 300), and the three highest (450, 500, 750). These may not be in the same row or column of the matrix. If preferred, reservoir parameters and probabilities may be laid out in separate matrices or programmed using simple matrix algebra. Low Case (25 x 0.09) + (75 x 0.18) + (100 x 0.15) 0.09 + 0.18 + 0.15
Probability 0.30 0.50 0.20
Recovery Factor Low Medium High
Medium
0.30 25
200
Area in square metres x 103 Low Medium High
0.5
OJ
22.4.2 Decision Matrices
Estimate 50 200 300
Low
(m'x 103 )
=73.214x 10'm' with a probability of 0.09 + 0.18 + 0.15 = 0.42 Medium Case (125 x 0.03) + (150 x 0.06) + (300 x 0.30)
Estimate
Probability
0.10 0.20 0.25
0.30 0.40 0.30
Calculations are performed using a matrix layout with one parameter and probabilities across the top and one parameter down the side. Each cell of the matrix contains the value of the parameter at the top left and the probability at the bottom right. The products ofthe parameters and probabilities are placed in the appropriate cell of the matrix. This method may be used for more than three-point (high-medium-low) estimates, but becomes more laborious.
0.03 + 0.06 + 0.30 = 263.462 x 103 m' with a probability of 0.03 + 0.06 + 0.30 = 0.39 High Case (450 x 0.12) + (500 x 0.05) + (750 x 0.02) 0.12 + 0.05 + 0.02 = 494.737 x 10' m' with a probability of 0.12 + 0.05 + 0.02 = 0.19 These values are entered into a matrix with recovery factor (RF) as the other parameter:
276
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _1
UNCERTAINTY AND RISK IN RESERVES EVALUATION Recovery Factor
NetOilx
Value
Low
73.214
Medium 263.462
High
494.737
Area
(10' m')
Probability
Low
0.10
7.321 0.3
Medium
0.20
High
0.25
0.117 52.692
0.168 18.304
0.3
26.346
0.126 14.643
0.4
0.390
0.420
0.156 65.865 0.117
0.126
0.190 49.474 0.057 98.947 0.076 123.684 0.057
Checksum: Sum of probabilities = I If the result of this calculation is to be used for further calculations (e.g., for economic high-medium-low calculations), new high-medium-low values would be generated: Low Medium High
13.545 x 10' m' 42.795 x 10' m' 89.105 x IO' m'
Probability 0.42 Probability 0.33 Probability 0.25
However, if this is the end point of the exercise, an expectation can be calculated: (13.545 x 0.42) + (42.795 x 0.33) + (89.105 x 0.25) = 42.088 x 103 m 3 While this method does not have the sophistication of a full stochastic simulation and requires some simplifying assumptions, it will usually provide a reasonable answer.
22.4.3 Decision Trees A decision tree is a graphical summary of the possible outcomes and probabilities of the events that comprise a project. It is a powerful analytical tool that allows the calculation of expectations and various risk-related parameters. There are several types, varying from simple trees with the decision nodes absent, to trees with stochastic decision nodes. The type chosen will depend on the particular problem being investigated; a simple tree that can be solved manually will suffice for most problems. McCray (1975a) and Newendorp (1975a) provide a detailed discussion ofthe construction and use ofdecision trees.
22.4.4 Probabilistic Simulation Probabilistic simulation (often referred to as Monte Carlo computer simulation)' is the combination of frequency distributions ofvariables in order to produce the frequency distribution of a final outcome. Decision matrices and most trees are relatively crude approaches to combining distributions, and this can be done much more thoroughly using simulations. Analytical approaches have also been developed that, under the right conditions. produce a similar result to simulation. They depend on the transformation ofthe frequency distributions of the various parameters to normal (or log-normal), the mean and variance of which can be. easily manipulated. Care must be taken to ensure that the transformations are valid, as significant errors can be introduced if they are not. For simulation, frequency distributions are generated for the significant parameters, a value is randomly selected from each one, and a calculation is carried out using these randomly selected values. The process is repeated many times (typically 300 - 1000), and the result is presented as a frequency distribution or an expectation curve. The most common method of selecting a random value is Monte Carlo sampling; the Latin Hypercube method/ has computational advantages, but is less commonly used. Programs of various levels of sophistication have made simulation a much easier
process.' "Monte Carlo" is a probabilistic simulation method. Probabilitydistribution functions are prepared for the parameters in an evaluation and, using random numbers generatedby a Monte Carlo (or similar) sampling process, values are selected from the distributions. A calculation is carried out using the selected values, and the process is repeated many times (typically 300 - 1000). The resulting values define a probability distribution function from which parameters such as median, mean, and mode may be determined. 2 Latin Hypercubeis a method of sampling a probability distribution by a stratified random sampling process. It performs the same function as Monte Carlo sampling, but with fewer samples required for the same result. 3 Computerprograms are commercially available (Palisade Corporation's sophisticated stand-alone PRISM, or spreadsheet add-on, @RISK) or can be found in the literature (McCray (1975b) p. 215, gives a program in FORTRAN: Garb (1988) presents a simple Monte Carlo program in BASIC; Crovelli and Balay (1991) describe a PASCAL program that is available from the USGS). 1
277
DETERMINATION OFOILAND GASRESERVES
The parameters being simulated must be mutually independent (e.g., pay thickness should not depend on porosity), or results may be seriously wrong. Methods of handling dependent parameters include combining them (e.g., using tbe product h rather than and h separately) or setting up dependencies as part of the simulation. This facility is available in some of the programs. Simulations can be carried out for different purposes, for both technical and financial reasons, and at different levels of complexity. It is possible, for example, to simulate a gas field development project that, in addition to the technical aspects, includes tbe possibility of different market levels or the impact of an embargo on price. A more detailed discussion of simulation for project evaluation is given in Newendorp (1975b) and McCray (1975c).
22.5
EVALUATION OF UNDEVELOPED LANDS
Uncertainty plays a major role in the estimation of undiscovered hydrocarbon volumes and their values. A variety of methods is available, many of whichespecially the statistical approaches-are still under active development. Details and further references can be found in Haun (1975); McCray (1975d); Newendorp (1975a); Megill (1984, 1985), Masters (1984); Rice (1986); Drew (1990); and Campbell (1970). Estimates of undiscovered hydrocarbon volumes are required at scales ranging from poorly known basins to single well offsets in known pools. The method adopted will depend on tbe scale and the information and time available for making the estimate. Most estimates will be for relatively small projects using a subjective estimate based on analogy. Larger projects will usually warrant the use of a more sophisticated approach. The methods available can be summarized as follows (after Miller, in Rice, 1986): Areal and Volumetric Yield Methods with Geologic Analogy. The area or volume of the petroliferous sediments in an unknown area is multiplied by the volume of hydrocarbons per unit area or volume in a known analogous area. This method depends critically on the identification and the validity of an appropriate analog. It is always difficult to know how good an analog is, and the result is uncertain, usually to an unknown degree. Areal and volumetric methods are of the most use when there is little other information, but once
addi~idonal informI' abtilon is available, other methods will provi e more re ra e results. Delphi or Subjective Consensus Assess " 10 Section 22.3.3. ment M e th 0 d s, ThiIS approach iIS descnbed Historical Performance or Behaviouristic Meth d These methods are based on the extrapolatio 0 Sf' hiIston.ca . I data, such as discovery and drilling ratesnand 0 field sl~es. The data are entered into mathematical models WhICh are then used to make extrapolations. D (1990) gives an interesting account of their evolut~W · ~ an d new wor k continues to appear. Geochemical Material Balance Methods. These methods estimate the volume of hydrocarbons generated, the volume involved in migration and loss, and the volumes trapped and lost. By their nature, geochemical material balance methods are useable only on a relatively large scale. Considerable information and an appropriate model are required for this method to be successful, and it has had limited use (Sluijk and Parker 1986). ' Combined (Integrated) Methods. Combinations ofthe above methods, often with sophisticated statistical and mathematical models, are becoming more common. In general, they involve the following: • Geological basin analysis • Play or prospect analysis techniques • Statistical, economic and supply projection models • More comprehensive petroleum province analog systems The Geological Survey of Canada (GSC) review of petroleum potential (Podruski et a!., 1987) used two approaches to the estimation ofremaining undiscovered volumes in western Canada. The discovery process model described by Lee and Wang (1983, 1985, 1986) and also described in a less mathematically daunting way by Drew (1990) is a statistical model that assumes that discoveries made to date represent a biased sample of the underlying pool population. To understand the characteristics of the population and make predictions, the discovery process is modelled, using the pool size and sequence of discovery. Economic cutoffs can be built into the model to determine the undiscovered volumes at various price levels. Both Podruski and Drew claim this to be tbe most reliable method. The second approach used by the GSC was a subjective probability model, using probabilistic (Monte Carlo) simulation. This may incorporate an assumption that
278
s
UNCERTAINTY AND RISK IN RESERVES EVALUATION
the underlying pool population has a log-normal distribution.
Davis, lC. 1986. Statistics and Data Analysis in Geology (2nd ed.). Wiley, New York, NY.
A recent series of papers, in a thematic issue of the AAPG Bulletin in 1993, which includes papers by Masters (1993), Houghton et al. (1993), Drew and Schuenemeyer (1993), Root and Mast (1993), Root and Attanasi (1993), Attanasi et al. (1993), provides an extensive summary ofthe current practices ofthe USGS on petroleum resource assessment. It is interesting to note that one of these papers (Houghton et aI., 1993) recommends the use of a modified Pareto distribution, as being better than the more traditional log-normal distribution for modelling pool sizes.
Drew, LJ. 1990. Oil and Gas Forecasting: Reflections ofa Petroleum Geologist. International Association for Mathematical Geology Studies in Geology, No.2, Oxford University Press, Oxford, UK.
The development ofmethods ofestimating undiscovered reserve volumes and values is an active field, with new papers continuing to appear in the literature.
References Attanasi, E.D., Bird, KJ., and Mast, R.F. 1993. "Economics and the National Oil and Gas Assessment: The Case of Onshore Northern Alaska." AAPG Bulletin, Vol. 77, No.3, p. 491. Campbell, A.D. 1984. "An Analysis of Bias and Reliability in Revisions of Previous Estimates of Proved Oil and Gas Reserve Quantity Information: Replication and Extension." Petroleum Accounting and Financial Management, Summer 1984. - - - . 1988. "An Analysis of Bias and Reliability in Revisions of Previous Estimates of Proved Oil and Gas Reserve Quantity Information: An Update." Petroleum Accounting and Financial Management, Spring 1988.
Drew, LJ., and Schuenemeyer, lH. 1993. "The Evaluation and Use of Discovery Process Models at the US Geological Survey." AAPG Bulletin, Vol. 77, No.3, p. 467. Garb, FA 1988. "Assessing Risk in Estimating Hydrocarbon Reserves and in Evaluating Hydrocarbon- Producing Properties." JPT, Jun. 1988,pp.765-778. Haun, J.D. (ed.) 1975. Methods ofEstimating the Volume ofUndiscovered Oil and Gas Resources. American Association of Petroleum Geologists, Studies in Geology No. I, p. 206.
Hohn, M.E. 1988. Geostatistics and Petroleum Geology. MacMillan, New York, NY, p. 264. Houghton, J.C., Dolton, G.L., Mast, R.F., Masters, C.D., and Root. D.H. 1993. US Geological Survey Estimation Procedure for Accumulation Size Distributions by Play." AAPG Bulletin, Vol. 77, No.3, p. 454. Isaaks, E.H., and Srivastava, R.M. 1989. An Introduction to Applied Geostatistics. Oxford University Press, Oxford, UK, p. 561. Kadane, J.B. 1990. "Comment: Codifying Chance." In: Mosteller, F., and Youtz, C., 1990. Lee, PJ., and Wang, P.C.c. 1983. "Probabilistic Formulation of a Method for the Evaluation of Petroleum Resources." Jour. ofthe Int. Soc.for Math. Geol., Vol. 15, pp. 163-181.
Campbell, J.M. (ed.) 1970. Oil and Gas Property Evaluation and Reserve Estimates. SPE Reprint Series, No.3. - - - . 1986. "Nontechnical Distortions in the Analysis and Management of Petroleum Investments." JCPT, Dec. 1986.
---.1985. "Prediction of Oil or Gas Pool Sizes when Discovery Record is Available." Jour. of the Int. Soc.for Math. Geol., Vol. 17, pp. 95-113.
Capen, E.C. 1976. "The Difficulty of Assessing Uncertainty." SPE Journal, Aug. 1976, pp. 843850; also in Megill, 1985.
- - - . 1986. "Evaluations of Petroleum Resources from Pool Size Distributions." In: Rice, D. D., 1986.
Clark, I. 1979. Practical Geostatistics. Applied Science Publishers, London, UK, p. 129.
Masters, C.D. 1993. "US Geological Survey Petroleum Resource Assessment Procedures." AAPG Bulletin, Vol. 77, No.3, p. 452.
Crovelli, R.A., and Balay, R.H. 1991. "A Microcomputer Program for Energy Assessment and Aggregation Using the Triangular Probability Distribution." Computers & Geosciences, Vol. 17., No.2, pp. 197-225.
- - - . (ed.) 1984. Petroleum Resource Assessment. International Union of Geological Sciences, Publication No. 17.
279
DETERMINATION OF OIL AND GASRESERVES
McCray, A.W. 1975a. Petroleum Evaluations and Economic Decisions. Prentice-Hall, Inc., Englewood Cliffs, NJ, pp. 3-4. - - - . 1975b. Petroleum Evaluations and Economic Decisions. Prentice-Hall, Inc., Englewood Cliffs, NJ, p. 215. - - - , . 1975c. Petroleum Evaluations and Economic Decisions. Prentice-Hall, Inc., Englewood Cliffs, NJ,Ch.8. - - - . 1975d. Petroleum Evaluations and Economic Decisions. Prentice-Hall, Inc., Englewood Cliffs, NJ, Ch. 7. Megill, R.E. 1984. An Introduction to Risk Analysis (2nd ed.). PennWell Publishing Co., Tulsa, OK. - - - . 1985. Evaluating and Managing Risk: A Collection ofReadings. SciData Publishing, Tulsa, OK. Miller, B.M. "Resource Appraisal Methods: Choice and Outcome." In Rice, 1986. Mosteller, F., and Youtz, C. 1990. "Quantifying Probabilistic Expressions." Statistical Science, Vol. 5, No. I, pp. 1-34. Newendorp, P. 1975a. Decision Analysisfor Petroleum Exploration. Petroleum Publishing Company, Tulsa, OK, Ch. 6. - - - . 1975b. Decision Analysis for Petroleum Exploration. Petroleum Publishing Company, Tulsa, OK, Ch. 7 & 8.
Podruski, J.A., Barclay, J.E., Hamblin, A.P., Lee, PJ., Osadetz, K.G., Procter, R.M., and Taylor, G.C. 1987. Conventional Oil Resources of Western Canada. Part I: Resource Endowment. Geologi_ cal Survey of Canada Paper 87-26, Minister of Supply and Services Canada. Rice, D.O. (ed.). 1986. Oil and Gas Assessment: Methods and Applications. American Association of Petroleum Geologists, AAPG Studies in Geology #21. Robinson, J.G. 1990. "Determination of Reserves and Values and Application of Risk." JCPT, Nov. 1990 Supplement. Rock, N.M.S. 1988. Numerical Geology. SpringerVerlag, New York, NY. Root, D.H., and Attanasi, E.D. 1993. "Small Fields in the National Oil and Gas Assessment." AAPG Bulletin, Vol. 77, No.3, p. 485. Root, D.H., and Mast, R.F. 1993. "Future Growth of Known Oil and Gas Fields." AAPG Bulletin, Vol. 77, No.3, p. 479. Sluijk, D., and Parker, J.R. 1986. "Comparison of Predrilling Predictions with Postdrilling Outcomes, Using Shell's Prospect Appraisal System." In Rice, D.O. (ed.), 1986. Spetzeler, C., and Stael von Holstein, C. 1975. "Probability Encoding in Decision Analysis." Management Science, Vol. 22, No.3, Nov. 1975, pp.344-347. Tversky, A., and Kahneman, D. 1985. "The Framing of Decisions and the Psychology of Choice." In Megill, R.E., 1985.
280
Chapter 23
THE REGULATORY ENVIRONMENT
23.1
INTRODUCTION
This chapter describes the regulatory environment for the petroleum industry in Canada. The regulatory activities, functions and objectives of both the provincial and the federal levels of government are described, as well as the necessary legislationand organizationalstructures. The focus is on Alberta, which is Canada's largest producer of oil and gas. The regulatory environment in the other producing provinces would, in general, be similar. Governments are involved in a number of different functions that have a direct influence on the development of oil and gas reserves: • Resource inventories • Mineral ownership • Economic development policies • Conservation control • Development, operating, and environmental regulation • Domestic supply assurance • Fiscal policies • Business regulation • International policies The provincial governments have jurisdiction over all aspects of the petroleum industry within provincial borders, except for lands under federaljurisdiction, such as Indian reservations and national parks. The federal government has jurisdiction over all frontier lands, including the Yukon and Northwest Territories, Hudson's Bay, and most of Canada's offshore areas. The federal government signed accords with the governments of Newfoundland and Nova Scotia, giving these provinces joint control with the federal government over offshore petroleum. The federal government has jurisdiction over interprovincial and international trade and commerce, which are of major importance to the petroleum industry.
Basic policy direction in all of the functions is established by political elements of government. At the provincial level these include the Legislature, the Premier and Cabinet, and the Minister of Energy; at the federal level, the Parliament, the Prime Minister and Cabinet, and the Minister of Energy, Mines and Resources. The basic policies are embodied in the acts and regulations that are administered by specialized government agencies. In Alberta, the principal agencies are the Energy Resources Conservation Board (ERCB), the Alberta Department of Energy, and the Alberta Petroleum Marketing Commission. The principal federal agencies are the Department of Energy, Mines and Resources, the National Energy Board (NEB), and the Department of Indian and Northern Affairs. Under the accords with Newfoundland and Nova Scotia, management and regulation are carried out by the Canada-Newfoundland Offshore Petroleum Board and the Canada-Nova Scotia Offshore Petroleum Board, which have equal representation from the federal government and the particular province. Reserves estimates are an important factor in many regulatory functions and policies. The government sources of reserves estimates, how these are used, and their effect on reserves development are discussed in this chapter.
23.2
RESOURCE ASSESSMENTS
Assessments of petroleum resources and reserves are needed by governments in order to carry out their functions relating to exploration, development, and the use of these resources. Both the provincial and federal governments conduct their own assessments of ultimate potential, recoverable reserves, and supply (rate ofproduction). Alberta's assessment is done primarily by the ERCB, which carries out continuing detailed evaluation of reserves and periodic assessments of ultimate potential. The Alberta Geological Survey also does some assessments.
281
-..,
DETERMINATION OF OIL ANDGAS RESERVES
The Geological Survey of Canada evaluates ultimate potential for the federal government, and the NEB evaluates reserves and supply. These two agencies also work with the Energy Sector of the Department of Energy, Mines and Resources in the Petroleum Resources Appraisal Panel. Because reserves assessments are frequently quite subjective and interpretive, and evolve as a result of emerging technology and changing economic conditions, governments seek comparative estimates from external sources. These sources include reserves assessments supporting gas removal and export licence applications, assessments by other organizations, and voluntary submissions by oil and gas companies. Governments are also very interested in reserves estimates for other jurisdictions (i.e., other provinces and other countries) with whom they compete for investment capital and for markets.
23.3
MINERAL OWNERSHIP
The majority of mineral rights in Canada are owned by either the provincial or the federal Crown. The remainder are held privately by individuals or corporations whose ownership originated from land and mineral rights granted a century or more ago to certain parties, notably the Canadian Pacific Railway Company and the Hudson's Bay Company. The Province of Alberta owns about 80 per cent of the mineral rights within its borders. The federal government is responsible for mineral rights in the territories (Yukon and Northwest Territories) and offshore (arctic, east coast, west coast, Hudson's Bay, and St. Lawrence), as well as in federal lands and Indian Reserves within the provinces. Governments manage these mineral rights on behalf of the citizens. The mineral rights are leased to private operators for development. Leasing is done by a competitive bidding process involving an initial acquisition cost (bonus), plus annual rental fees and royalties (share) on production. Decisions regarding initial acquisition costs usually take into account estimates of reserves under the tracts involved. In Alberta, the Department of Energy is responsible for mineral rights disposition, rentals and royalties under the Mines and Minerals Act. The department relies on competitive bidding in the initial disposition of rights (Crown sales). Ifa well has not been drilled during the initial term ofthe lease, under certain circumstances the lease may be extended if the lands are considered to be capable of economic production.
282
Government's royalty interest can be taken in-kind (the actual oil or gas, rather than monetary proceeds of its sale), which results in direct involvement in marketing of the product. In Alberta, the marketing of Alberta's royalty oil is carried out by the Alberta Petroleum Marketing Commission (APMC), which has a mandate to ensure that the greatest possible benefits are secured from the sale of Alberta's oil and gas. Royalty on gas is not taken in-kind, but the APMC closely monitors all gas sales from Alberta under the Natural Gas Marketing Act. The APMC also represents Alberta at national and international regulatory proceedings. The Alberta government has an interest in acting to maintain or increase petroleum product prices because of the direct royalty income and also because of the overall economic development benefits. In the early 1970s,the Alberta government brought about an increase in the price of its gas, which at that time was significantly underpriced on a heating value basis relative to oil. In the subsequent energy crisis, Alberta sought to have the price of its oil and gas follow the rapidly escalating world prices. Leasing of federal oil and gas rights is done under terms of the Canada Petroleum Resources Act. Rights to explore are granted after competitive bidding based on the proposed exploration expenditure during the initial term of the licence. The Department of Indian and Northern Affairs has the responsibility of managing these rights in frontier lands north of the 60th parallel, and the Department of Energy, Mines and Resources has responsibility south of the 60th, except for the east-coast offshore accord areas, which are under the Canada-Newfoundland and Canada-Nova Scotia Offshore Petroleum Boards. The Department of Indian and Northern Affairs assists in the management of mineral rights in Indian Reserves. The levels of royalties can affect whether a particular reserve is developed and becomes proven, or remains in a less certain category of reserve. Royalty reduction or royalty holidays are commonly used to promote development of the petroleum industry or a particular sector of the industry. Similarly, the oil and gas price levels, which can be influenced to some extent by governments, significantly affect the rate of development. When prices are restrained, the rate slows; when prices escalate, the rate increases.
23.4
ECONOMIC DEVELOPMENT POLICIES
Most governments have an interest in overall economic development, and the governments usually concentrate
---I THE REGULATORY ENVIRONMENT
on the development of industries having a large potential. Decisions regarding the promotion of petroleum industry development are based on estimates of ultimate potential, including assessment of conditions necessary for these to be economic viable reserves. Petroleum development is very capital-intensive. The methods used by governments to attract capital include the following: • Minimizing of royalty and taxation • Subsidization • Loan guarantees • Equity participation by government • Funding of infrastructure construction Funding of research
fashion. In 1985 the two levels ofgovernment agreed to deregulate and allow market-responsive pricing.
23.5
CONSERVATION CONTROLS
23.5.1 Field Development and Production Conservation Oil and gas reserves can be lost both in the reservoir and on the surface as a result of wasteful production practices. The governments of the producing provinces have the following kinds of legislation to minimize wasteful practices: • Limits on excessive gas production from oil reservoirs • Requirements to implement enhanced recovery schemes
• Assistance with market development
• Production rate limits from wells or pools
• Maintenance of petroleum product prices
• Requirements to gather and conserve solution gas produced with oil
Maintenance ofpolitical, business, fiscal, and social stability • Provision of business and technical information • Publication of estimates of available resource potential Industrial diversification and decentralization are two other usual government objectives that are embodied in many policies that affect oil industry development. As an example of a government policy meant to stimulate the development ofthe Canadian oil industry, the National Oil Policy (196 I) reserved the Canadian market in Ontario and westward for Canadian oil at a time when cheap offshore supplies were available. An example of government financing support was the construction of the TransCanada gas pipeline (1957) to stimulate development of the then-fledgling gas industry and provide an alternative energy source to Ontario. Development of the Athabasca oil sands has been supported in a variety of ways including equity participation, loan guarantees, reduced royalties and taxation, and the direct funding ofresearch and testing ofvarious recovery methods. When world petroleum prices were rapidly escalating during the 1970s, Alberta and the producing provinces sought to have their petroleum prices follow world levels, but the federal government, reflecting the interests ofthe consuming provinces, wanted to restrain the rate of price increases to domestic consumers. Therefore, from 1975 to 1985, natural gas prices were controlled by agreements negotiated between the federal government and the governments of the producing provinces. Oil prices were controlled in a similar
In Alberta, conservation requirements are stipulated in the Oil and Gas Conservation Act and Regulations, which are administered by the ERCB. The application of specific measures to individual wells and pools frequently depends on the reserves estimates for those wells and pools. Conservation controls serve to increase proved reserves in developed pools, but can slow the development of other projects that are competing for limited available capital. Regulation is necessary because the economic rate of return of a conservation project is sometimes less than for the same project without controls.
23.5.2 Consumer Demand Conservation Particularly during the energy crisis of the 1970s, all levels of government were involved in programs to reduce the demand ofpetroleum products. Many ofthese were in the form of advertising campaigns and incentive programs to reduce waste and increase efficiency by the individual consumers. Development of alternative fuels was actively supported. In Alberta, the ERCB regulates manufacturing industries that use natural gas, and requires gas useage to be efficient.
23.6
DEVELOPMENT, OPERATING, AND ENVIRONMENTAL REGULATIONS
The provincial and federal governments impose a variety of detailed regulations relating to the construction and operation of oil and gas production facilities. These regulations are aimed at achieving safe, orderly, efficient, and equitable development and operation of
283
DETERMINATION OF OILAND GASRESERVES
facilities, and minimizing their social and environmental impacts. Some of these regulations are specific to the petroleum industry while others apply to all industries.
requiring NEB approval, the NEB coordinates the EARP reviews carried out by all the federal departments that are involved.
In Alberta, the regulations specific to the petroleum industry include the following acts and the regulations pertaining to them: the Energy Resources Conservation Act, the Oil and Gas Conservation Act, the Oil Sands Conservation Act, and the Pipelines Act. The regulations are administered by the ERCB and apply to all oil and gas wells, pipelines, and production and processing facilities. The regulations cover such aspects as demonstration of need, sites and routes, sizing and design, construction and operational practices, monitoring and reporting, and ultimate decommissioning offacilities. Applications for each individual facility and operation must show that all regulations and standards will be met. The applications are also subject to scrutiny by the public, including affected landowners and residents, special interest groups, and competing oil companies. Public hearings are held when issues dictate or when the issues cannot be resolved by private negotiation. When all requirements and concerns have been met, approvals, permits and licences are issued. Further specific approvals are required from other provincial government departments. Department of Environment authorizations are required for pollutant emissions and waste disposal, watercourse crossings, and land surface disturbance and reclamation. Where public lands are involved, land use authorizations must be obtained from the Department ofForestry, Lands and Wildlife. Development permits must be obtained from municipal authorities. Generic provincial regulations regarding worker and public safety, building and construction standards and codes, and industrial water usage apply to all industries including the oil industry.
Development regulations often result in increased development costs for reserves. In particular, pollution and environmental control requirements are becoming an increasingly significant factor in the cost of petroleum development and production.
The ERCB, the Alberta Department of Environment, and other provincial agencies do regular inspections and compliance monitoring during the operating lives ofoil and gas facilities and operations. In federal lands, similar functions are exercised under the Oil and Gas Production and Conservation Act by the NEB and the Offshore Petroleum Boards. The NEB is also responsible for regulation of pipelines that cross provincial and international borders. The Federal Environmental Assessment and Review Process (EARP) applies to projects that are on federal lands, receive federal funding, require approvals from federal departments, or are undertaken directly by a federal department. For hydrocarbon development projects
23.7
DOMESTIC SUPPLY ASSURANCE
The provincial and federal governments both review proposals for removal or export of certain energy products from their jurisdictions, to ensure that domestic long-term needs are provided for and that the export is in the best interests of their jurisdictions. In Alberta, the removal of natural gas is subject to the Gas Resources Preservation Act, which is administered by the ERCB, and is subject to subsequent further approval by the provincial government. Alberta's current removal criteria requires that a IS-year supply be reserved for the core market within the province (principally residential and commercial consumers) in the form ofestablished reserves ("proven" plus a portion of "probable") before removals are permitted. The entire amount ofgas approved for removal must be in the form of established reserves and under the contractual control ofthe permit holder at the time the permit is issued. This requirement causes reserves to be moved from the "possible" category to "proven" and "probable." In the absence of this criterion, the supply for the later stages of some long-term supply contracts would likely rely on "possible" reserves. Gas removal from Alberta is subject to further approval by the Minister of Energy or the Provincial Cabinet, both of which consider such matters as gas pricing and market practices, commitments, and destinations. At the federal level, the National Energy Board is responsible for deciding export applications. One of its current criteria is that sufficient established reserves must be under contract to cover the volume licensed for export. In addition, gas purchasers in Canada are given an opportunity to formally complain about any proposed export if they have been unable to obtain gas on terms and conditions similar to the proposed export. The ability to control cross-border movement of gas allows some capability to control prices. Both Alberta and Canada acted to ensure higher prices for exported oil and gas during the "energy crisis" ofthe 1970s when prices were escalating worldwide.
284
-------
iiI
THE REGULATORY ENVIRONMENT
23.8
FISCAL POLICIES
Governments can vary taxation levels, apply special taxes, and grant tax incentives to achieve the objectives of deficit or balanced budgets, and to stimulate economic growth (especially in localities, areas and regions of lagging economies) and industries of specific importance. A number ofdifferent taxes are levied on the producing sector of the petroleum industry by municipal, provincial and federal governments. Municipal governments levy property taxes on petroleum facilities and real estate. Both the provincial and federal government levy tax on all corporate income. In addition, the Alberta government collects a freehold mineral tax on all freehold mineral leases (mineral rights not held by government), based on their production. The federal and provincial governments both have the ability to apply special taxes to the petroleum industry. The Alberta government grants a special royalty tax credit of up to 2.5 million dollars annually, which is a significant benefit to small petroleum companies. During the period 1972 to 1984, the Alberta government provided exploratory seismic and drilling incentives through drilling credits and royalty holiday programs. These were designed to promote exploration for new reserves and to maintain industry activity during periods of economic downturn. During this same period, Alberta significantly increased its royalties on oil and gas. The federal government offered tax incentives for frontier exploration in the 1960s and 1970s. The National Energy Program brought in by the federal government in 1981 provided incentive payments for exploration and development expenditures (Petroleum Incentive Program) giving particular advantage to companies of predominantly Canadian ownership. The National Energy Program also significantly increased federal taxation of oil and gas production, particularly through the Petroleum and Gas Revenue Tax. Fiscal policies affect the timing of reserves development. Governments within Canada and throughout the world are in competition to attract and retain investment capital. A high economic rent may cause exploration capital to move to other provinces or other countries where the economic rent is lower. Petroleum industry capital is quite mobile. Even small companies tend to look on a worldwide basis, and they can invest away from their home operating area by taking a minority working interest in a project sponsored by a larger company.
23.9
BUSINESS REGULATIONS
Price- or fee-setting regulation is necessary where normal business competition is not present. This is the case with many pipelines. For field gathering systems and for gas processing plants in Alberta, a production owner may apply to the ERCB to have these facilities declared to be "common" and then to the Public Utilities Board to have useage fees set. Tolls and tariffs on interprovincial transmission lines are under the jurisdiction of the National Energy Board. The prices for natural gas distributed to end-users by local utility companies are subject to the approval by some type of public utilities board in each of the provinces. These boards protect the interests of the consumer and ensure that rates are justified. Right-of-entry and land compensation legislation has been established by the provinces so that mineral owners cannot be prevented by surface owners from recovering the minerals. In Alberta, this function is carried out by the Surface Rights Board. The provinces each have some type of securities and exchange commission that regulates corporate matters. Public corporations are required to publish annual financial statements listing, among other things, the assets of the corporation. Oil and gas reserves are the primary assets of most petroleum companies. Consistency in the method of estimating the volume and value of these reserves is important. The Federal Competition Act (FCA) protects and promotes competitive processes, and is administered by Consumer and Corporate Affairs Canada. The act has both criminal and non-criminal provisions. The latter are prosecuted by the Attorney General ofCanada. Criminal offences include conspiracy, bid-rigging, price discrimination, predatory pricing, price maintenance, misleading advertising, and deceptive marketing practices. Certain other activities and practices are subject to review, but are not criminal matters. For example, companies proposing to merge must notify Consumer and Corporate Affairs Canada if the companies exceed a certain size of assets or gross revenues. If a proposed merger is found to prevent or substantially lessen competition, the merger may be conditioned or prohibited.
23.10 INTERNATIONAL POLICIES Petroleum is usually considered to be a strategic commodity. Hence, countries try to protect themselves from any serious supply disruptions that may result from 285
~
DETERMINATION OF OIL AND GAS RESERVES ......:,.:,.. : i
business, political, or natural events. Canada has a large oil and gas potential and has followed the strategy of developing supply through the incentive ofallowing exports to the USA. Exploration and development of frontier areas could not proceed on the basis of the domestic Canadian market alone. Canada has exported oil and gas to the USA for decades, and the US views Canada as a secure supplier. Currently, approximately one-third ofCanada's oil production goes to the USA. Oil is being exported to the west and midwest regions of the USA at the same time offshore oil is imported into Quebec and the Maritimes, resulting in a near balance ofexports and imports. More than 40 percent of Canada's gas production is exported to various parts of the US including California, the midwest, and recently the northeast. The producing provinces are usually more anxious to increase exports than are the consuming provinces. The federal government has the challenge of balancing the interests of both groups. Canada's close political and economic ties to the US led to the Free Trade Agreement which came into effect in January 1989. For Canadian oil and gas producers, this agreement provides access to US markets free of export or import taxes or duties. Except in national defence emergencies, export restrictions may be applied only under limited circumstances and in a proportionate manner. Incentives for exploration and development are still allowed. Recently, Canada, the USA, and Mexico negotiated the North American Free Trade Agreement (NAFTA) which is scheduled to come into effect on January I, 1994.
This agreement will allow for relatively free movement of oil and gas between these three countries and will allow the Canadian and US companies to participate in supply and services ofthe Mexican petroleum industry but not ownership of Mexican oil and gas resources. ' The review of company take-overs and mergers by Consumer and Corporate Affairs Canada described in Section 23.9 applies even if foreign-owned companies are involved. Canada is signatory to the Agreement on an International Energy Program, along with the USA and a number of European countries. This group has a plan for distribution of available oil in a supply crisis. In issues of worldwide public concern, Canada seeks to do its part. Energy conservation was such an issue during the energy crisis of the '70s. Environmental concerns, including petroleum transport failures and atmospheric emissions, are issues at the current time. In a similar vein, Canada feels obligated to provide technical assistance to developing countries to develop their petroleum industries. Canada and the producing provinces provide expert advisers and trainers in response to specific requests and under on-going international aid programs. Government interfacing with more developed countries is often in the form oftechnical exchanges and discussions about business and regulatory systems. Some countries seek assistance from Canada specifically, not just because it has a high level ofexpertise and technological development, but because the country seeking aid wants to avoid a tie to the USA. Both the federal and provincial governments are active in trade development initiatives in other countries.
286
s
Chapter 24
CRUDE OIL MARKETS
24.1
INTRODUCTION
This chapter provides an overview of Canadian crude oil markets, with particular focus on Alberta and other westem Canadian provinces. More detailed information can be found in the references cited in this chapter. Despite the slump in activity since 1986, the Canadian oil industry is an important element of the domestic economy in terms of direct employment, total revenue and trade surplus. Crude oil satisfies nearly 40 percent of the total domestic energy demand. According to the Petroleum Communication Foundation (1992), the bulk of crude oil was used for various modes of transportation (65 percent), heating and electricity generation (25 percent), and manufacture of oil-based products such as asphalt, lubricants and various petrochemicals (10 percent). In the global context, Canada is a mediumsize oil producer, supplying less than 3 percent ofworld production. The history of the Canadian oil industry dates back to 1947, when the discovery ofthe Leduc field triggered a resurgence of exploration activity in Alberta. During the 1950s, the industry operated in a free market environment characterized by essentially no government regulation. The period of the laissez-faire approach towards the oil industry came to an end in 1961, when the National Oil Policy divided Canadian markets along the Ottawa Valley line. Markets to the east of the line were to use cheaper imported crudes, while markets to the west were to be supplied with domestic crudes. The oil crisis in 1973-74 precipitated further market regulation involving an oil price freeze, an export tax, reductions in exploration write-offs and depletion allowances, and the establishment of Petro-Canada as a national oil company. Market intervention reached its peak in 1980 when, on the heels ofthe second oil crisis, the federal government introduced the National Energy Program (NEP), which provided for oil price controls, several new federal taxes, Canadian ownership targets, and incentives for fuel switching. This first federal
attempt to tax resources under provincial jurisdiction was countered by very strong opposition from Alberta. Election of the conservative government in 1984 and the decline in oil prices paved the way for the Western Accord between the federal government and the energyproducing provinces. The signing ofthe accord on June I, 1985 marked the dawn of crude oil deregulation and the demise of both the NEP and 12 years of administered pricing. All NEP taxes were either phased out or eliminated, and oil price controls were lifted. Canadian producers were free to compete in the international market place and reap the rewards of unrestricted sales opportunities through the direct negotiation ofcontracts with refiners and marketers. In the process, Alberta modified its prorationing program, virtually returning control of production levels to producing companies. In Canada, the term "conventional" crude oil usually refers to light, medium and heavy crudes from the western Canadian sedimentary basin-the traditional source for most Canadian production. The distinction among these three classes ofcrudes is based mainly on gravity, with specific gravity cutoff rates differing regionally in the absence of one widely recognized international standard. For instance, the minimum gravity for light crude ranges from 28·API in Canada to 32·API in the US and 36· API overseas. Medium oil is defined by the World Energy Conference as having gravity between 22· and 31·API (Petroleum Communication Foundation, 1992). In Canada, generally no distinction is made between light and medium oil. Heavy oil is typically defined as crude with the API gravity between 12· and 28·API, although Alberta's Energy Resource Conservation Board (ERCB) uses a lower maximum of25·API. Since Canadian pipelines generally require oil to have a gravity of at least 21·API, some of the heaviest grades must be blended with condensate or natural gas liquids to be shipped by pipeline.
287
~
DETERMINATION OF OIL ANDGAS RESERVES
"Nonconventional crude" comprises synthetic and frontier oil. "Synthetic" oil is heavy oil and oil sands bitumen refined to make a product similar to high-quality light crude oil. "Frontier" oil includes resources off the eastern coast or north of the 60th parallel in the Arctic. Nonconventional crudes differ from conventional in that they are more difficult to recover and cannot be shipped to a refinery without processing or preparation.
Interprovinclal From Norman Wells
Rainbow
Geographically, oil production is heavily concentrated in the western Canadian sedimentary basin. In 1992, Alberta, which is by far the largest oil producing province, accounted for 8I percent ofthe 1.73 million barrels per day produced in Canada. Saskatchewan accounted for another 13 percent and the remainder came from British Columbia, Northwest Territories, Manitoba and Ontario and Nova Scotia. Conventional light oil dominates the Canadian crude slate, representing halfoftotal production. The other half comprises heavy oil (29 percent), synthetic oil (13 percent) and pentanes (8 percent). Historically, Canadian crude oil production has been relatively stable since 1980, hovering closely around the 1992 level. Underlying this almost flat overall performance were divergent trends for conventional and nonconventional crude oils, conventional declining at 5 percent per year (mainly light crude) and nonconventional increasing by more than half since 1988.
24.2
•
Calgary
lnterprovlncletto Eastern Canada andU.S.
Continental ToU.S. Rock MIn.States
Figure 24.2-1
Major Alberta Pipeline Systems
TRANSPORTATION NETWORK
Canadian crude reaches domestic and export markets through a vast network of pipelines. In Alberta, a capillary-like system of gathering lines throughout the province collects and transports field crude production to a smaller number of feeder pipelines. Trucking is a marginal mode of transportation and is used locally in areas where batteries are not connected to the gathering lines. Most ofthe feeder pipelines conjoin at Edmonton (Figure 24.2-1) to serve the Alberta refining market and further link to two major interprovincial pipeline systems: Interprovincial Pipe Line (IPL) and Trans Mountain Pipeline (TMPL). These and most other Canadian pipelines are "common carriers", that is, public utilities for hire obligated by law to provide equitable and nondiscriminatory pipeline access to all interested parties. IPL operates the largest and most complex crude oil pipeline system in North America, stretching over 3700 kilometres from Edmonton to Montreal (Figure 24.2-2). It transports up to 35 different types of liquid hydrocarbons to refineries in eastern Canada and the
Great Lakes region of the US Midwest. In addition, IPL is linked to the Wascana system, which provides access south to the Rocky Mountain markets of Wyoming and Colorado. During 1992, average throughput on the IPL system was 1.45 million barrels per day, with deliveries split almost evenly between domestic and export markets. ~
TMPL extends over 1300 kilometres from Edmonton to delivery locations in the Vancouver area. The system can transport up to 190 thousand barrels per day ofcrude oil, partially processed oil, and petroleum products from Alberta. TMPL also receives small volumes of crude from northern B.C. via the West Coast Pipe Line connection at Kamloops. TMPL's marine terminal at Westridge, B.C. is capable of loading barges serving the US West Coast and small tankers providing access to Pacific Rim markets. TMPL also operates a lateral link from Sumas, B.C. to Anacortes, Washington, where four refineries are located. In 1992, the system delivered 161 thousand barrels per day to domestic locations
288
s
CRUDE OILMARKETS
Beaufort Sea
.
Taylor " (2860)'"
\
,
,
,
Edmonton I • .: (56300) • Prince George':... r .....
~1500)/ \},~. __
'\,
t-: " ,',,' ~ ·· ... ,.,.-
"
'
Loydminster ; ,,(3700) : Regina (7200)
.. ,
'b ,-JI
<::::;:::::> Come-By-Chance (15100)
"
Calgary Vancouver- .. -(5180) . -Moose (22700)
•
Jaw
(2110) •SaintJohn (27000)
Legend
•
Refinery Locations (m3/d)
Rainbow Pipeline
- - - Interprovincial Pipeline Portland-Montreal Pipeline
Wood River
TransMountain Pipeline
Proposed Pipeline Loop/Capline/Chicap from Gulf Coast
•
I I I
Source: National Energy Board. 19918.
Figure 24.2-2 Maior Crude Oil Pipelines and Refining Areas
and another 41 thousand barrels per day to export destinations. A smaller pipeline, the Rangeland, transports oil south from Edmonton into the Montana market. The line has a capacity of approximately 90 thousand barrels per day and has recently operated at rates approaching that capacity. The Portland-Montreal pipeline is the main oil import line that brings offshore crudes to the Montreal market. In 1992, the line delivered 166 thousand barrels per day of oil, equivalent to 67 percent of its maximum
capacity and over 90 percent of Montreal's oil requirements. Marginal volumes of US and overseas grades are also imported into Ontario via IPL's Lakehead portion, which ties in at Chicago with two major US pipeline routes: the Capline/Chicap system from Louisiana Gulf Coast and the Arco system from Texas Gulf Coast via Cushing, Oklahoma. Since the beginning of deregulation in 1985, pipeline capacity constraints have had occasional impact on Canadian crude oil pricing and production. Prior to IPL's 1987 expansion, insufficient capacity of the system necessitated diversions of Canadian crude to lower 289
DETERMINATION OFOIL AND GAS RESERVES
valued markets and even to the shut-in of wellhead production. More recently, in March 1991, a leak on Lakehead's Line 9 restricted the throughput and caused persistent prorationing of the nominated volumes. Apportionment continued through 1991 and early 1992, as the capacity was restricted to 80 percent by an order from the US Department of Transport and, additionally, by line closures during hydrostatic tests. In the face of these unprecedented high levels of apportionment, producers and shippers attempted to protect their access to IPL capacity by over-estimating crude supply. That practice led to even higher levels of apportionment. which, in tum, resulted in inequitable pipeline and market access. By mid- 1991, the forecasting and nomination systems used to schedule feeder pipelines and IPL broke down. Subsequently, an industry working group was formed to address the problem. The group developed modified procedures designed to eliminate overnominations through stricter monitoring of battery production and penalties for inflated forecasts. These procedures came into effect in March of 1992, reducing IPL apportionment only temporarily. Among other developments, IPL's extension from Sarnia to Montreal was re-opened in July of 1992, after being mothballed for one year. The initiative came from a group of Alberta producers and marketers who decided to move 20 to 30 thousand barrels per day ofheavy crude to Montreal. Around the same time, the Bow River Pipeline completed construction of a 55-mile pipeline across the US border, to ease access from the southern Alberta fields to the Billings market. This pipeline expansion was triggered by the addition of a heavy crude coker at Conoco's Billings refinery. The current capacity is around 24 thousand barrels per day, but can be expanded to about 42 thousand barrels per day. With the exception ofa few privately owned feeder lines that are not common carriers, Canadian pipelines are regulated by a host of government agencies. Pipelines crossing the US or provincial boundaries, such as IPL and TMPL, come under the jurisdiction ofthe NEB. As an independent federal regulatory tribunal, the NEB is responsible for the issuance of export licences for oil, natural gas and electricity; the certification of interprovincial and international pipelines and power lines; and the setting of pipeline tolls and tariffs (National Energy Board, 199Ib). Pipelines functioning within provincial boundaries are generally under provincial jurisdiction. For example, the construction and operation of the province's feeder pipelines are regulated by Alberta's ERCB and Public Utilities Board.
Setting tolls and tariffs is a key component of regulation and is intended to protect public interest against monopolistic or discriminatory practices ofpipeline companies. That protection is aimed at establishing "just" and "reasonable" tolls, which "under substantially similar circumstances are charged equally to all persons" (National Energy Board Act, 1985). The main standard of reasonable tolls is the cost of service, meaning necessary cost, reasonably or prudently incurred, inclUding the cost of capital. This involves consideration of the capital structure of a pipeline company and its operating costs and of the necessity to attract capital through a fair rate of return. The ancillary rules of cost-causality and user-pay imply that costs should be assigned directly to specific classes of service or customers or geographic areas, and that the users bear financial responsibility for the costs caused by the delivery of their particular commodity.
24.3
MAJOR MARKETS
Crude oil must be refined to the various forms of petroleum products before it can be utilized by the endusers. Thus, the refineries are essentially the only direct recipients of crude oil and, as such, determine the market for it. The refinery requirements are in tum driven by the level of inventories and sales of petroleum products to the consumers. The main product categories include motor gasoline, middle distillates, heavy fuel oil, and petrochemical feedstock. Seasonal nature of demand for these products dictates seasonal variations in refinery modes of operation and the optimal composition offeedstock crudes. During the 1980s, Canadian refiners faced volatile feedstock costs, reduced oil demand, changing product specifications and demand slate. The industry responded through rationalization, which included plant closures and refinery upgradings. As a result, eleven refineries were closed and two reduced in size. Over the past few years, Canada's refining capacity has stabilized at around 1.9 million barrels per day, down from 2.3 million barrels per day in 1980. In the process, the Canadian refining industry has become highly competitive and capitalintensive, resulting in the gradual erosion of the profit margins. In 1992, the refinery utilization rate dropped to 80 percent, down from 85 percent in 1990, as sluggish demand forced refiners to trim their crude runs (Energy, Mines and Resources Canada, 1993). Traditionally, Canadian refining centres west of Montreal have been supplied exclusively with western Canadian crude oil, while those east of Montreal have
290
--------------------_.",.,
CRUDE OIL MARKETS
relied heavily on water-borne imports of mostly light crude from offshore sources. The reliance of Atlantic refineries on imports has increased steadily, reaching almost 100 percent in the past few years. Montreal refiners have obtained their feedstock crude from both domestic and overseas sourcesbut, most recently,have increasingly favoured cheaper overseas crudes from the North Sea,West Africa and Latin America. Overall, Canada has been a net oil exporter, with the surplus of mainly heavy crudes declining gradually through the late 1980s, before increasing to 190 thousand barrels per day in 1991 and 289 thousand barrels per day in 1992 on the heels of sluggish domestic demand. The bulk of Canadian refining capacity is located in Ontario and the prairie provinces (Figure 24.2-2). Consequently, these two regions are the largest domestic markets for Canadianoil, with the receiptsreaching 26 and 21 percent of total 1992production respectively. British Columbia consumed another 9 percent of Canadian crude in 1992 (mostly from Alberta), while deliveriesto Quebec and the Atlantic provincesconstituted a mere 0.3 percent (Energy,Mines and Resources Canada, 1993). The remaining 44 percent of Canadian crude was destined for exports. The US Midwest was the primaryexportmarket for Canadiancrude,accounting for approximately three-quartersof all exports. The US Rocky Mountain and the US East Coast accepted the bulk of the remaining export barrels. In 1992, Canadian crude oil exports were split almost evenlybetweenlight andheavy crudes.However, heavy oilproducerswere substantially more dependent on foreign marketsthan light oil producers. Heavyoil exports amounted to three-quarters of total supply, while the comparable figure for lightoil was only 38percent. This strong dependence on export markets for heavy oil is caused by limited demand from Canadian refineries, which are designed to run predominantly light crudes. Although a large number of northern tier American refiners use Canadian heavycrudes,overhalfof the total exports is purchased by three large refiners: Koch at Minneapolis, and two Chicago refineries owned by Mobil andAmoco (Table24.3-1). Sincelate 1980s, these and other refineries (including Newgrade at Regina) have gone through debottlenecking, which has resulted in increased demand for Canadian heavy crudes. This growthhas been partially offset by the shutdown of the Sarnia-Montreal line, and a switch by the Uno-Ven (Union) refinery in Chicago to Venezuelan feedstock, following a 50 percent acquisition of the refinery by Petroleos de Venezuela SA (PDVSA).
Canadian crude is sold to the end-users directly by the producers, or throughthe Alberta Petroleum Marketing Commission (APMC) and several commercial marketing entities. The APMC is a provincial crown corporationand the largest marketer of Canadian crude oil, supplying it to a wide base of refiners throughout Canada and the northern tier of the United States. As agent for the Alberta Crown, the APMC is responsible for gatheringand marketingcrude oil royalty taken inkind from provincial Crown leases. It also markets Alberta's 16.74 percent equity share in Syncrude, and offers contractmarketingservices to Alberta producers (Alberta Petroleum Marketing Commission, 1992). The USRockyMountain regionis one of fivegeographical districts, delineated in 1950 by the Petroleum Administration for Defense (PAD) for the purpose of administration, and is often referred to as PADD IV. In recentyears, PADDIV has offeredthe highest netbacks forAlbertacrude, butrelatively limiteddemand. Bycontrast, the US Midwest (PADD II) and Ontario markets have been the mainrecipientsof Alberta crude, together accounting for over half of Albertaproduction. The US EastCoast(PADD I) andparticularly the US West Coast (PADD V) have been the marginal markets, both in terms of relative volumes and netbacks. Closure of the Sarnia-Montreal extension in mid-1991 and IPL's persistentcapacityconstraints sincethenhave led to the development of oil surplus in the traditional markets for Canadiancrudes. This encouraged Alberta producers to pursue opportunities in nontraditional markets in order to stabilize prices and avoid shut-in. Consequently, increased volumes of light crude were moved to Wyoming, the US West Coast and the PacificRim countries. For example, in its annual report the Alberta Petroleum Marketing Commission (1992) reported selling 900 thousand barrels of light royalty crude to the Chinese Petroleum Corporation in Taiwan during the fourth quarter of 1991.
24.4
NORTH AMERICAN PRICING
Deregulation coincided with an eraof substantially lower world oil prices in the aftermathofthe 1986price crash. AverageOECD I import prices have fallen from an average of over US $26 per barrel in 1985 to US $14 per barrel in 1986, and fluctuated in a US $15-20 range in recentyears. TheOPEC' "basket"price-another global indicator representing the average price for a basket of 'Organization forEconomic Co-operation and
Development. 'Organization of Petroleum Exporting Countries. 291
DETERMINATION OFOIL ANDGAS RESERVES
seven OPEC crudes and one Mexican crude-moved in tandem with OECD import prices. In North America, prices of West Texas Intermediate (WTI) crude have followed the world trend, falling from over US $30 in late 1985 to below US $12 in the summer of 1986, and fluctuating between US $15 and 23 through the summer of 1990 (Figure 24.4-1). In late 1990, WTI prices briefly soared over US $40 (on a daily basis), the highest level since 1982, on the heels of the Middle East tensions. Since March of 1991, WTI prices have exhibited remarkable stability, hovering in a narrow range around US $20. WTI enjoyed price premiums against major international crudes, particularly against heavier and more sour grades. These premiums have more than doubled since 1987, reflecting falling WTI output, US pipeline and refining bottlenecks, wider sweet/sour differentials and higher tanker rates.
Table 24.3-1
With Chicago constituting the key export market for Canadian crude, WTI is also the benchmark for Canadian light sweet oil. Since 1986, Canadian refiner postings have tracked WTI spot prices very closely. The FOB parity value for Alberta crude of equivalent quality has been based on WTI price at Chicago, netted back to Edmonton. In 1992, this value fluctuated around the transportation cost differential of US $0.85 per barrel below WTI (Figure 24.4-2), modified occasionally by "market discounts" reflecting local pipeline and
Importers of Canadian Heavy Crude
Company
Location Crude
Indiana
Since the establishment of the New York Mercantile Exchange (NYMEX), WTI has been the benchmark for North American and overseas light crudes. WTI is the deliverable grade of crude oil specified in the NYMEX futures contra~t. WTI prices are quoted at Cushing, Oklahoma, which IS the mam gathenng terminus for pipelines shipping US domestic crudes north to Chicago and other Midwest refining centers.
Rated Capacity (bed x 103) Asphalt Coking Cracking
Whiting Laketon
350.0 8.3
28.0
Clark Mobil Uno-Yen Marathan
Wood River Joliet Lemont Detroit
57.0 180.0 177.0 70
14.5 38.0 27.9
26.0 98.0 58.0 27.0
3.6 18.0
0.0 84.8 8.1 4.0
Minnesota
Koch Ashland
Rosemount St. Paul
218.5 67.1
58.0
55.0 23.0
35.0 14.0
157.6 3.4
Montana
Cenex Conoco Exxon Montana Refining
Laure Billings Billings
14.0 7.7
12.0 19.0 25.9
10.0 6.5 11.0
14.8 0.0· 10.0
-
2.4
1.2
2.3
BP Ashland
Toledo Canton
120.7 66.0
15.0
-
90.0 25.0
7.0 12.0
1.0 2.3
Washington US Oil
Tacoma
32.8
-
-
8.0
2.7
Wisconsin
Supedor
32.0
-
11.0
13.5
5.6
Illinois
Michigan
Ohio
Murphy
Great Falls
40.4 49.5 42.0 7.0
-
-
-
45.0 3.5
62.1 5.4
Amoco Laketon
-
145.0
Canadian Heavy Crude Usage in 1991 (bed x 103)
-. -
Source: Scott, 1992. • Coker startup in 1992.
292
s
CRUDE OIL MARKETS
refining constraints. That price relationship is expected to hold in the near future, as the positive impact of the phase-out of US import duties will be offset by rising IPL pipeline tariffs.
US$Jbbl 40.00
._----
30.00
---------------. ---- .-------------------.-- - --_ ••• -------- •• __•••••• --
10.00 -------------------------. --_ •. ------------ ••. -------- •• --_ ••••
0,00+---+---<----+---+-->---+--_JaB
J87
Source:Reuters.
Figure 24.4-1
JaB
JaB
a so
J91
J92
J93
Trading Month Averages
NYMEX WTI Prices at Cushing
Unlike sweet crude prices, the Canadian postings for sour and heavy grades have not consistently tracked their US and international benchmarks. Canadian prices for heavier grades have typically been lower and more volatile due to the limited Canadian market, seasonality of demand, aggressive competition from Mexican and Venezuelan heavy grades, and lower desirability of heavy feedstock. In 1991, the situation was further exacerbated by increased Canadian production of heavy crudes. Lower prices for heavy crudes reflect their inferior physical characteristics as compared to light crudes.
Tariff & loss Carrying cost Total
$US 1.22 $US 0.06 $US 1.28 Tariff & loss Carrying cost Total
U.S. importfee Custom user fee U.S. import fees
$US 1.26 $US 0.07 $US 1.33
$US 0.02 $US 0.02 $US 0.04 CHICAGO WTI + $US 0.52
Price used in pipeline loss and carrying cost calculations $US 21.74/bbl Exchange rate Canadian interestrate U.S. interestrate (Prime rate =1%)
Tariff & loss Carrying cost Total
$US 0.48 $US 0.04 $US 0.52
1.1917
5.94% 5.00%
Source: Alberta Petroleum Marketing Commission, 1992.
Figure 24.4-2 Alberta Crude Oil Pricing, Chicago Market (July 1992)
293
.
'-'
DETERMINATION OF OIL AND GASRESERVES
Heavy crudes yield significantly higher volumes of heavy components at the standard distillation cuts. Heavy components can be used either to produce lowpriced heavy oil products (i.e., residual fuel oil or asphalt) or run through sophisticated catalytic. or thermal conversion units to obtain lighter products (i.e., gasoline or jet fuel). Since lighter products fetch higher prices, but are also more expensive to produce from heavy crudes, the refiners' choice ofthe feedstock crudes is determined by relative product values and operating costs. The refinery coking differentials between light and heavy crudes, which represent the difference in "gross product worth" net of operating costs, constitute a floor for light/heavy crude price differentials. When the existing heavy oil conversion capacity is fully utilized and the demand for heavy products is limited, processing of incremental heavy oil requires installation of new conversion units. The refiners will consider such an investment ifthey can expect to recover associated costs through low prices of heavy oil feedstock. This threshold differential is more difficult to pinpoint as it is contingent on the type of conversion unit and location. The National Energy Board (I991c) estimates that the additional cost of upgrading ranges from US $6 per barrel for existing US refineries to US $10 per barrel for a stand-alone upgrader in Alberta. In summary, the price differentials for light and heavy crudes are driven primarily by the supply of heavy crudes, the demand for heavy products, and the economics of converting these crudes to lighter products. The demand for heavy products is related to economic activity, weather patterns, environmental regulations, competition from natural gas, and technological progress. These differentials are also affected by changes in world crude slates, available conversion capacity (planned vs. required), and transportation logistics. The refinery posted prices are typically set for heavy oil blends rather than for pure heavy crudes, which often require diluent to be shipped through the pipelines. Differentials against the reference light crude for Canadian heavy oil blends such as Bow River or Lloydminster have typically been around US $5-6 per barrel. These differentials widened to US $9 per barrel in early 1991, due to such factors as increased supply of domestic heavy crudes, the closure of the Sarnia-Montreal pipeline, warm winter weather, and natural gas substitution. By the summer of 1992, the differentials narrowed close to the typical levels, as a result of incremental demand by the newly constructed
"".".. c>
Lloydminster upgrader and the new coker at Conoco's Billings refinery.
24.5
PRICE RISK MANAGEMENT
With the deregulation and commoditization ofcrude oil in Canada, the producers have been exposed to international price volatility, thus creating the need to minimize price risks through the use of various instru_ ments that spread the risk over a large number ofmarket participants. The reduction ofthe price risk over a specific period of time is commonly referred to as hedging. The main hedging instruments include energy futures, options, and swaps.
24.5.1 Futures Futures are developed from forward contracts, that is, individually negotiated contracts for the future delivery of commodities. The uniqueness of each forward contract, limited transferability, and the lack ofa third-party guarantee of performance led to the creation of futures contracts. A "futures" contract is an agreement to buy or sell a standard quantity and quality ofa specific commodity at a fixed price, time and place, under the rules of a recognized exchange, guaranteed by a third party known as a clearing house (Arshi, 1992). The energy futures were established on the NYMEX in 1978. These were followed by the gasoline futures in 1981, crude oil futures in 1983 and propane futures in 1987. At present, NYMEX futures are the main short-term risk management vehicle for oil prices, with more than 30 million contracts traded annually. Each NYMEX crude oil futures contract represents an obligation to deliver one thousand barrels ofWTI crude at Cushing on a specified future date. A company wishing to protect its cash flow can use the futures to pre-sell a specified portion of its annual production by taking a short position on NYMEX. The company can sell futures contracts ifprices rise above target levels or when prices are expected to decline. If, subsequent to sellin.g these contracts, the prices fall, lower prices from physical sales are offset by profits from the futures "paper" trade. Alternatively, if prices rise, futures losses are offset by higher prices from physical sales. The NYMEX contract owes its popularity to relatively high liquidity (the highest number of contracts traded) and transparency (prices are broadly disseminated to the industry and quoted in the press). For these reasons, the NYMEX contract is the most accepted by the buyers and sellers as the benchmark for North American and international oil transactions. In particular, close correlation ofCanadian oil prices with NYMEX futures prices
294
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _A
CRUDE OIL MARKETS
makes the futures contract an excel1enthedging vehicle for Canadian producers. WTI futures contract prices are available on a real time basis up to 36 months out and provide participants with the opportunity to lock in oil prices any time during the trading hours.
24.5.2 Options Futures protect against unfavourable price movements at the expense oflost opportunity to benefit from favourable price movements. Crude oil options expand the range of hedging strategies by offering limited quantifiable risk and the potential to gain from favourable price movements as wel1. The first NYMEX option contract was launched in November 1986 for crude oil futures, fol1owed in June 1987 by an option on heating oil futures. These powerful financial instruments complement the energy futures and greatly enhance liquidity and trading opportunities at futures and options markets. A "put option" buyer pays a premium for the right to sel1 at a specific price for a specific period oftime. Therefore, a put option strategy can provide a guaranteed price floor. A "call option" gives the holder the right to purchase a futures contract at a specified price during the life of the option which, in effect, provides a guaranteed price ceiling. The simplest hedging strategy is to purchase put options to be exercised if WTI futures prices fal1 below a predetermined strike price. This strategy establishes a guaranteed floor price without sacrificing potential upward price gains. A more sophisticated strategy cal1ed a "fence option" involves establishing both a floor and a ceiling price.
24.5.3 Swaps "Swaps" are over-the-counter financial transactions that allow producers and consumers to transfer price risk to a financial intermediary. Liquidity problems associated with the purchase or sale offutures contracts in the more distant months have led to the development ofthese socal1edoil price swaps. The intermediary can either hold the unbalanced risk portion, match the position to an opposite counterpart, or use futures and options to balance the risk. The provider of swaps offers customtailored price insurance in a variety of crudes and products, with guaranteed maximum or minimum prices, according to the need, and protection from other market uncertainties. Oil swaps took off in 1988 when oil prices were falling and oil producers wanted a guaranteed revenue. The swaps market is now wel1-developed with brokerage companies and banks providing forward pricing in
crude oil for terms of up to ten years. Although most transactions are limited to less than two years. The volume of swap market is difficult to estimate due to close competition and the secrecy of swap business.
24.6
OUTLOOK AND CHALLENGES
Canada has a large resource base and enhanced access to the world's largest energy-consuming market, but the industry is facing chal1enges from deteriorating geology and rising production costs. There are strong indications that conventional oil production from the mature western Canadian basin may be in an irreversible decline. Hence, future supplies will have to come increasingly from nonconventional sources or imports. As a result, it is expected that the quality of crude and the regional distribution of supply will shift dramatical1y. Heavy oil will increase its share at the expense of light oil, while conventional supplies from western Canada will be increasingly replaced by nonconventional supplies of synthetic and frontier oil. Based on the latest projections by the National Energy Board (1991c), synthetic oil is expected to account for 18 percent of Canadian production by year 2000, while crude oil from the East Coast offshore (mainly Hibernia) is expected to provide about 12 percent. Consequently, the share of western Canadian conventional light oil in total supply may fal1 from over three-quarters in 1991 to only one-third by year 2010. Growing supplies of heavy oil will require equivalent growth in easily accessible markets, posing serious chal1enges for Alberta producers in the areas of transportation (reduced pipeline space, increased demand for diluent), marketing and refining (limited upgrading facilities and environmental concerns). There is some opportunity to increase sales to the established major export markets. New capacity could also be added in these markets, but this would require substantial up-front investment. Optimization of current available capacity, fol1owed by conversion of current light refining capacity to handle heavy crudes, seems to offer the lower cost solution to expected heavy oil refining bottlenecks. Some additional "grass roots" upgraders may also be required early in the next decade. The outlook for world oil prices is rather bleak. There is growing consensus that future world capacity additions will be more than sufficient to satisfy the world's sluggish demand for oil, which will be increasingly constrained by environmental regulations. As a consequence, real oil prices are general1y expected to show little, if any, growth over the next decade. Low oil prices
295
-
DETERMINATION OFOIL AND GASRESERVES
will have a detrimental impact on the size of established reserves, as well as on future production and industry cash flows. To survive in the continued soft price environment, Canadian producers will be hard-pressed to minimize production costs by employing leading edge technology and focusing on narrower areas of expertise. As well, vast amounts of capital will be required both upstream and downstream to meet increasingly stringent environmental standards. Marketing efforts will likely continue to be frustrated by growing refining and pipeline constraints, while governments will be faced with shrinking attainable economic rents and growing pressure to soften their crude oil royalty terms.
References Alberta Petroleum Marketing Commission. 1992. Annual Report (1991). Calgary, AB, p. 8. Arshi, A.A. 1992. "Energy Swaps as Profit Motive Instruments in Oil Markets". OPEC Review, Summer 1992, pp. 201-212. Energy, Mines and Resources Canada. 1993. The Canadian Oil Market. Vol. IX, No.1. Minister of Supply and Services Canada (1993), ISSN 0829-3732, p. II.
296
National Energy Board. 1991a. Canadian Energy. Supply and Demand 1990-2010. Minister of Supply and Services Canada, Calgary, AB, June 1991, Cat. No. NE 23-15/199IE, ISBN 0-66218956-6, p. 232. - - - . 1991b. Annual Report. Minister of Supply and Services Canada (1992), Cat. No. NE 11991E, ISBN 0-662-19372-5. - - - . 1991c. Canadian Energy. Supply and Demand 1990-2010. Minister of Supply and Services Canada (1991), Cat. No. NE 23-15/1991E, ISBN 0-662-18956-6, p. 228. National Energy Board Act. R.S. 1985. c.N-?, Cat. No. YX76-N7/1992, ISBN 0-662-58945-9, Part IV, par. 62, p. 37. Petroleum Communication Foundation. 1992. "Crude Oil." The Backgrounder Series, Calgary, AB. Scott, G.R. 1992. "Canadian Heavy Oil and Bitumen-Some New and Old Ideas." Paper presented at AOSTRA-Heavy Oil Assoc. conference, Calgary, AB, Jun. 1992.
Chapter 25
NATURAL GAS MARKETS
25.1
INTRODUCTION
Since the deregulation of gas markets in 1986, a very rapid evolution has occurred in the marketing, transportation, and government regulation ofnatural gas. The environment of gas markets has developed from that of a virtual monopoly held by a very small number of aggregators into a large spectrum of sales opportunities for producers falling into three general categories: direct purchases, aggregator purchases, and hedging opportunities such as storage purchases and participation in futures markets. Similarly the role of reserves estimates has changed considerably. No longer do the reserves provide the sole underpinning ofa contractual arrangement between the buyer and the seller; reserves are now frequently only one of the supply characteristics providing the buyer with assurance that his requirements will be met. Production forecasting is the synthesis of all of the factors and variables that drive the producer activities of exploring, developing, and selling natural gas. The market factors that affect production forecasting will be defined and discussed in this chapter by reviewing the Canadian and US market environment during the following periods: (I) the pre-deregulation era, before November, 1986 in Canada and before the mid-1980s in the United States, (2) the current era, and (3) the (expected) future. Demand forces exerted by various types of markets and buyers will be described followed by a discussion of production forecasting.
25.2 THE MARKET ENVIRONMENT 25.2.1 Review of Pre-Deregulation Era Market demand forces had a very strong, but somewhat indirect, role during the years prior to deregulation. Gas was purchased from producers by a small number of aggregators in the United States and Canada, and these aggregators were usually affiliates of pipeline companies or utilities. The aggregators pooled the producers' volumes and then resold them to utilities and local
distribution companies (LDCs), who in turn supplied the core and noncore markets. "Core markets" are defined for purposes of this discussion as "the group of consumers who have no ability to use alternative energy sources, primarily use gas for space heating, and have a high security requirement." In addition, the aggregators performed all of the intermediate steps between producer purchase and burner tip including transporting the gas from gas plant to end user and obtaining governmental regulatory approvals (such as Alberta Energy Resources Conservation Board (ERCB) removal permits, National Energy Board (NEB) export licences, and US import authorizations). Most importantly, the aggregator provided the contractual link that joined producer and end user. This contractual link coupled with the regulatory permits provided security of supply for the end user and for the exporting and importing geographical regions. The producer was given the assurance that the take levels and prices under his purchase contract would be maintained, often through some sort of "take-or-pay" or "take-or-release" mechanism. The end user was provided with assurances as to security of supply through the reserves pool that the aggregator had under contract. The exporting provincial and federal governments were provided with assurances ofsupply security through the surplus test mechanisms. This arrangement worked well enough considering the overall philosophy of the end users and regulators at the time who regarded the gas reserves of western Canada and the United States as a finite resource, limited entirely by the technology then currently available. The issue of supply security was paramount and was met entirely by the known reserves inventory. Accordingly, producers' contracts with aggregators were usually long term and had a daily contract quantity based on a 15 to 20 year reserve life. The end user did not deal directly with the producer, but the aggregator's purchase and sales contracts and regulatory mechanisms acted as a buffer that filtered market signals. This regulatory
297
~
DETERMINATION OF OIL AND GASRESERVES
mechanism has now been viewed as partially responsible for creating the large productive capacity surpluses during the last two decades, otherwise known as the infamous "gas bubble." This era came to an end with the following changes to regulatory approvals: Canada I.
NEB amendments to export licences in 1984, which allowed US purchasers to negotiate prices subject to the Toronto City Gate Price as a floor
2. The federal - provincial agreement on "Natural Gas Markets and Prices" ofMarch, 1985, which dropped price tests in favour of a price monitoring mechanism United States
1. The enactment of the Natural Gas Policy Act (NGPA) of 1978, which commenced a phased decontrol of wellhead gas prices 2. Various amendments by the US Federal Energy Regulatory Commission (FERC) to the Natural Gas Act (NGA) in the form ofFERC orders. The intention of these orders was as follows: • To ensure that producing states would regulate the physical production of gas and control intrastate marketing matters • To protect overall public interest where gas production transportation sales to end users involves two or more states • To establish a framework for contract demand conversions by merchant pipelines and for the creation ofgas inventory charges that would ultimately allow customers to purchase gas in a reliable and competitive fashion from as many suppliers as they wished The more important FERC orders that furthered the process were as follows: • May, 1984. FERC Order 380, which outlawed the collection ofminimum commodity bills, leaving LDCs free to reduce their minimum purchase obligations from the merchant interstate pipelines. • October, 1985. FERC Order 436, which introduced a voluntary open access program allowing LDCs to convert their service from sales to transportation, but ignored any resulting take-or-pay implications.
298
• June, 1986. FERC Order 451, which adminis_ tratively commenced deregulation ofinterstate gas as defined in the NGPA. • December, 1986. FERC Opinion 256, which addressed the problem of different pipeline rate calculations utilized by US and Canadian pipelines. Most Canadian pipelines use the "full fixed variable" rate design whereby all fixed costs namely, operating, maintenance, depreciation' debt costs, income taxes, and return on equity are included in the demand charge, and variable costs are included in the commodity charge. Most US pipelines utilize the "modified fixed variable" rate . design whereby income taxes and the retum on equity are included in the commodity charge rather than in the demand charge. This FERC opinion attempted to solve this problem by disallowing the pass-through of Canadian pipeline charges except for the ''prebuild'' portions of the Alaska Natural Gas Transmission System (i.e., the Foothills Pipeline System). • August, 1987. FERC Order 500, which addressed the take-or-pay implications in pipeline company-producer contracts that were unresolved in FERC Order 436. Open access pipelines were ordered to offer volumetric takeor-pay credits to shippers on their pipelines. The entire gas producing and transmission industry had to absorb the large take-or-pay liabilities which had been incurred from 1987 onwards as market forces forced the process ofreforming gas purchase contracts. • 1991. Proposed Mega-NOPR (Notice of Proposed Rulemaking), which forces mandatory unbundling on all pipeline companies, thus requiring them to offer transportation, storage, and balancing services on an individual basis to shippers. • 1992. FERC Order 636A, which implements the unbundling process by specifying the steps pipelines can go through to offer transportation, storage or merchant services to customers as well as offering rights on upstream pipelines.
25.2.2 Review of Current Era Deregulation in Canada commenced officially on November 1, 1986, and pent-up market forces were unleashed that caused structural changes to supply contracts, markets, transportation, and government regulatory requirements. The most obvious change
i'i'.".... ','
NATURALGAS MARKETS
occurred with prices: export prices from Canada were no longer set by the federal govermnent, but rather reflected competitive market forces. The customers and producers were now to be closely linked, without the artificial buffers that blurred market forces. Deregulation has progressed at different rates in the areas of marketing, supply, and transportation and in a different fashion in the US as compared to Canada. United States Deregulation
Deregulation in the US is occurring at a somewhat uneven pace. The end users themselves have adopted deregulation fairly quickly; for several years various end-users have purchased volumes directly from producers. However, transportation deregulation has been much slower than in Canada, and a number of US pipelines still retain their merchant function while a number of others are not yet open-access carriers. This appears largely due to the industry responses to FERC Orders 380 and 436, which gave LDCs and shippers options as to purchases and transportation. Another significant factor is, of course, that Canada has only one interprovincial carrier, TransCanada Pipelines, whereas a large number of pipelines exist in the US. One of the more significant events taking place in the US during 1991 was the California Public Utilities Commission's (Cpuq move to dismantle the Pacific Gas and Electric (PG&E) monopoly in northern California. This is an example of an individual state commission overturning freely negotiated contracts between two parties: PG&E with its affiliated purchasing arm Alberta and Southern, on the one hand, and the Alberta producer group on the other. It appears that this action is being taken by the CPUC in order to expedite the transition of Pacific Gas Transmission (PGT) to a competitive open access carrier. The same process is occurring with other US pipelines; however, the process was much more disruptive in the case of PGT due to the unique northern California gas system, where a large dedicated Alberta supply is connected to a single pipeline owned by an end user with a complete sales and distribution monopoly. Historically the US domestic supply-demand balance has reflected more of a market approach than in Canada where a mandatory surplus test created an artificial supply-demand ratio. The current US reserves-toproduction ratio, RIP, is approximately 8 to 10; the Canadian equivalent RIP is approximately 15 to 20. Approximately 40 to 60 percent of US sales volumes and approximately 20 to 40 percent of Canadian sales
volumesare purchased through short-term contracts (less than one year). Thus a significant number of sales arrangements are not based on traditional reserve-based contracts; instead, term, interruptible or deliverability types of contracts may be used. This reserves-toproduction philosophy, coupled with the clear reliance by end-userson short-term "spot market" price contracts, demonstrates the deregulated nature of the end-user segment of the industry. However, a significant number of merchant pipelines are not yet deregulated. The pace of this deregulation will be a function of the implementation speed of the Mega-NOPR and the progress of individual states-the most visible being California-towards full deregulation. This will not be a straightforward process; literally thousands of LDCmerchant contracts will be ultimately replaced with the following: • Direct LDC-producer sales contracts • LDC or producer transportation service contracts • LDC or balancing service contracts • Combinations of all of these Canadian Deregulation
The market environment in Canada is significantly different than in the US. Intraprovincial transportation is virtually fully open access from a contractual sense. The same is true of interprovincial pipelines; however, there is less physical access due to the fact that only one pipeline services markets east ofthe Alberta border, and a small number of pipeline systems carry volumes into the US market areas. Domestic markets, with the exception of some restrictions on core markets, are fully open to all producers, but there is limited access due to the magnitude ofthe long-term purchase contracts currently in place between the LDCs and their suppliers. A further restriction on open access is political: Alberta has not yet developed a policy of allowing producers to directly sell to eastern Canadian core markets on terms of less than 10 years. The intention is to maintain security of supply and prevent further erosion ofthe historic core sales arrangements by only allowing long-term contracted supplies to sell to the eastern Canadian core market. However, the net effect has been market displacement via the acquisition of direct short-term purchase supplies from Saskatchewan (and to a certain extent from B.C.) by eastern Canadian end users. The mix of markets now available covers a full spectrum ranging from short-term with no reserves dedication, to long-term underpinned by corporate warranty (with or without reserves dedication). The trend 299
DETERMINATION OFOIL AND GAS RESERVES
towards shorter term sales and faster pool depletions has caused Canada's RIP to steadily decline during the last decade from approximately 25 to just below 20. This phenomenon will likely continue during the 1990s, approaching a stable value in the range of 8 to 10 by the tum of the century. Many events currently underway will ultimately playa role in determining the market demand forces discussed. These events need to be viewed in the context of the preceding discussions. I. US storage continues to mask the true daily supplydemand relationships. The actual US deliverability capability is still difficult to determine'. 2. Canada and US statistics show sizeable replacements of production during the last 3 to 4 years, in spite of dropping prices and oversupply signals. 3. The phenomenon of "overmarketing," i.e., producers being very aggressive in marketing has led to: • NOVA excess receipt capacity • Over-subscription of NOVA export capacity 4. Supply distortions such as the US subsidy of coal seam gas development result in producer netbacks currently higher than current spot prices for natural gas. 5. Futures markets monthly closing prices are correlating consistently with actual monthly spot prices. 6. The political and regulatory "out of sync" phenomenon; the Alberta government is still proceeding slowly to establish a core market definition with Ontario. 7. Producer price expectations vs. marketplace price expectations. Very recently cogeneration projects have been proposed with fixed-price, 20-year contracts, i.e., no formulae, no "openers." 8. The FERC Mega-NOPR and subsequent implementation orders are an attempt to complete the full deregulation of gas from wellhead to burner tip by forcing pipelines to completely unbundle their sales and transportation services. 9. The "prorationing experiment" being tested by the states ofTexas and Oklahoma is an attempt to force gas supply and demand into a closer balance. This is an interesting attempt at re-introducing government regulation at the same time as the effort towards complete deregulation is still continuing.
25.2.3 Preview of Future Era The current era in Canada and the US appears to be a transition phase. The role ofthe traditional marketers is changing rapidly; market types are becoming less definable; and producers are reacting in different ways to this changing environment. It is expected that the next era of gas markets will commence when the following occur:
• Canadian core markets can purchase supplies from any source under freely negotiated terms, as opposed to terms set by regulatory bodies. • US pipelines become mostly open access carriers which allow capacity brokering. • Sufficient export pipelines are constructed out of Alberta to reduce the excess deliverability inside the province. • Buyers and sellers in the US and Canada can effect gas sales arrangements, and pipeline systems can provide transportation arrangements on a com-mercial basis only without regulatory impediments. An equilibrium phase may then be in place, to the extent that any market can be accessed by any source of supply, subject only to "ordinary" economic supplydemand relationships. As a result of the experience of the last three decades, the remaining reserves in Canada and the US may be viewed as an inventory that is continuously being replaced rather than as fixed entity.
a
25.3
MARKET MECHANISMS AND MARKET FORCES
This section discusses market forces in the context of a North American reserves base dependent only upon economics.
25.3.1 Market Types and Market Mechanisms Markets are generally of two types: core and noncore. Various Canadian provinces and US states have yet to establish formal definitions for core markets although they have been attempting to do so for several years. The definition of core markets developed in Section 25.2.1 will be used throughout this discussion. Market mechanisms are illustrated in Figure 25.3-1, which illustrates the basic steps that a producer must go through to place gas onstream to an ex-Alberta market. The following are the fundamental requirements: 1. Establish reserve deliverability; in other words, define the volume to be sold.
300
______________________1
NATURAL GASMARKETS
Producer (Alberta)
ShortAlberta-ERCB Term Removal Pipeline (Less than Permit 2 Years) NOVA
Eastern Canadian Market
Ex-Alberta Pipeline
I
Ex-cana~
Pipeline
(Federal) NEB Short-Term Export Licence
US DOE
f-- Short-Term Import fAuthorization
US Market
Long-Term Ex-Alberta Pipeline
,, ,, ,, ,,
Consuming Province Import Review
Eastern Canadian Market
Ex-Canada Pipeline I. 2. 3. 4. 5. 6. 7. 8.
Basic Requirements Transportation in Alberta ERCB Removal Permit Transportation Ex-Alberta NEB Export Licence DOE Import Authorization FERC Review US Transportation Markets
L
NEB Long-Term Export Licence
US DOE Import Authorization
FERC Review
If facilities required
US Market
Figure 25.3-1 Commercial and Regulatory Mechanisms for Ex-Alberta Markets
2. Identify a market and negotiate a purchase contract. 3. Procure transportation in Alberta to move the gas from the production point to the Alberta border. 4. Obtain an ERCB removal permit, either short-term (less than 2 years) or long-term, which requires a demonstration of reserves, market details, transportation arrangements, and an analysis ofthe social and economic impact on the province. 5. Obtain transportation outside of Alberta, either to eastern Canada or to the United States. 6. Obtainan NEB export licence, either short-term(less than 2 years) or long-term, which requires demonstrationofmarkets, transportation arrangements,and reserves and an assessment ofthe impact (optional) on the effect on Canadian energy markets of this export plus an estimate of project revenues. 7. Obtain Department of Energy (DOE) import authorization, either short-term (2 years or less) or
long-term, which requires identification of the market and specifics of the sales contract. 8. Obtain FERC approval if interstate transportation facilities are required. The foregoing very briefly outlines the mechanics involved in selling volumes to an ex-Alberta market. The degree of complexity of regulatory approval applications, will obviously be a function of the size of the market and the amount, if any, of new transportation facilities required in Canada or the US. The marketing options are illustrated in Figure 25.3-2, which shows some of the options available to a producer. Most important from a reserves point of view is the degree to which liabilities are passed through to the producer and whether the liability is satisfied by reserves dedication alone or whether a corporate guarantee must be made as to supply make-up should a shortfall occur. The manner in which reserves and markets are linked is shown in the simple examples of Figure 25.3-3. In
301
DETERMINATION OF OIL AND GAS RESERVES
Aggregator Pro ducer
I
• • · · •
· • · · • ·
Takes title Pooled supply & markets Market swings levelled out Regulatory permits Transportation
Broker / Agent Prod ucer
I
• Pass-through title · Obtains permits in producer's name · Mayor may not hold transportion
End-User Residential Commercial Industrial Electrical utilities Cogenerators May hold transportation
Markets Core or non core may hold transportation
Figure 25.3-2 Gas Marketing Options Figure 25.3-3a, the reserves of an individual producer (or producers) provide the underpinning or supply security to the customer either directly or through an aggregator, broker or agent. This supply security mechanism can be one of two main types: a reserves underpinning or a financial warranty. The reserves comfort is based upon the pooling procedure ofan aggregator or, if long-term (greater than 2 years in Alberta's case) export licences are required, then the appropriate regulatory bodies will be "inserted." The effect of a regulatory "screen" in the sales arrangement is to provide a second security blanket for the customer. Regardless of the particular mechanism in place, the producer's reserves still provide the overall supply assurances. In Figure 25.3-3b, the customer is provided not with a reserves base for security but rather with a financial guarantee on the part of the supplier to pay for replacement volumes in the event of a supply failure. In this case, the assumption is that sufficient gas supplies are available and the question ofobtaining replacement volumes is simply a matter of paying for them. Therefore, provided that there is overall comfort as to the reserves stock, and total daily delivery capability, there is no apparent need for a direct connection between the producer's reserves and the customer. By choosing a more direct contractual link to the end user, the producer has done several things:
• Established more direct control over the value of his reserves. • Reduced the "aggregated pool" effect, i.e., the individual producer's particular reserve entity will not be rolled into a pool where it would share in the take level and price characteristics of a basket of markets which the aggregator has. Thus the specific reserve entity will enjoy a unique price-and/or-take level market. • Placed a portion of his assets, either from a reserves value or other financial asset point of view, at risk for the market. Failure to deliver either on a daily or contract term basis will make a producer liable and directly impact the producer's financial situation. Thus the upside of direct marketing vis-a-vis take level and price must be balanced off against the downside, namely, corporate warranty costs and demand c~arge payments to pipelines, all of which impact negatIvely on the value of the producer's reserves.
25.3.2 Market Demand Forces The foregoing discussion has laid the groundwork for a discussion of market demand forces and their specific effect on reserves and reserves value. From a global perspective, the North American reserves-market connection takes on a "chicken and egg" charactenstIc (as indicated also in the diagram that follows):
302
_____________________c.
NATURAL GASMARKETS
(a) Reserves Underpinning
Producer Reserves
or via Broker or Agent
Aggregator Pooled Reserves
Volumes
Producer Reserves
Customer (Individual or Pool)
Volumes Directly Dedicated
Individual Producer Reserves
Volumes
Aggregator Pooled Reserves
Volumes
Customer (Individual or Pool)
Volumes
Regulatory Body Licences, Permits
Volumes
Customer (Individual or Pool)
(b) Deliverability Warranty or Corporate Warranty Underpinning Producer
Deliverability Warranty (Payment For Replacement Deliveries)
Customer
Figure 25.3-3 Reserves Connection to Markets • The market actions establish the value of the reserves, and affect exploration, development, and acquisitions.
has not been quantified here, but amounts to a sizeable dollar figure.
• The reserves characteristics provide the quality assurances that will support a high quality (meaning high price) contract. • The size of the reserves stock pushes on the market, creating the forces that re-establish the value of the reserves.
It is reasonable to assume that reserves in remote geographical locations, such as Canada's arctic regions, have likely been discounted severely during the last 5 years. This loss in value translates directly into a real loss of proven reserves. In other words, reserves formerly classified as potential and probable and any quantities of proven reserves burdened with a high development cost have been removed from the reserves stock.
Price and Price Trends : - I Demand and Demand Trends
~
I
Producer Actions - Technology - Cost rationalization
I
Market Actions ~
Reserves Stock and ' - - - - - Delivery Capability Commencing with Canadian deregulation in 1986, and continuing with the issuance of key FERC orders from 1984 onwards, the market forces have triggered a series of continuous downward revisions to asset values throughout North America. This overall loss of value
During this same period of time, however, significant cost reductions and technological improvements in exploration and development have resulted in new reserves additions and reserves appreciation to existing pools; these have mitigated and perhaps even outweighed the discounting effect of decreasing market prices. The net effect has been a virtual replacement of gas production each year in the United States (which is evidenced by the static RiP ratio during the 1980s), and replacement of a significant fraction of Canadian production each year during the 1980s. Therefore, the quantum of the ultimate stock of North America's remaining proven
303 !. ..
"".,""'"" osOL ' " ' "
",,::-;III'
plus probable plus possible reserves acts as an overall dampener.
overall producing sector responses to changes i demand and price. n
This apparently endless supply reduces the quality and hence the degree of force that an individual producer or group ofproducers can exert on the market, resulting in stronger market forces and forcing the seesaw in the market's favour.
Developing a production forecast on an individual_ producer-of-individual-field basis is somewhat more straightforward. A producer would have to consider the following variables in developing a production forecast: I. Gas purchase contract conditions
25.3.3 Production Forecasting
• Price expectations
Production forecasting is the act of reconciling the expected market demand environment with the expected supply environment. The market demand scenario has already been developed in this chapter on the assumption that the North American gas supply base is large enough to be considered unlimited. Thus production forecasting, from a global point ofview, appears to be a reasonably simple task. One need only assimilate and accurately forecast the interactive effects of all of the regulatory and political market demand phenomena described in this chapter, and then match production capability to this complex model!
• Take level expectations
A simplistic total North American production forecast would be based upon the following: • An established forecast ofrequirements and expected market price • An estimate of established reserves and productive capability
• An estimate of supply costs • An estimate of future finding rates-perhaps best derived by statistical analysis ofdrilling activity, discovery rates per well, drilling costs-all modified by gas market prices It follows that a supply-demand curve can then be drawn as the intersection of consumer gas demand with supply capability.
A more sophisticated forecasting model would incorporate such variables as: • Pipeline restrictions • Political decisions or government edicts that set prices, production rates, or fuel substitution • Complex demand forecasts for each type of gas user • Individual forecasts for major pools or producing regions Regardless of the degree of sophistication of the production forecast mode, simplifying assumptions must be made regarding price, consumer demand, and
• Daily volume nomination levels and modifica_ tions to this level (i.e., restoration ofvolume levels previously reduced by deliverability testing) • Some expectation ofthe purchaser's performance in view of all of the global market demand and regulatory items 2. Existing field conditions • Reserve and deliverability information • Drilling, completion, and well tie-in costs • Operating costs 3. Undeveloped field conditions • Capital and operating cost expectations 4. Producer corporate objections • Financial health and corporate economic guidelines • Type of reserves available to produce, e.g., raw land, which requires "full cycle" costs, including land; geophysical drilling, including risk; and surface facilities costs or acquired reserves, which may have been purchased at a discount from the full cycle costs and thus may be connected at relatively less cost The field forecast can then be prepared by matching individual well or field physical capability and market demand with corporate economic criteria. The producer can then use this forecast to make the appropriate drilling and other capital expenditure decisions.
25.4
THE ROLE OF RESERVES
The foregoing discussion of market forces and production forecasting has provided at least a "thumbnail" sketch of the process. It is worth noting what has happened to the role ofreserves since deregulation. Reserves are one of the cornerstones, directly or implied, of a sales contract between a buyer and a seller. Any user of gas must have some degree of assurance as to supply continuity. Reserves can take on different roles in the marketplace, as follows:
304
c
NATURALGAS MARKETS
Negotiating Tool. A large reserves stock owned by a producer or controlled by a broker or aggregator carries a significant weight in sales contract negotiations. Supply Indicator. A buyer would naturally be interested in the seller's reserves supply, but the absolute size of the reserve base may be ofless importance than the seller's replacement ability. Security. An adequate reserves supply, or adequate assurance of replacement ability, provides support for financing the construction of facilities such as pipelines and end user equipment. Reserves-to-Production Ratio Yardstick. The reserves-to-production ratio, RIP, of a seller provides an indication of efficiency, longevity, and marketing strength as well as perhaps relative financial health. The RIP ratio for a region or country gives a picture ofoverall security of supply, as well as an indication of emerging trends such as oversupply or undersupply problems. An understanding by buyers of the RIP for producers and aggregators or brokers and countries as a whole is crucial to establishing sound contractual arrangements. A buyer's needs for supply security need to fit
a producer's needs for economic turnover of his inventory ofreserves. Currently, most producers have a need to sell reserves at approximately a 7 to 12 year turnover rate, or a I :3000 to I :4500 RIP ratio. Sellers, on the other hand, can have a need for 20 or even 30 years of dedicated reserves. Thus, sales arrangements need to satisfy both these needs.
25.5
CONCLUSIONS
Market demand forces are obviously the very basis for the gas industry; markets are simply the demand for the commodity the gas producers are trying to sell. The demand forces are not necessarily in balance, from a classical textbook supply-demand law consideration, with either the customers' or the producers' needs. Production forecasting is becoming an increasingly difficult task. A forecaster needs to consider all of the following: regulatory and political uncertainties, merchant pipeline deregulation, government jurisdictional disputes over market deregulation, market price uncertainties, and changing relationships between finding costs, technology, and producers' corporate objectives. The complexities of the physical characteristics of the reservoir may be ultimately less difficult to forecast than the market factors.
305
Chapter 26
USES OF RESERVES EVALUATIONS
26.1
INTRODUCTION
The estimates of reserves and forecasts of production from those reserves are widely used within the oil and gas industry and by organizations working with the industry. The users fall into two well-defined groups: the first uses only the volumes resulting from the process (Section 26.2); the second uses the values derived from the cash flows generated from the forecasts ofproduction and other variables (Section 26.3). The development of various economic yardsticks is explained in Section 26.4.
26.2
USERS OF RESERVES VOLUMES AND PRODUCTION FORECASTS
A wide range of users requires only estimates of reserves and forecasts of production to operate their businesses and make their decisions. The users with this requirement would include the following: Producers, who use the forecasts of production and life of facilities to size the equipment put in place to produce the reserves and to deliver them to their market Pipeline companies, who require estimates of both current and potential reserves, together with forecasts of current and potential maximum future production, to decide whether new or extended pipelines could be justified and, if so, the sizes needed Governments, who need information on reserves and production to implement and modify legislation and policy on resource development and security of energy supply Gas marketers, who need to know how much gas they have under contract and the physical limits on the way in which their contracted gas can be produced
26.2.1 Producers Producers of oil and gas must have realistic estimates of recoverable reserves and forecasts ofproduction to plan the development of hydrocarbon discoveries.
The sizing of equipment, from the diameter of the well drilled to the size of the plant required to process the production stream, and everything in between, depends upon the level of reserves and the resulting production. It is important that the estimates be as realistic as possible, as over-enthusiastic estimates result in unneeded capacity and expense, while low estimates result in bottlenecks that limit production and are costly to overcome. It is important to use reserves estimates based on detailed engineeering studies.
26.2.2 Transporters Companies (either producing or pipeline) that transport oil and gas to their final markets need to know both the current level of reserves and the production their facilities will handle, and also the ultimate reserves and production their facilities may be required to move. Pipeline companies need reserves and production data not only for their own planning and development, but also for supporting government applications to build new pipelines or expand existing ones. The building ofpipelines requires government approval to prevent their unnecessary proliferation and, as a consequence, their tariffs are set by the government in dollars, recognizing their operating costs, the recovery of capital, and a return on any unrecovered capital costs. Projections of future tariffs require forecasts ofproduction to estimate the amount ofproductthat will be moved through the facilities; from this, charges per unit ofproduct moved can be predicted. Pipelines that are under- or over-sized may result from inadequate estimates of current and future reserves and production.
26.2.3 Governments Governments must have detailed knowledge of the currently developed resource base and also the ultimate resource that is available for development in order to plan the best legislation and policies for resource developments.
306
s
USES OF RESERVES EVALUATIONS
A knowledge of both present and future reserves is necessary for the best resource development legislation and policies, and also for the development oflong-range energy policy. Identifying sources of supply to meet future demands is dependent upon realistic forecasts of current and potential reserves.
26.2.4 Gas Marketers Brokers of natural gas must know with the greatest precision possible the reserves they have under contract and the changing rate at which the reserves can be produced, so that they can match these reserves with the markets they are servicing. Marketers also need estimates ofreserves and forecasts ofproduction for government approvals ofexports from the producing provinces and also of exports to other countries. The use of unrealistic estimates of reserves can lead to either excessive reserves with no markets or a shortage of supply, so that markets are lost.
26.2.5 Other Users Many other groups benefit from estimates of hydrocarbon volumes, including stockholders, the investment community in general, and consumer groups.
26.3
DEVELOPING VALUES FROM RESERVES ESTIMATES
The cash flows that result from estimating reserves and forecasting production, prices, royalties, costs, and taxes are discounted to provide a range of net present values. This provides a very powerful tool for investment decisions, whether they are internal capital investment decisions regarding land, drilling, or production facilities; decisions to buy or sell producing properties; decisions to acquire or merge with other corporations; or decisions to lend money using future production as the security. If realistic discounted forecasts offuture cash flows are not included in the information on which investment decisions are made, the investments are not likely to be intelligent or successful. The Canadian oil and gas industry runs on discounted dollars. This situation is not unique to the Canadian oil and gas industry. It must always be remembered that the estimation of reserves should never be considered as the end of the evaluation process. Unless the resulting economics are understood, reserves may be incorrectly estimated, and the risks and uncertainties may not be fully understood. The result may be reserves that are incorrectly classified as proved, probable or possible.
In corporate investment decision-making, the most common internal investment decisions involve the drilling ofdevelopment wells and the building ofproduction facilities. While risk may exist with these investments, it is usually not addressed explicitly, but the minimum value used to accept or reject an opportunity is set to be sufficiently conservative to allow for the risk. In this way, the targeted rate of return can be achieved on a corporate basis.
26.3.1 Profitability Indices The following are some ofthe profitability indices used in Canada (but certainly not the only ones): • Payout period • Return on investment • Rate of return • Discounted return on investment • Net present value The use and limitations of each index are described in the following subsections. Payout Period
Payout period is the time required to recover the investment in a project. It can be calculated on a discounted or undiscounted basis. In some cases discounted payouts are used in the evaluations. It is important to note that payout is measured from the time ofthe first significant investment, not from the time at which production starts. Payout period is usually one of the indices used in the selection of investment opportunities. A commonly used index is a payout period that is a maximum of 3 years for a development project (where there is some risk) orno more than 4 to 4Y2 years for fully developed producing properties (where the risk is considered to be minimal). The choice of payout period, however, varies from investor to investor. Ifthe payout period for a development project is greater than 3 years, it should not necessarily cause the project to be rejected, but should be a warning that the project must be better than average to overcome the longer period required to recover the investment. A useful rule of thumb is that projects with payout periods ofgreater than 3 years should not represent more than 10 percent, or at a maximum 20 percent, of the total investment budget in anyone year. The company has to be able to stay in business until payout is complete.
307
~I DETERMINATION OF OIL AND GAS RESERVES
Only those investments required prior to payout are included in the calculation of payout period. It is usually calculated using the before-tax cash flow. Return on Investment
Return on investment (ROI) is also known as the undiscounted profit-to-investment ratio. It is simply defined as the undiscounted net revenue (also called cash flow) divided by the initial investment. It has the serious disadvantage of ignoring the time value of money, but is still used by some in selecting suitable investments. It can draw attention to projects with short lives or unusual cash flow profiles if it yields a value that is not in keeping with the other indices calculated.
becomes the rate of return at which the cash flow must be discounted to give a present worth value of zero. This index takes the time value of money into account , but suffers from two serious deficiencies. First, the rates of return can be very high (in the order of 30 to 70 percent) with the result that little weight is given to any cash flow after 10 years. This causes projects with high cash flows in the short term to appear to be more desirable and may cause long-life projects to appear to be less desirable, with the result that the wrong investment decisions may be made.
The ROI may be calculated for a project requiring investment on either a gross basis (the investment is not included in the cash flow) or a net basis (the investment is included in the cash flow). The normal procedure is to calculate the net ROI, but the gross ROI will be exactly 1.0 more than the net ROJ. A hurdle of $4.00 per $1.00 invested (ROI =4.0) is often used for the net ROI, but this may vary between companies. A net ROI of less than 3.0 should raise questions about the desirability of an investment project.
Second, each project has a different ROR, so that different weights (discounted values) are given to dollars earned at the same time in the future. Projects being compared also have different lives so it becomes a problem to decide if a project with a ROR of 50 percent and a IS-year life is better or worse than a project with a ROR of 40 percent and a 25-year life. Assumptions on re-investment can be made for the project with the shorter life to try to solve the problem; however, a better measure of the relative desirability of investment opportunities is to calculate the discounted return on investment (DROl), as described in the next subsection, in which a constant discount rate is used.
The reason that the ROI is not recommended as a limiting index is that the cash flow profiles for oil industry investments usually vary significantly over their lives; consequently, the relationships between cash flow in each year vary widely. In industries where the annual cash flow relationships from various projects remain relatively constant, the time value of money may be downplayed. This is not the case in the oil industry.
The ROR is a good and commonly used hurdle for culling projects that are to be rejected from those that are acceptable. In Canada, the minimum RORnormally required would be in the range of 15 to 20 percent, using the after-tax cash flow, the most common minimum being 18 percent after taxes. The rate ofreturn required is a function of the cost of capital and the overall risk involved in investing in oil and gas.
The ROI can be calculated using the gross cash flow (without deducting the investment), or the net cash flow (after deducting the investment), or the total investment required, or only the initial investment, but there is no one "correct" way. The important thing is to calculate all ROIs in a particular company in the same way.
Discounted Return on Investment
The discounted return on investment (DROl) is defined as the present worth value ofthe future income using an externally- derived discount rate, divided by the investment, or:
Rate of Return
In recent years, rate of return has been one of the more widely used profitability indices and has been called by many different names, including internal rate of return, internal yield, discounted rate ofreturn, discounted cash flow rate ofreturn, profitability index, and marginal efficiency ofcapital. In this chapter, the term rate ofreturn (ROR) will be used, and it is defined as the rate at which the future cash flow must be discounted to give a present worth value that is equal to the investment. If the investment is included in the cash flow, then the definition
308
DROI=
present worth . mvestment
This index is also known by many names including the discounted profit-to-investment ratio (DPR), the present worth ratio (PWR), and capital productivity index (CPI). The discount rate that should be used is whatever the marketplace is using to determine the value of oil and gas assets. This rate is currently in the order of 12 to 15 percent, applied to the after-tax cash flow stream; however, the rate will change as the following change:
'~,
I ,
I
USES OFRESERVES EVALUATIONS
the cost of capital, the perception of risk in forecasting prices, and the supply of and demand for properties. The DROI is a very powerful index for the ranking of investment opportunities. If the current marketplacederived discount rate is used, then the value of the numerator in the DROI ratio represents the market value of the investment being considered. The denominator is the investment required to generate this value. Consequently, the index is the number of dollars of asset value to be added per dollar of investment. The DROI overcomes the problems inherent in the ROR as a ranking tool since the same discount rate is used for all projects, and the discount rate is realistic. If the investment is included in the cash flow stream (from which the value in the numerator is derived), then a value of zero indicates a rate of return equal to that used to discount the cash flow. Any positive value indicates a rate ofreturn greater than the discount rate used. Conversely, a negative value indicates a rate of return that is less than the discount rate used and does not necessarily indicate a loss. Sometimes, the DROI is increased by \.0 ("DROI plus 1.0") to calculate the index when the investment has been included in the cash flow stream. In this way, the investment is removed from the cash flow stream, which in effect is adding an amount equal to the denominator (the investment) to the numerator (the value). It is measuring the value of the asset after the investment has been made as a function ofthe investment required. The "DROI plus 1.0" index is the multiplier to be applied to the investment to give the value of the asset expected to be developed as a result of the investment. Ifthe "DROI plus \.0" index is used, then a result of \.0 means that $\.00 of value is being added for each $\.00 spent. Ifthe value is greater than 1.0, then an asset with more value than the investment being made is expected. Because there is some risk with most investments in the oil business, a DROI that is considerably greater than zero ("DROI plus \.0" considerably greater than \.0) would usually be required before an investment would rank above other investment opportunities. Capital availabillity certainly sets the minimum DROI, but the usual range would be from \.0 to 3.0 (2.0 to 4.0 for "DROI plus \.0"). A low risk opportunity, such as the purchase of producing reserves, would usually have a DROI of \.0, if the investment is not included in the numerator; that is, the value ofthe asset acquired is equal to the investment.
The value in the numerator is the asset value and, therefore, all future investments must be included in the cash flow; however, there is some argument as to whether the investments used in the denominator should be only those required to start the project or the total of the investments (suitably discounted) required during its life. The DROI is used as a ranking tool and is used to select projects for inclusion in the current budget and, for this reason, it can be argued that only the funds to be budgeted to the star! of production should be used in the denominator. Ifthis is correct, then the use ofall the investments (discounted, of course) in the denominator could cause the rejection ofdesirable projects. This could be a serious problem with the use of the DROI. It should be remembered that the DROI is the value of the asset added divided by the cost required to add that asset. If the "DROI plus \.0" is used, then none of the investment in the denominator is included in the numerator, if only the investment to the start ofproduction is used as the divisor.
Net Present Value
The present value equation compares income and costs at time zero. Net present value (NPV) is equal to the present value calculated at a selected discount rate minus the investment. Using Cash Flow Forecasts in Investment Decision-Making
When the value of an investment opportunity is being determined, sunk costs are never taken into account. If they are, the wrong investment decision may be made. An example ofthis would be the decision to complete a well uphole when the lower target is found to be dry. The well may have cost $1,000,000 to drill, but the completion of an uphole zone for $100,000 would add an asset worth $200,000. Ifthe sunk costs are included, a request to spend $1,I00,000 for a $200,000 asset would be rejected. The correct request would be for $100,000 to develop the $200,000 asset. At the end of the day, if the $100,000 is approved, the net out-of-pocket cost would be $900,000 ($1,000,000 plus $100,000 less $200,000), but if the additional $100,000 is not spent then the cost would be $1,000,000. Sunk costs cannot be recovered so they are never included. Only future revenue and costs should be included in the basis for an investment decision. An interesting down-side check on an investment opportunity is to use trial and error to determine, using constant prices and costs, the minimum product price
309
DETERMINAnON OF OIL AND GASRESERVES
that will yield an acceptable rate of return, usually 10 percent after taxes, because ofthe use ofconstant prices and costs. This test will give an indication of the sensitivity of a project to price variations. If a project cannot tolerate much reduction in price, it may be too risky to pursue. Most cash flow forecasts used to justify an investment opportunity consider the incremental effect that the new project will have on the total corporation. For this reason, only those incremental costs and revenues that result from the project are included in the cash flow forecast. General and administrative expenses (G&A) usually will not increase as a result of the addition of the incremental charges created by a new project (such as a well or gas plant) and, therefore, no G&A costs are included. However, if the incremental charge being considered will result in increased or even decreased G&A charges, then these charges should be included in the cash flow forecasts. There is an argument that some G&A should be included, ifnot in the first year, certainly after that time. While the current wells (excluding the incremental increase being contemplated) will meet the current G&A, the new project will be charged with its share in the future. Of course, those entities currently meeting the G&A cost will disappear with time and only the new investments will be available to meet the G&A. This point should be considered, as ignoring future G&A costs could present a more optimistic basis for a new incremental investment than is really justified. It is important to note that charges for depreciation, depletion and amortization (DD&A), which are included in financial statements, are noncash items and are, therefore, not included in the cash flow forecast. DD&A is an accounting means of charging the revenue stream with a share of the capital required to generate the revenue stream. Any capital costs are included in the cash flow forecast in full at the time at which they are incurred.
26.3.2 Incremental Economics This section does not refer to the incremental nature of most investment opportunities analyzed from a corporate perspective, but deals with the method that should be used to evaluate an additional investment that may enhance the performance of an existing project. For instance, a waterflood will increase the recovery from an existing pool as the result of additional investment in injection wells and water-injection facilities.
310
How should the incremental effect of the additi I . bd ined?. ona investment e etermine The procedure is simple. A forecast of cash flow' e . 1if no mvestment . IS prepare d lor the project is made (primary case). A cash flow forecast is then prepared fo the project ifthe investment is made (waterflood case)f with the investment being included in the cash flow fore~ cast. In the next step, the cash flow for the primary case is subtracted from the cash flow for the waterflood case. The resulting incremental cash flow provides the basis for making the investment decision. Because the investment is included in the incremental cash flow forecast, the rate of' return is the discount rate which gives a present worth value of zero. Incremental investment opportunities often yield high rates of return. The incremental annual cash flow determined by subtracting one cash flow from the other will usually be negative for some of the life ofthe incremental project. This does not mean that the project is losing money and should be discontinued, but that the upgraded project, after the investment, has lower cash flows in later years than would have been the case had the incremental investment not been made.
26.3.3 Acceleration Projects An acceleration project is one in which an investment is made, not to get any additional revenue, but merely to get the same, or even a reduced, cash flow sooner: then on a discounted cash flow basis, the investment has the potential to earn a rate of return. The big problem is that true acceleration projects usually have at least two rates ofreturn. The investment made only brings the cash flow ahead. It does not add any extra cash flow and, in fact, the cash flow may be less through lower prices and higher costs. Coupled with the'Tnvestment, this will result in a negative undiscounted cash flow for the incremental acceleration investment. The investment required for the acceleration project is evaluated in the same way as any incremental investment; the cash flow forecast for the project without the investment is subtracted from the cash flow for the project if the investment is made. As the discount rate is increased, the present worth value goes from negative to positive. Experience has shown that this usually occurs at discount rates of 5 to IS percent. As the discount rate is further increased, the present worth value continues to increase (the opposite of the normal effect of discounting) until the discount rate generally gets to the range of 35 to 50 percent, at which
USES OF RESERVES EVALUATIONS
point the present worth value begins to decline. It returns to zero at a discount rate usually in the order of 100 to 300 percent. These general ranges do not apply to all projects. By definition, the rate of return is the discount rate required to give a present worth value of zero, which occurs usually between 5 and 15 percent and between 100 and 300 percent; that is, there are two rates of return. Which is the "correct" rate ofreturn? Well, no one has answered that question yet. One appears to be too low and the other too high. The only practical way to check an acceleration project appears to be to determine its sensitivity to price reductions. If a small reduction in price results in no positive values for the discounted cash flow, then the project may not be worthwhile. The problem is to make sure that an acceleration project is recognized as such by the people making the analysis and the investment decision. If a present worth value at a selected discount rate is used as the ranking index (such as a DROI at a 15 percent discount rate), an acceleration project might not qualify, but if a higher discount rate is used, the DROI would increase and the project could become acceptable. For this reason, all projects (whether acceleration or not) should have a profile of present worth values calculated for a range of discount rates (usually from 0 to 50 percent), so that multiple rates of return and increasing present worth values with increasing discount rates can be recognized. Most acceleration projects are masked by the inclusion of additional cash flow with the accelerated cash flow; in fact, most incremental projects include an element of acceleration. A waterflood project includes increased rates of production (that is, acceleration) as well as additional recovery of reserves, The most common acceleration project in Canada is one in which infill wells are drilled. Such an opportunity could be a true acceleration project if it only results in getting the same reserves sooner with the increased well count. Normally, there is also an increase in the reserves to be recovered and this will mask the acceleration aspects of the project. The main thing to remember is that the new wells will affect the production performance and reserves to be recovered from the existing wells. For this reason, the economics of the new infill wells should never be looked at on a stand-alone basis. The project should be evaluated by looking at the performance ofthe total project without any infill drilling and subtracting that cash flow from the cash flow that will
result from the total project when the infill wells are drilled. This incremental cash flow is the way in which the investment should be analyzed.
26.4
USES OF THE VALUES DERIVED FROM RESERVES ESTIMATES
26.4.1 Valuing Oil and Gas Companies The assets of an oil company consist of its inventory of reserves of oil, gas and related products and its land holdings. The balance sheet of an oil company records the value of the reserves and lands as the price paid to acquire or develop those assets. However, the price paid to develop reserves or purchase unexplored lands is unlikely to be an indication of their current value. To overcome this problem, the value carried under "Property, Plant and Equipment" on the asset side of the balance sheet should be replaced with the going concern value of the remaining reserves and lands. Goingconcern value (GCV) is the value assigned to an asset that is already owned and that will be kept for future exploitation. It is not the value to be expected ifthe asset is sold directly. The determination of the GCV of reserves would be based on a projection of future cash flow, using estimates of reserves and forecasts of production, costs, prices and royalties for each separate interest. These cash flows would then be summed and an after-tax calculation made, using not only future investments to protect income from taxes, but also any unused tax pools held by the company such as: • Canadian exploration expense (CEE) • Canadian development expense (CDE) • Canadian oil and gas property expense (COGPE) • Tangibles subject to capital cost allowance (CCA) Any prepayments to be repaid in the future, and all abandonment and reclamation costs, would also be included in the cash flow forecasts. The after-tax cash flow discounted at an acceptable rate, currently between 12 and 15 percent, would give the value to replace the remaining undepreciated investment carried under "Property, Plant and Equipment." All future taxes and costs are included in the cash flow forecast, as well as any repayments of prepayments received, so the amounts carried as "Deferred Taxes," "Site Restoration Costs" and "Deferred Revenues" on the liability side ofthe balance sheet should be removed. The value of any unexplored lands would then be determined and added to the value of the reserves. No
311
value would be assigned separately to the production facilities, such as gas plants and treating facilities, as the value of these facilities is included in the value assigned to the reserves. The only time that a facility would be valued separately would be when the interest in the facility is not the same as the interest in the reserves being processed through it. The sum of all the assets less all the liabilities would then give the GCV of the corporation. The value per share would be this value divided by the number of shares outstanding. This would be the basis used to make an offer to acquire an oil and gas company.
26.4.2 Sale of Resource Properties The Canadian definition of resource properties for income tax purposes includes oil and gas reserves (and the wells to produce them) and unexplored lands. The direct purchase of resource properties (but not the purchase of the shares of a company that owns resource properties) creates a tax advantage for the buyer in the form of the Canadian oil and gas property expense (COGPE). This is equal to the price paid and may be written off at the rate of 10 percent ofthe declining balance each year against all taxable income from any source. The creation of the COGPE by the act of buying the resource property means that a buyer is acquiring two income streams, one from the production and sale of the oil or gas, and the other from the tax savings generated by the write-off of the COGPE. Because of the competitive nature of the marketplace, the buyer must be prepared to pay the full value ofboth the reserves (or potential reserves in the case of unexplored lands) and the COGPE, ifhe expects to acquire the property. This value would be described as the market value (MV) of the asset. A term often used is fair market value (FMV) which is defined as the price that a willing buyer would pay to a willing seller, ifneither is under any compulsion to buy or sell and both are competent and have reasonable knowledge ofthe facts. Because all of these limitations are seldom met, the term market value (MY), rather than fair market value, has been used. The value of the COGPE is a function of the value of the reserves (or lands) and the write-off rate, tax rate and discount rate. Equation (I) gives the multiplier, F, to be applied to the value of the reserves (or lands) to determine the value of both income streams (using midperiod discounting):
312
F=
I 1-(WRxTR)x (I +i)'"' WR+i
where WR TR Thus
MV
= =
0)
fractional write-off rate fractional tax rate fractional discount rate GCVxF
(2)
where MV market value ($) GCV = going concern value ($) F = multiplier In Alberta, where the current total tax rate is 44.34 percent, the multipliers would be: 1.2711 for a 12 percent discount rate 1.2349 for a 15 percent discount rate With the value of the COGPE adding 23 to 27 percent to the value of the reserves (or lands), it is too big to be ignored. Similarly, if a COGPE is not created because the corporation, not the resource property, is purchased, a value that includes the COGPE will be far too high. These distinctions must be understood if the correct values are to be assigned. The seller of properties that create a COGPE for the buyer is required to take the total proceeds of the sale (not just the gain in value) into income for the purpose of calculating the taxable income in the year in which the sale occurs, and it is income that is not eligible for the resource allowance; however, up to 100 percent of any unused Canadian development expense or COGPE balances may be used to offset the proceeds ofthe sale, in addition to 100 percent of any unused CEE balances. Often a seller ofproperties will buy properties with approximately the same value before or in the same year as the sale. Ip this way, he has a COGPE balance that he can use to render the sale a tax-free event. A complicating factor in the determination of MV is the fact that part of the price paid is classified as the price of the tangible assets (generally, the production facilities) acquired with the resource properties. These tangible assets can usually be written off at the rate of 12.5 percent of the value in the year of acquisition and at 25 percent ofthe declining balance each yearthereafter. Some tangible assets are written off at other rates. The multiplier for this higher write-offrate would be in the order of 1.32 to 1.38 for discount rates ofl5 and 12 percent. The fraction of the total price considered to be
USES OFRESERVES EVALUATIONS
for the purchase of tangible assets is negotiated by the buyer and the seller, but the fraction must be reasonable to be acceptable to the taxman. Because of the different tax consequences of a direct purchase and a purchase of a company owning oil and gas assets, it is not recommended that GCV and MV be determined from before-tax cash flows. The MV for oil reserves in Alberta is generally found to be equal to a before-tax cash flow discounted at 20 percent (this is the same value that would normally be determined by discounting the after-tax cash flow at 12 percent). In Saskatchewan, a cash flow before taxes from an oil property would have to be discounted at 25 to 30 percent, depending on whether the Crown royalty is new or old, to give a 12 percent after-tax rate of return. Similarly, the cash flow before taxes for gas properties in both Alberta and Saskatchewan should be discounted at 18 percent to give a value equal to the after-tax cash flow discounted at 12 percent. If GCVs are to be based on before-tax cash flows, the discount rate would range from 22 to 34 percent, depending on the province and royalty classification. It is dangerous to ignore taxes when determining the values of Canadian resource properties.
26.4.3 Evaluation of Unexplored Lands and Exploration Wells The procedure for estimating the value of interests held in unexplored lands by oil and gas companies starts by determining the value of discoveries of oil or gas reserves that could be made on each spread oflands. After this value is determined, the probability of success is determined. Then, the cost of failure is estimated; the probability of failure is equal to one minus the probability of success. The difference between the product of the value and probability of success and the product of the cost and probability of failure gives the riskweighted value, or expected monetary value (EMV) of the land on which oil or gas may be found. It must always be recognized that the prices at which lands that are close to those being evaluated have traded may differ greatly from the value determined using this procedure because of variations in the geological prospects under the lands. The value of success would be determined in the same way that reserves are valued. The size of the reserves that can be expected to be discovered would be estimated, the production forecast, and an after-tax cash flow forecast prepared and discounted as previously described. The cost of failure would be the cost of a dry
hole, after taking into account the value of any tax advantages resulting from the expenditure. The values of unproved properties are based on aftertax values and costs. Consequently, the value ofany tax advantages to be gained by exploring for and developing reserves in the future is either included in the values assigned to the lands presently held or will be included in the prices paid for lands that are acquired in the future. If the values assigned to the reserves ofa company are based on the assumption that future investments in exploration and development will eliminate any tax payments on income from the production of the reserves, then the value of these future tax advantages will have been taken twice. When unexplored lands are purchased, the price paid not only represents the risk-weighted value these lands are expected to contribute, but also includes the value of the COGPE created by the acquisition. Consequently, iflands already held are being evaluated, their value will be equal to the price currently set in the marketplace less the value of the COGPE, which can be determined using the formula set out in Section 26.4.2 on the sale of resource properties. If the price paid in Alberta is $100 per acre, it would represent $78.70 for the land and $21.30 for the COGPE if a 12 percent discount rate were used. These values would change to $81.00 and $19.00, respectively, with a 15 percent discount rate. Exploratory wells would be evaluated in the same manner. The value and probability of success would be determined and the cost offailure, after taxes, estimated. Then, the EMV would be calculated as the difference between the product ofthe value and probability ofsuccess and the product ofthe cost and probability offailure (i.e., one minus the probability of success), The EMV is the risk-weighted contribution the exploration well could be expected to make to the assets of the corporation, in excess ofthe investment required. If the exercise is repeated frequently .and the probabilities of success are assessed realistically, then the sum of all the EMVs for a year should be the level ofvalue ofthe assets added by exploration in the year, in excess of the investment required. A possible index for ranking exploration wells would be to divide the EMV by the after-tax dry hole cost. This is similar to the DROI (Section 26.3.1) and is commonly referred to as the profit/risk capital ratio. Another index that can be used for project ranking, the profitability index (P/I), is calculated by dividing the EMV by the maximum capital exposure (present worth
313
c
DETERMINATION OFOIL AND GAS RESERVES
value of total capital, after taxes), which indicates the expected net dollar value to be received for each after-tax dollar of capital expended.
26.4.4 Lending and Borrowing Financial institutions are permitted by law to take a pledge of reserves and assignment of production proceeds as security for funds advanced to the production owner. The following discussion effectively assumes that a term loan is made available for a particular reserve. (In the case of a company with diverse production interests, annual principal payments may not be required.) To assess the maximum loan that can be secured by production, the lender may look at a series of values to determinethe amount of the loan.A lendermay require the production offered as security to come from more than one well, each of which has a production history. Forecasts of reserves, production, and cash flow, usually supplementedby an independent engineer's report, are used by the lender. The loan life is normally limited to no more than the periodrequiredto recoverhalf the reserves (the reserves half-life), with the loan amount as a percentage of the discounted value of production. Loans secured by production would normallybe based on reserveshalf-life witha maximum term that is establishedby the lendingpolicy of the financial institutions. Reasons for including the reserves half-life are (I) to ensure that when the loan is repaid, the property will still have sufficient value that the borrower will look after it while repaying the loan, and (2) to ensure that the loan will still be repaid if the reserves estimate is not accurate. If there is little value left when the loan is repaid, the lender cannot expect the borrower to have much interest in the property. When the amount and the term of the loan have been determined, the annual minimumprincipaland interest payments are determined under different repayment optionsandpossiblyunderdifferent priceforecasts. Due to the risk of changes in legislationand tax, any Alberta royalty tax credit is calculatedseparately. The price to be expected from a distressed sale of the properties may be estimated by the lender in a number of ways. It may range from using 80 percent of the after-tax cash flow discounted at 15 percent, with the COGPEadded in, to taking 50 percentof the before-tax cash flow discounted at 15 percent. Whatever method is used, the end result is a reasonably consistent
314
estimate of the price to be received under distr d esse conditi itions. Lenders may ignore income taxes in determin' Iendimg I"irmts unIess the company is fully taxablea mg d fi n a . Iarge portion 0 Itstotal cash flow is being dedicated t loanrepayments. Underthese circumstances, the tax ~ ligations of the company will be forecast, recogniz~ng that all interest payments would reduce the taxable i _ come. It is realistic to assume that only the cash flo: after taxes would be available to repay principal. The actualloan size is negotiatedbetweenthe borrower and the lender, using the foregoing calculations as a guide.
26.4.5 Auditing Evaluations For manyyears,the industryhas attempted to introduce auditing techniques as the basis for an independent evaluatorto offer an opinion ofthe reservesdetermined internally by a company. In the auditing of financial statements, procedures are established and thenchecked to see if they are being followed by reviewing random and material samples of the various financial transactions. If the samples show that the procedures are being followed, then the financial statements, after adjustments, are considered to reflect the finances of the company. The lack of the development of an equivalent system for reserves reportssuggests that the auditing procedures used for financial accounts may not be applicable to reserves reports. Procedurescan certainlybe established and checked,but because of the need for interpretation ofthe data in the estimationof reserves,the checking of a sample could lead to an erroneous conclusion concerningthe accuracyofthe reserves estimates. It would be very easy for a comparisonof a sampleto lead to the conclusion that the reserves should be half or double those determined by the companybecauseeach evaluator could interpret the available data differently. Of course, it would also be possible to agree after reviewing a sample, but disagree significantly after reviewing the total reserve base. Because of this problem of interpretation leading to different answers using the same data by different engineers who are both knowledgeable and objective, it is common for external evaluators, when asked to offeran objectiveopinionofthe reservesof a company, to independently evaluate the reserves which represent 70 to 80 percent of the total reserves (or value) of the company. This would normally require 20 to 30 percent of the properties in which the company has
USES OFRESERVES EVALUATIONS
interests to be evaluated. The ratio of external to internal reserves determined for the 70 to 80 percent of the reserves is then applied to the internal total to give the order ofmagnitude ofreserves that the auditor would expect to determine if he carried out a complete, independent evaluation. The anticipated difference should not be more than plus or minus 10 percent, with a desired difference of less than plus or minus 5 percent. These are the differences that can be expected as the result of interpretation by different evaluators. Differences in the total of greater than 10 percent would normally require review and explanation. No two evaluators will agree on the reserves for all of the properties of a company, but the differences will be higher for some properties and lower for others so that the sum of all the estimates by each evaluator will usually be within the 10 percent range suggested. Two evaluators will seldom estimate the same reserves for a given property. The best opinion that can be expected from an audit is that any differences would be within an acceptable range. Attempts are still being made for external evaluators to use samples to develop an opinion ofthe "correctness" of the reserves estimates by company staff. This does not appear to have a statistically sound basis for drawing such a conclusion due to the need for data interpretation. It is a reality that no one can accurately predict the reserves that will be recovered from a well until it has been abandoned. This lack ofprecision appears to limit the use of audits based on small samples to arrive at a soundly based opinion of the accuracy of reserves estimates.
26.4.6 Securities Reporting The agencies regulating public companies in both the US and Canada require annual reporting ofreserves and may require additional information on the future cash flows expected from those reserves. The US agency is the Securities and Exchange Commission (SEC), and the most important agency in Canada is the Ontario Securities Commission (OSC), which regulates the Toronto Stock Exchange on which a number of Canadian companies are listed.
the opening and closing net proved reserves volumes, together with an analysis of the change broken into the following: • Revisions • Discoveries and extensions • Purchases • Sales Production Net proved developed reserves are also to be disclosed. Form 10K also requires disclosure of the discounted future net cash flows, which are based on a forecast of production from the proved reserves with constant prices and constant costs and after deducting income taxes, using any remaining tax pools and all future investments required to produce the proved reserves. The discount rate used is 10 percent. A quarterly report is also submitted to the SEC on Form 10Q, but this report does not require reserves to be disclosed or the basis for the ceiling test, even though one is required in the preparation ofthe financial statements reported in Fonn I OQ. Recently, the SEC has agreed to accept the disclosure made to the Canadian equivalent ofthe SEC (generally, the OSC), instead ofFonn 10K, for companies listed in both Canada and the US. The Canadian disclosure does not include a discounted cash flow; however, reconciliation to US accounting principles, including the ceiling test, will be required at least to the end of 1993. In a prospectus issued by an oil and gas company listed on a US exchange, net remaining proved reserves at a date generally within 12 months ofthe date of the prospectus must be included. Any material changes in reserves since their determination must be disclosed. A discounted cash flow, using constant prices and costs, after taxes, and at a discount rate of 10 percent, must be included. The reserves reported in the prospectus must be supported by a detailed report on the determination of those reserves. The SEC does not require an independent reserve report; however, in practice, except for the major companies, underwriters usually insist on an independent report.
US Securities and Exchange Commission
Ontario Securities Commission
The SEC requires companies listed on a US stock exchange to submit annual reports that include a reconciliation of the changes in the net (after royalties) remaining proved reserves of the company. This information is reported on SEC Form 10K and includes
The OSC requires an annual report in which the gross (before royalties) remaining proved reserves must be included. It includes the opening and closing balances and a reconciliation that breaks the change in reserves
315
~I DETERMINATION OF OIL AND GASRESERVES
into the same five categories used for the report to the SEC. A prospectus for an oil and gas company listed on the OSC must include a statement ofgross remaining proved and probable reserves at a date generally within 12 months of the date of issue of the prospectus. Material changes in reserves, since their determination, must be disclosed. The OSC also requires a discounted cash flow, after taxes, using constant prices and costs and a discount rate of 10 percent. A cash flow, after taxes, using escalated prices and costs and higher discount rates may also be included, at the choice of the issuer. The reserves reported in the prospectus must be supported by a detailed report on the determination of those reserves. The report must be prepared by a registered professional engineer, or a registered professional geologist, who is independent of the issuer of the prospectus, but in-house reports may be acceptable from large, well-established companies at the discretion of the OSC.
26.4.7 Accounting Requirements Accountants use reserve estimations and subsequent evaluations in preparing financial reports and audits. The main uses of evaluations by accountants are for the purposes of ceiling tests and depletion calculations. Ceiling Tests
Two methods ofreporting the costs ofexploring for and developing oil and gas reserves are available to corporations: "full-cost" accounting and "successful efforts" accounting. In full-cost accounting, all costs associated with exploration and development are capitalized and written off over the life of all the reserves in each country. With successful-efforts accounting, only those investments in exploration and development that are successful in finding reserves are capitalized and written off over the life of the reserves on a property-by-property basis; all costs ofunsuccessful wells are expensed in the year in which they occur. Full-Cost Accounting
Full-cost accounting, presents a less conservative, but perhaps more accurate, picture of the financial results of a company. It is justified by the argument that all investments, both successful and unsuccessful, lead to the development of the company's total "inventory" of oil and gas reserves. The users of full-cost accounting could experience a run of bad luck that results in costs which are greater than
316
the value of the reserves added or are even greater th the total net income to be earned by producing the :: serves. To ensure that the balance sheet does t include highly overstated values for the company'sn~1 and gas reserves, a "ceiling test" must be performed quarterly for US registered companies and at least annually for Canadian registered companies. This test determines the maximum value that can be carried for the net cost of reserves under "Property, Plant and Equipment" on the balance sheet and, ifthe actual value carried is in excess of the ceiling test value, then the actual value must be written down to the ceiling test value. This is a one-way street, so that ifthe ceiling test value is greater, there is no change to the value carried. In Canada, a company using full-cost accounting is subject to a ceiling test with a value based on a cash flow forecast prepared from the forecasts ofproduction from the proved reserves, using constant prices and costs. The company's estimated future interest and administrative costs and income taxes are deducted. Income taxes take into account the remaining tax pools and any future investments. The resulting undiscounted cash flow is then compared with the value of the net cost of the reserves carried on the balance sheet (less deferred income taxes) and, ifthe cash flow is lower, then a writedown to the lower amount must be made. This ceiling test is a cost-recovery test, as it limits the value of the reserves on the balance sheet to no more than the net estimated cash flow to be recovered from the production of those reserves. The constant price used may be either the price at the end of the reporting period or the average price for the period. In the US, the ceiling test is a value comparison test for companies using full-cost accounting. The ceiling test value for a full-cost company is determined by taking the estimates of proved reserves and forecasts of production for the company and preparing a cash flow forecast using constant prices and costs. This cash flow, after taxes, is then discounted at 10 percent per annum and, ifthe value carried for the net costs of the reserves on the balance sheet (less deferred income taxes) is more than this present worth value, a write-down to this value must be made. It is a rough estimate of the value of the reserves, and this constant price and cost-based cash flow, discounted at 10 percent, must be reported in the annual report submitted to the SEC. The 10 percent discount rate used recognizes that inflation is not included in the cash flow forecast because constant prices and constant costs are
:""" .
':
.
USES OFRESERVES EVALUATIONS
used. The price used is that in effect on the last day of the reporting period.
26.4.8 Establishing Finding and Replacement Costs
The volatility of oil and gas prices over the last five years has resulted in significant write-downs, especially in the US where the price to be used must be the price on the last day of the reporting period and the ceiling test must be applied quarterly. A very low price at the end ofone period can cause an extreme write-down that can never be recovered even if prices bounce back in future quarters.
One of the important uses of estimates of reserves is in determining the average cost of replacing reserves. It must be determined whether the exploration and development expenditures are adding reserves for a cost that is less than or equal to their value. Ifthe cost ofreserves additions is in excess of their value, then the company will not be making a rate of return that is equal to or greater than the discount rate used to determine value.
Successful-Efforts Accounting
The finding cost that is often calculated as a measure of exploration success is determined by dividing the exploration costs by the proved reserves discovered. A calculation that only includes exploration costs gives a completely useless (and possibly misleading) result as discovered reserves are of no use until they are fully developed. The cost ofa fully developed unit of reserves ranges from 1.3 to 3.0 times the cost of finding the reserves. For this reason, a low finding cost could lead to the mistaken conclusion that the exploration program was successful even though the fully developed cost was in excess of the value of the reserves added. The opposite is also true, where high exploration costs might indicate an apparent lack of success, but low development costs, in fact, develop reserves with a cost that is less than their value.
Canadian and US companies using successful efforts accounting are subject to a ceiling test, but the constant price, constant cost cash flow used is not discounted or reduced by any interest, administrative, or income tax costs. In both Canada and the US, unproven properties are included in the ceiling test value. The value included is the cost, reduced by any impairment of value due to either exploration on the lands or to declines in market values. Depletion Calculations
Accountants also employ estimates of proved reserves as the most commonly used method of amortizing or writing off the cost of capital investments. This "depletion" calculation divides the remaining undepleted or undepreciated capital costs by the remaining reserves to determine a depletion rate. These net capital costs are depleted (that is, reduced) by the product of the production during the depletion period and the depletion rate. In this way, all the capital costs are written off by the time the last production is received. To calculate the depletion rate and charge, all the reserves and production must be converted to the same units, usually barrels of oil equivalent (BOE), which must be based on either heating values (6 x 103 cf per BOE) or relative selling prices. Having determined the rate per BOE, then the rate per barrel of oil and per thousand cubic feet of gas can be back-calculated. The depletion charge is calculated at the end of the period, using the remaining net investment, and before the current period charge for depletion and remaining proved reserves plus production for the period are calculated. Depletion is usually calculated on a quarterlybasis, and the annual charge is the sum of the quarters. The depletion calculation must be updated whenever there is a material change to the rate, as a result of either new investment or changes in reserves quantities.
Only replacement costs give a useful measure of success. They are calculated by dividing the total of all exploration and development costs (including all plants and production facilities) by the proved reserves added. The reserves other than oil would be converted to barrels of oil equivalent (BOE) to make this calculation. This conversion is discussed in Section 26.4.9. Replacement costs are calculated on before-tax costs, but values are based on after-tax calculations. Using a mix of tax write-offs for an average program of exploration and development, the after-tax cost ofreplacement would be in the order of 65 percent of the before-tax cost for a company that can use all the tax write-offs. The current average going concern values of oil and gas reserves in Alberta are approximately $7.00 per barrel and $0.35 per thousand cubic feet. This means that the before-tax replacement cost ofreserves must be less than $10.00 per barrel and $0.55 per thousand cubic feet if a rate of return of 12 percent is to be earned, after taxes. In recent years, only 60 percent of the companies developing reserves in Canada have been able. to add reserves at costs that are this low. Proved reserves may be added before all the costs of development have been incurred or may be added after 317
7
DETERMINATION OF OIL AND GASRESERVES
the investments are made, such as with enhanced recovery schemes. Because costs and reserves additions may be out of synchronization, it is preferable to calculate replacement costs on a three- to five-year rolling average.
energy as gas from the well.head to the burner tip as it costs to move the same unit of energy as oil. Conse_ quently, the value at the wellhead (or in the ground) is in no way related to energy content.
Studies have shown that annual replacement costs for a company can range from $4.00 per barrel one year to $280.00 per barrel the next year, while on a five-year rolling average, the costs are $10.00 per barrel, which means that the company is earning an acceptable rate of return. If replacement costs are higher than values for an extended period, it is an indication that serious financial problems will probably develop for the company some time in the future.
6 x 103 cf per BOE, but it actually ranges from about
The average values set out in the foregoing discussion may not be applicable to reserves added by particular companies; consequently, the actual value ofthe reserves added by a company should be determined in detail for comparison with the replacement cost.
26.4.9 Estimating Barrels of Oil Equivalent From time to time, it becomes necessary to divide costs among reserves of oil, gas and related products, or to report reserves ofoil, gas and related products asa common unit. This is done by converting reserves that are not oil to barrels of oil equivalent (BOE). Often the reason for the conversion is not understood and, therefore, the conversion is made incorrectly. Conversions to BOE are. usually made for one of the following reasons:
The energy equivalent often (but incorrectly) used is 5.8 x 103 cf per BOE for light oil to approximately 6.3 x 103 cfper ~OE for hea~ oil. The.use of energy content as the basis for conversion can give misleading results which, in turn, could lead to wrong decisions about success and value. It is recommended that this BOE never be used for the purpose of value conversion. Because most conversions to BOE are for the purpose of value comparison, the values of oil and gas reserves should be used to calculate the conversion ratio. Currently, "average" values for fully developed reserves of oil and gas in Alberta are in the order of $7.00 per barrel and $0.35 per thousand cubic feet. This gives a conversion ratio of 20 x 103 cf per BOE. The equivalent barrel of oil is a barrel oflight oil. The value-based conversion rates have varied as shown in Table 26.4-1 over the last eight years for "average" reserves. No company owns average reserves, so that the conversion rate for a particular company could be higher or lower, depending on the actual reserves owned.
Table 26.4-1 Conversion Rates Year
(103 cf/stb)
1. To calculate the depletion charges used to write off the investment in reserves 2. To report reserves volumes using a common unit 3. To calculate the replacement cost per unit ofreserves added 4. To calculate the acquisition cost per unit ofreserves purchased The conversion for the first reason must use either heating value (6 x 103 cf per BOE) or relative selling prices, as established by accounting principles. Each of the other three reasons is for the purpose of making a value comparison and, therefore, conversions based on the relative values of the reserves to be converted compared to the value of oil reserves should be used, if possible. Energy equivalence should not be used for the purpose of value conversion, as energy equivalence is only of significance at the burner tip. It costs approximately five times as much to move a unit of
318
BOE
~
1984 1985 1986 1987 1988 1989 1990 1991
20 17 17 18 16 14 16 20
Price is not the most desirable basis for determining conversion to BOE, as the cost of producing energy in the form of gas is usually different per dollar of income than the cost of producing oil. Therefore, value, which is the difference between price and cost, will usually yield conversion rates that are different from those based solely on price. If values are available, they are to be preferred; however, values are not generally available for related plant products and, therefore, price provides the only useful way to convert reserves of ethane, propane, butanes, pentanes plus and sulphur to BOE.
USES OF RESERVES EVALUAnONS
Example
Conversion Factors
Reserves ofcrude oil, gas and related products as shown in the following table were purchased several years ago for $197.8 million.
Product
Wellhead Price
Crude oil Ethane Propane Butanes Pentanes plus Sulphur
$19.70/bbl 7.00/bbl 7.50/bbl 11.50/bbl 19.50/bbl 75.00/1t
Reserves Purchases Company Interest Gross Reserves (before deducting royalties)
Product Crude oil Naturalgas Ethane Propane Butanes Pentanesplus Sulphur
8,200,000 165,400 740,000 610,000 370,000 670,000 670,000
X
bbl 10' cf bbl bbl bbl bbl It
What did this price represent in the way of value per unit of reserves? At the time, the conversion of gas to BOE on a value-equivalence basis was 14 x 103 cf of gas for each barrel of oil. No value data was available for the related gas products, so the wellhead price had to be used to determine the conversion ratios. At the time, the prices shown in the table ofconversion factors were in effect. Using these conversion factors, with the conversion factor of 14 x 103 cfper BOE for gas and the reserves set out earlier, the total reserves in barrels of oil
Conversion to BOE Value-Equivalence (Crude Oil Price IProduct Price) 2.81 2.63 1.71 1.01 0.26
equivalent were calculated. These are shown in the following table. By using the BOE at 23,966,233 barrels and the purchase price at $197.8 million, the value paid per BOE was $8.25 per BOE. The purchase price can be allocated to each product to determine the respective value per unit of reserves. The table on the next page shows the value per unit of reserves. The assignment of value is extremely sensitive to the conversion ofother products to BOE. Ifthe price equivalence for gas had been used instead of the value equivalence to convert to BOE, the conversion factor would have been 12 x 103 cf per BOE and the price paid for the crude oil reserves would be reduced from $8.25 per barrel to $7.65 per barrel, but the price paid for the gas reserves would be increased from $0.59 to $0.64 per thousand cubic feet.
Total Reserves in Barrels of Oil Equivalent Product
(1) Company Interest
(2) Conversion
toBOE
Reserves
Crude oil Natural gas Ethane Propane Butanes Pentanes plus Sulphur Total
8,200,000 bbl 165,400,000 x 10' cf 740,000 bbl 610,000 bbl 370,000 bbl 670,000 bbl 670,000 It
1.00 14.00 x 10' 2.81 2.63 1.71 1.01 0.26
ef/BOE bbl/BOE bbl/BOE
bbllBOE bbl/BOE
It/BOE
Barrels of Oil Equivalent (BOE) Col (1) + Col (2) 8,200,000 11,814,286 263,345 231,939 216,374 663,366 2,576,923 23,966,233
319
iM
il'1 DETERMINATION OF OIL ANDGAS RESERVES
Value per Unit of Reserves
Product
ROE
Purchase Price of Reserves
Crude oil Natural gas Ethane Propane Butanes Pentanes plus Sulphur
8,200,000 11,814,286 263,345 231,939 216,374 663,366 2,576,923
$67,676,885 97,506,595 2,173,460 1,914,257 1,785,795 5,474,944 21,268,064
Total
23,966,233
$197,800,000
Although it has no relevance except at the burner tip, the energy equivalence is used from time to time to convert reserves of gas to BOE. The equivalence used is usually 6 x 10' cfper BOE. Ifthis had been used in this case, the price paid for the oil reserves would be reduced from $8.25 per barrel to $5.00 per barrel, but the price paid for the gas reserves would be increased from $0.59 to $0.83 per thousand cubic feet. This is quite an unrealistic assignment of the price paid to the different reserves. If the values for related products were included with the value of the natural gas, then the value of the gas and related products at $130,123, 115 ($197,800,000 less the value of oil at $67,676,885) divided by the natural gas reserves of 165,400,000 x I OJ cfwould yield a unit value of$0.79 per thousand cubic feet. These values are market values, not going concern values, as the value ofthe COGPE created by the direct purchase was included in the price paid.
It is easy to understand why reserves quoted in BOE that are calculated using a conversion of 6 x 10J cf per BOE would vastly overstate the apparent reserves of the company, when the conversion should have been based on a conversion rate of 20 x 10J cf per BOE. Reserves are usually used as a proxy for value and the use of a 6 to I conversion ratio, instead of a 20 to I conversion ratio, will yield gas reserves in BOE that are more than three times what they should be. Conversions to BOE are really being made to barrels of light oil equivalent. Because reserves of medium and
320
Company Interest Reserves
Value Per Unit
8,200,000 bbl 165,499,000 x 10' cf 740,000 bbl 610,000 bbl 370,000 bbl 670,000 bbl 670,000 It
8.25/bbl 0.59/10' cf 2.94/bbl 3.14/bbl 4.83/bbl 8.17/bbl 31.74/lt
heavy oil have values which may be only 25 percent (or less) of the value of a barrel of light oil, it is perhaps time for the industry to break out oil reserves into light, medium and heavy categories, or convert medium and heavy oil reserves to "barrels oflight oil equivalent," as a better indication of the real value of the inventory of oil and gas reserves.
26.4.10 Estimating Net-Back Calculations A calculation used from time to time takes the price received for a unit ofproduction at a particular time and deducts from it all production costs, royalties, and taxes to give the net-back to the producer from each barrel or thousand cubic feet. This net-back calculation is a dangerous way to identify value or to determine the effect of changes in price, costs, royalties or taxes. A net-back calculation only looks at the situation at one time and ignores anything that may occur at some other time. It is also difficult, if not impossible, to make realistic tax calculations because the write-off of tax pools is not a function of a unit of production. Net-back calculations ignore changes in the future in production rates, royalties and operating costs and also ignore any capital costs. A net-back calculation merely looks at the revenue per unit of production at a particular time and some of the costs at that time. For this reason, the results can be manipulated to give the answer being sought. It is a poor measure of what is likely to happen over a period of time and, consequently, is not recommended as a useful basis for making investment decisions.
BIOGRAPHIES OF AUTHORS
Complete names of technical societies are given in the Acronyms.
Dr. Roberto Aguilera - B.Sc., Petroleum Engineering, Universidad de America, Bogota, Colombia; M.Eng. and Ph.D., Colorado School of Mines, Golden, CO - Currently, President, Servipetrol Ltd. - Authorof Naturally FracturedReservoirs and co-author of The Technology ofArtificialLifl Methods and Horizontal Wells, and papers in technical journals
- International lectureron naturallyfractured reservoirs - Memberof AAPG, ACIPET, APEGGA, CWLS, Petroleum Society of CIM, SPE, SPWLA; past Directorof Petroleum Society, Calgary Section
Dr. Soheil Asgarpour - B.Sc., Mechanical Engineering, Tehran University, Tehran, Iran; Ph.D., Rice University, Houston, TX - Currently, Technical Leader of Northern Alberta and BritishColumbia Group, Gulf Canada Resources Limited - Authorof over 35 technicalpapers on heat transfer, solar energy, reservoirengineering, enhanced oil recovery, production engineering, and risk analysis - Memberof APEGGA, Petroleum Society of CIM, PNA, SPE; National Directorof Petroleum Societyof CIM since 1990; Chairman of Editorial Board of JCPT 1991-1993
Barry R. Ashton - B.Sc., Chemical Engineering, University of Alberta, Edmonton, AB - Currently, SeniorPartner, Ashton Jenkins and Associates Ltd., which specializes in reservoir studies and reserve evaluations - Experience includes Chairman of Reserve Evaluation Committee of Encor Inc. as part of the $1 billion property rationalization agreement reached between Encor, Amoco Canada, and Maligne Resources in 1992 - Member of APEGGA, SPE, Petroleum Society ofCIM Dr. Anthony D. Au - Ph.D., ComputerScience, University of Utah, Salt Lake City, UT - Currently, Vice-President, Servipetrol Ltd. _ Specialties include fractured reservoirs and reservoir simulation - Co-author of over 20 technicalpapers on reservoir simulation technology - Memberof APEGGA, Petroleum Societyof CIM, SPE; Directorof Petroleum Societyof CIM C. BrentAustin - B.E., Mining Engineering, Technical University of Nova Scotia, Halifax, NS - Currently, Advisor, Petrophysics, PanCanadian Petroleum Limited _ Experience includes international consulting in areasofformation evaluation and economic analysis - Memberof APEGGA, CanadianInstitute of Management, Petroleum Society of CIM, SPWLA
l 321
DETERMINAnON OFOILANDGASRESERVES
N. Guy Berndtsson · B.Sc., Chemical Engineering, University of Alberta, Edmonton, AB • Currently, Board Member, Energy Resources Conservation Board - Experience includes formation evaluation and oil and gas reserves determination · Co-author and presenter of technical papers on crude oil reserves, enhanced recovery techniques and economics, water and waste disposal wells, and well servicing • Member of APEGGA, Petroleum Society of CIM, Technical Advisory Committee for the Petroleum Recovery Institute Robin G. Bertram - B.Sc., Petroleum Engineering, University of Alberta, Edmonton, AB · Currently, Operations Engineer, Talisman Energy Inc. • Experience includes formation evaluation, oil and gas reserves determination, and project management - Member of APEGGA, Petroleum Society of CIM, SPE Janusz Bielecki • Masters Degrees in Energy Economics, University of Calgary, Calgary, AB, and International Law, University of Warsaw, Poland · Currently, evaluator of supply costs and the impact of new technologies (i.e., horizontal drilling) on petroleum reserves and supply, National Energy Board Experience in economic evaluation, market analysis, price and supply forecasting, energy regulation and strategic planning - Member of International Association of Energy Economists
'''-1
I
Keith M. Braaten · B.Sc., Mechanical Engineering (Distinction), University of Saskatchewan, Saskatoon, SK · Currently, Technical Consultant and Partner Coles Gilbert Associates Ltd. ' • Specialties include reservoir engineering and economic studies of secondary and tertiary enhanced oil recovery schemes, fractured reservoirs, gas cycling schemes, heavy oil reservoirs, and shallow gas reserves • Member of APEGGA, Petroleum Society of CIM, SPE Keith D. Brown · B. Chern. Eng., Technical University of Nova Scotia, Halifax, NS • Currently, Manager, Oil and Gas Evaluations, Royal Bank of Canada, directing the technical evaluation of the bank's energy portfolio • Experience includes gas processing, construction, and economic evaluations · Member of APEGGA Mike J. Brusset • B.Sc., Geological Engineering, University of Oklahoma, Norman, OK • Currently, President, Brusset Consultants Ltd. • Specialties include formation evaluation, oil and gas reservoir performance studies, and economic evaluations of oil and gas reserves · Lecturer for 13 years on economic evaluations in petroleum exploration and engineering at U ofC • Member of APEGGA, Petroleum Society of CIM
322
s
- B.Sc., Geology, University of Alberta, Edmonton, AB
- B.Sc., Mineral Engineering (Petroleum) and MBA, University of Alberta, Edmonton, AB
- Currently, Executive Vice-President, Sproule Associates Limited
- Currently, Director, Natural Gas Business Centre, Shell Canada Limited, responsible for natural gas marketing and business development activity
- Specialties include geological interpretation of oil and gas reservoirs, reserves estimates, and analysis of exploration prospects based on assessmant of geological risk and economic evaluation - Co-author of papers and course notes on subsurface mapping of oil and gas reservoirs, potential oil and gas resources, economic evaluation and risk analysis of Canadian oil and gas properties - Member of APEGGA, CSPG, Petroleum Society ofCIM Graham R. Campbell - B.Sc., Physics, University of Waterloo, Waterloo, ON; M.Sc., University of British Columbia, Vancouver, BC - Currently, Director-General, Energy Resources, National Energy Board, responsible for assessment of Canada's oil and gas supply, monitoring of upstream industry activity and new resource extraction technologies, and regulation of geological and geophysical activities in the north - Member of AAPG, APEGGA Noel A. Cleland - B.E. (Honours), Mining Engineering, University of Sydney, Australia - Currently, Senior Engineering Consultant and Director, Sproule Associates Limited
- Experience includes technical and managerial positions in petroleum engineering, as well as assignments in economics and corporate strategies - Member of APEGGA, Petroleum Society of CIM, SPE G.J. (Gerry) DeSorcy - B.Sc., Petroleum Engineering, University of Alberta, Edmonton, AB - Currently, Energy Consultant - Formerly, first Chairman of Alberta Natural Resources Conservation Board, and previously, Manager of Gas Department, Board Member, and Chairman, Alberta Energy Resources Conservation Board - Member of APEGGA, Petroleum Society of CIM John Drury - B.A., Honours Science, University of Toronto, Toronto, ON - Currently, independent consultant and technical consultant to the Ontario Securities Commission - Experience includes mining geology and the regulatory industry - Member of APEO, CSEG, GAC; Life Member of CIM Dr. David C. Elliott
- Author of papers on petroleum economics and the Canadian petroleum industry and lecturer on petroleum economics at U. of C.
- B.Sc. and Ph.D., Geology, University of Birmingham, U.K.; B.Math., University of Waterloo, Waterloo, ON
- Member of APEGGA, Austro-Asian Institute of Mining and Metallurgy, Petroleum Society of CIM, SPEE; Past President of APEGGA
- Currently, Consultant, Geosgil Consulting
- Recipient of Frank Spragins Award from APEGGA in 1990, and Petroleum Society's Distinguished Service Award in 1984; CIM Distinguished Lecturer in 1974
- Experience includes oil and gas field development in Canada and overseas; statistical, geostatistical and mathematical applications - Member of AAPG, APEGGA, CSPG, Geological Society (UK), IAMG, Petroleum Society of CIM
323
Robert V. Etcheverry - B.Sc., Mechanical Engineering, University of Saskatchewan, Saskatoon, SK - Currently, General Manager, Production and Engineering, CN Exploration Inc. - Experience includes technical, research, supervisory, and managerial positions in the oil and gas industry in Canada and the US - Member of Petroleum Society ofCIM, PNA, SPE; Director, Calgary Section, Petroleum Society of CIM Merlin B. Field - B.Sc., Petroleum Engineering, Oklahoma University, Norman, OK - Currently, Consulting Reservoir Engineer - Specialties include petroleum property optimization, project evaluations, numerical simulation, computer applications, and enhanced recovery - Member of APEGGA, SPE RonM. Fish
Mam Chand Gupta B.Sc. and M:Sc., Physical Chemistry, University of Agra, India; D.LC., all Technology and M.Sc., Petroleum Reservoir Engineering, Imperial College, University of London, London UK ' - Currently, President and Consultant, GM International Oil and Gas Consulting Corporation - Experience includes supervisory, advisory and technical positions in reserves determination formation evaluation, oil and gas well testing, oil and gas field development, enhanced oil and gas recovery studies, economic evaluation, oil and gas property evaluation Instructor of drilling and production engineering courses at SAlT and U. of C. Co-author of one technical paper - Member of APEGGA, Petroleum Society of CIM Dave Hemphill - B.Sc., Mining Engineering, University of Alberta, Edmonton, AB
- B.Sc., Geological Engineering, University of Manitoba, Winnipeg, MB
- Currently, Petroleum Engineering, Shell Canada Limited
- Currently, Reservoir Engineering Advisor, New Pool Development Group, Imperial Oil Limited, Resources Division - Experience includes reservoir engineering, production engineering, development, operations, research and oil sands
- Experience includes uranium mining industry, development geology, oil and gas reserves definition, and field development optirriization
- Member of APEGGA J.D. (Joe) Giegerich - B.Sc., Mining Engineering, University of British Columbia, Vancouver, BC - Currently, retired from Chevron Canada Resources - Experience includes drilling, production, and reservoir engineering, and estimating hydrocarbon reserves - Member of APEGGA, Petroleum Society of CIM; past Chairman, Edmonton Section and two-term Director, Petroleum Society of CIM
324
- Member of APEGGA, CSPG, SPE John M. Hewitt - M.A., Mechanical Science, Queens' College, Cambridge University, UK Currently, Consulting Partner, Martin Petroleum & Associates Experience includes oil and gas property and company evaluations, reserves determination, reservoir engineering and enhanced recovery studies, royalty and regulatory problems, government presentations, and acquisition analysis - Member of APEGGA
BIOGRAPHIES OFAUTHORS
William E. (Bill) Kerr - B.Sc. (Honors), Petroleum Engineering, University of Wyoming, Laramie, WY Currently, Operations Manager, Joss Energy Experience includes production, operations, drilling, reservoir, evaluations, and enhanced recovery - Member of APEGGA, Petroleum Society of CIM, SPE Harold R. Keushnig - B.Sc., Chemical Engineering, University of Alberta, Edmonton, AB - Currently, Manager, Gas Department, Alberta Energy Resources Conservation Board, responsible for well spacing, reservoir recovery and recovery estimation, conservation, processing, and removal of gas from the province Member of APEGGA, CGPA, Canadian Potential Gas Committee, Petroleum Society of CIM Gobi Kular - B.Sc., Petroleum Engineering, University of Montana, Missoula, MT; course work toward M.Sc., University of Calgary, Calgary, AB - Currently, President, Advanced Petroleum Technologies - Experience includes international consulting on major reservoir studies in Canada, North Africa and the Middle East - Author and presenter of several technical papers on pressure transient well-test design and analysis, fracture design, project monitoring, and enhanced oil recovery - Member of APEGGA, Petroleum Society of CIM, SPE
Craig F. Lamb - B.Sc. and M.Sc., Geology, University of Manitoba, Winnipeg, MB; M.B.A., University of Calgary, Calgary, AB - Currently, President of Lonach Consulting Ltd. - Specialities include core analysis, fractured reservoir studies, design of technical training programs, and geoscience management - Author of nine publications - Member of AAPG, APEGGA, CSPG, Petroleum Society of CIM, SCA R.V. (Bob) Lang - B.Sc., Chemical Engineering, University of Alberta, Edmonton, AB - Currently, independent petroleum consultant - Specialties include reserves determination and reserves reporting - Member of APEGGA, Petroleum Society of CIM, SPE - Recipient of two merit awards from CPA, one service award from AGA Dana B. Laustsen - B.Sc. (Distinction), Mechanical Engineering, University of Calgary, Calgary, AB - Currently, Consultant and Director, Coles Gilbert Associates Ltd. - Specialties include economic analyses and reservoir engineering studies of waterflood optimization, miscible flood performance, gas storage, and tight gas deliverability models - Member of APEGGA, Petroleum Society of CIM William V. Mandolidis - B. Applied Science and Chemical Engineering, University of Toronto, Toronto, ON - Currently, Coordinator, Corporate Planning and Business Development, Saskatchewan Oil and Gas Corp. - Experience includes surface and subsurface petroleum engineering, reserves evaluation and reporting, economic and business analysis, and corporate strategy development - Member of APEGGA, Petroleum Society of CIM
325
Michael E. McCormack - B.Sc., Chemical Engineering, University of Calgary, Calgary, AB - Currently, Founder and Consultant, Fractical Solutions Inc. - Specialties include the application of fractal mathematics and computer solutions to petroleum engineering problems, fluid dynamics, heat transfer, well design, and surface facilities optimization - Author and presenter of several papers on well engineering Member of APEGGA, Petroleum Society of CIM, SPE
David C. Poon - M.Sc., Chemical Engineering, University of Calgary, Calgary, AB · Currently, Engineering Consultant, D.C. Poon Consulting Inc. · Specialities include geostatistics, reservoir simulation, well testing, and horizontal wells • International lecturer on enhanced oil recovery, thermal well testing, and water management - Member of APEGGA, CHOA, CSChE, Petroleum Society of CIM, SPE - Recipient of best paper award in 1990 from Petroleum Society of CIM Dr. Ross A. Purvis
Raymond A. Mireault - B.Sc., Agricultural Engineering, University of Manitoba, Winnipeg, MB · Currently, Senior Engineer, Southern Business Unit, Reserves Additions Team, Gulf Canada Resources Limited • Experience includes conventional oil and gas exploration, development and production, coal bed methane, horizontal drilling, and super sour gas · Member of APEGGA, CSEG, CSPG, Petroleum Society of CIM Margaret Nielsen - B.sc., Mechanical Engineering, University of Alberta, Edmonton, AB · Currently, Business Planning and Performance, Petro-Canada · Experience includes evaluating and developing gas, oil and heavy oil reserves under primary recovery, waterfloods, cycling schemes, miscible floods, and fire floods.
· B.Sc., University of Oklahoma, Norman, OK; M.Sc., University of Wyoming, Laramie, WY; Ph.D., University of Alberta, Edmonton, AB - Currently, Manager, Oil Department, Energy Resources Conservation Board - Experience includes production engineering and the development of technical software for phase behavior, petrophysics, decline curve, and other reservoir engineering applications, and teaching - Member of APEGGA, Petroleum Society of CIM, SPE Tim J. Reimer · B.Sc., Chemical Engineering, University of Calgary, Calgary, AB - Currently, Manager, Gas Contracts and Supply, Pan Alberta Gas Ltd. · Experience includes gas plant design and operations, joint interest, economic evaluations, exploitation and development. - Member of APEGGA, CGPA, Petroleum Society ofCIM,PNA
· Member of APEGGA W.D. (Bill) Robertson _ B.Comm., University of Alberta, Edmonton, AB _ Currently, Co-Chairman and Partner, Oil and Gas Industry Specialist Group, Price Waterhouse · Specialty is oil and gas accounting - Member of Petroleum Society of CIM; past Director and Officer, Petroleum Accountants' Society; Council Member, Institute of Chartered Accountants of Alberta
326
BIOGRAPHIES OFAUTHORS
J. Glenn Robinson - B.Sc. (Honours), Civil Engineering, Queen's University, Kingston, ON - Currently, President, Sproule Associates Limited - Experience includes exploitation engineering, production geology, petrophysical engineering, reservoir engineering, mathematical reservoir simulation, economic evaluations, and risk analysis - Lecturer on evaluation and risk analysis of Canadian oil and gas properties - Member of APEGGA, CWLS, Petroleum Society of CIM, SPE, SPEE, SPWLA
Dr. Ashok K. Singhal - Ph.D., Petroleum Engineering, University of California, Berkley, CA - Currently, Group Leader, Gas Flooding, Petroleum Recovery Institute, Calgary, AB - Experience includes horizontal well applications, enhanced oil recovery, and reservoir engineering - Member of APEGGA, Petroleum Society of CIM, SPE David W. Turt - B. App. Sc., Civil Engineering, University of Toronto, Toronto, ON .
- M.Sc., Statistics and M.Eng., Chemical Engineering, University of Calgary, Calgary, AB
- Currently, Vice-President, Engineering, Oil and Gas Department, Bank of Montreal, responsible for assessment of large domestic and international oil and gas project loan packages
- Currently, Manager, Strategic Action Planning, Petro-Canada
- Member of APEGGA, Petroleum Society of CIM, SPE, SPEE
Darlene A. Sheldon
- Experience includes reservoir engineering, oil and gas development and evaluations, information systems, economics, and asset rationalization - Author ofthree technical papers Member of AWES, Petroleum Society of CIM, SPE Dr. Phillip M. Sigmund - B.A.Sc., Chemical Engineering, University of Waterloo, Waterloo, ON; Ph.D., Chemical Engineering, University of Texas, Austin, TX - Currently, designing specialty oil extraction processes and building associated research equipment - Experience includes contract research and consulting projects related to improved reservoir management and fluid extraction processes - Author of several technical papers - Member of AIChE, APEGGA, Petroleum Society of CIM
George A. Warne - B.Sc., Electrical Engineering, University of Alberta, Edmonton, AB - Currently, Energy Resource Consultant and Secretary-Treasurer of the Canadian Association of the World Petroleum Congresses - Experience includes energy resource regulation and management, and reservoir engineering - Lecturer on energy resource regulation and management - Member of APEGGA, Petroleum Society of CIM, SPE Andy Warren - B.Sc., Civil Engineering, Queen's University, Kingston, ON Currently, Assistant Manager of Engineering, Oil Department, Alberta Energy Resources Conservation Board - Experience includes tight gas, shallow gas, small gas pools, oil pool reserves, well testing, and pool depletion strategies - Member of APEGGA, Petroleum Society of CIM,SPE
327 d
ACRONYMS
Acronyms for technical terms used in the monograph are listed first, followed by the acronyms for technical journals and societies, then for universities and colleges, and last, for countries, provinces and states. AACRT
adjusted attributed Canadian royalties and taxes
EARP
Environmental Assessment and Review Process
ACRI
attributed Canadian royalty income
EMV
expected monetary value
EOR
enhanced oil recovery
FCA
Federal Competition Act (US)
ADP AMP
average daily production .Alberta market price
APMC
Alberta Petroleum Marketing Commission
FDC
compensated formation density
AOF
absolute open flow
FERC
AOS
Alberta oil sands
Federal Energy Regulatory Commission (US)
AARTC
adjusted Alberta royalty tax credit
FOB
freight on board
BHT
FVF
formation volume factor
BOE
bottom-hole temperature barrels of oil equivalent
G&A
BTR
break-through ratio
GCA
general and administrative (costs) gas cost allowance
CAL
caliper
GCV
going concern value
CCA
capital cost allowance
GOR
gas-oil ratio
CDE
Canadian development expense
GORR
gross overriding royalty
CEC
cation exchange capacity
GPSA
Gas Processors Suppliers Association
CEDOE
Canadian exploration and development overhead expense
GR
gamma ray (API)
GSC
Geological Survey of Canada
CEE
Canadian exploration expense
GSP
CDF
cumulative distribution function
HClP
gas select price hydrocarbon in place
CGR
condensate-gas ratio
HPV
hydrocarbon pore volume
CGL CNL
conglomerate compensated neutron log
1FT lLd
interfacial tension
COGPE
Canadian oil and gas property expense
lLm
CPI
capital productivity index
IMPES
CPM
Critical path method
Implicit Pressure Explicit Saturation Method
CPUC
California Public Utilities Commission
IPL
Interprovincial Pipe Line
CSU DD&A
cyber service unit depreciation, depletion and amortization
IRR K.B
internal rate of return kelly bushing
DNPBI
discounted net profit before investment
LDC
local distribution company
DOE
Department of Energy (US)
LGR
liquid-gas ratio
DPHI
LPG
liquefied petroleum gas
DPR
density porosity discounted profit-to-investment ratio
MCM
multiple-contact miscibility
DROI
discounted return on investment
MMP
minimum miscibility pressure
DST
drillstem test
MV
market value
deep induction resistivity medium induction resistivity
329
t
,.1 DETERMINATION OFOILANDGASRESERVES
NAFTA
North American Free Trade Agreement
RIP
NEP
National Energy Program
NGL
natural gas liquids
SACROC a carbonate unit on the Kelly-Snyder field in Texas
NGPA
Natural Gas Policy Act (US)
SEC
NML
nuclear magnetic log
Securities and Exchange Commission (US)
NOPR
Notice of Proposed Rulemaking
SFL
NORR
net overriding royalty
SG
spherically focused laterolog specific gravity
NOVA
an Alberta corporation
SNP
sidewall neutron porosity
NPHI
neutron porosity
SP
spontaneous potential
NPV
net present value
TMPL
Trans Mountain Pipeline
NYMEX
New York Mercantile Exchange
TOP
take-or-pay (gas)
OECD
Organization for Economic Co-operation and Development
TVD
true vertical depth
TWP
township
reserves-to-production ratio
OGIP
original gas in place
USBM
United States Bureau of Mines
OOIP
original oil in place
VGR
viscous gravity ratio
OPEC
Organization of Petroleum Exporting Countries
WACC
weighted average cost of capital
WAG
water alternating gas injection
OSLO
a consortium of companies
WAR
water-air ratio
OSC
Ontario Securities Commission
WFT
wire line formation test
PAD
Petroleum Administration for Defense (US)
WGR
water-gas ratio
WOR
water-oil ratio
PADD
Petroleum Administration for Defense District (US)
WTI
West Texas Intermediate
PERT PDF
Program Evaluation and Review Technique probability distribution function
PDVSA
Petroleos de Venezuela S.A.
PG&E
Pacific Gas and Electric
PGRT
Technical Journals, Societies, and Institutions AAPG
American Association of Petroleum Gologists
ACIPET
Association of Colombian Petroleum Engineers
petroleum and gas revenue tax
AGA
American Gas Association
PGT
Pacific Gas Transmission
AIChE
PII
profitability index
American Institute of Chemical Engineering
PV
pore volume
AlME
PVT
pressure-volume-temperature
American Institute of Mechanical Engineering
PWR
present worth ratio
P-X
pressure composition diagram
RBA
rising bubble apparatus
RF RFT
recovery factor
RGE
range
ROI
return on investment
CHOA
RP
resource profits
CIM
330
repeat formation tester
APEGGA Association of Professional Engineers, Geologists and Geophysicists of Alberta APEO
Association of Professional Engineers of Ontario
API
American Petroleum Institute
AWES
Association of Women in Engineering and Science Canadian Heavy Oil Association Canadian Institute of Mining, Metallurgy and Petroleum
ACRONYMS CSChE
Canadian Society of Chemical Engineering
Universities and Colleges U of A
University of Alberta
CGPA
Canadian Gas Processors Association
U of C
University of Calgary
CSPG
Canadian Society of Petroleum Geologists
SAlT
Southern Alberta Institute of Technology
CSEG
Canadian Society of Exploration Geophysicists
CWLS
Canadian Well Logging Society
GAC
Geological Association of Canada
GPSA
Gas Processors Suppliers Association
IAMG
International Association for Mathematical Geology
IHRDC
International Human Resource Development Corporation
JCPT
Journal of Canadian Petroleum Technology
Countries, Canadian Provinces, American States AB
Alberta
AZ
CA
Arizona California
CO
Colorado
MB MT
Manitoba Montana
SK
Saskatchewan
NS
Nova Scotia
OK
Oklahoma
O&GJ
Journal of Petroleum Technology Oil and Gas Journal
ON
Ontario
OGCI
Oil and Gas Consultants International
TX
Texas
PNA SCA SPE SPEE
Petroleum Joint Venture Association
UK
United Kingdom
Society of Core Analysts Society of Petroleum Engineers
US, USA United States of America WY Wyoming
JPT
Society of Petroleum Evaluation Engineers
SPEJ
Society of Petroleum Engineers Journal
SPWLA
Society of Professional Well Log Analysts
331 t
GLOSSARY
Acidizing. A method of well stimulation using acid (to increase productivity); conducted mostly in carbonates. Acoustic log. A measurement ofthe interval transittime of compressional seismic waves in rocks near the wellbore of a liquid-filled borehole; used chiefly for estimating porosity and lithology; also referred to as sonic log. Analogous fields. Fields having similar properties that are at a more advanced stage of development or production history than the field of specific interest, and that may provide concepts or patterns to assist in the interpretation of more limited data. Anhydrite. A granular, white or light-colored evaporite mineral (CaSO.), often found together with rock salt. Annulus. The space around the tubing in a wellbore, the outer wall of which may be the wall of either the borehole or the casing. Aquifer. A stratum or zone below the surface of the earth capable of producing water. Arithmetic mean. The average obtained by dividing the sum of a distribution by the number of its addends. Asphaltene. Any ofthe dark solid constituents ofcrude oils and other bitumens that are soluble in carbon disulphide but insoluble in paraffin naphthas. Beta model. A numerical simulator used to model black oil systems; also referred to as black oil model. Bias. A systematic deviation from the actual value or distribution; a combination of two effects: displacement bias and variability bias. Bitumen. Refer to Crude bitumen. Black oil. Refers to a system in which the volume of fluid is primarily a function of reservoir pressure and constant temperature. A system that is not a black oil system includes compositional variables. Black oil model. Refer to Beta model. Bomb. A thick-walled container, usually steel, used to hold samples of oil or gas under pressure.
Bottom-hole pressure. The pressure in a well at a point opposite the producing formation as recorded by a bottom-hole pressure recorder. Bottom-hole temperature. The temperature in a well at a point opposite the producing formation. Bottom water. Sand layers at the bottom of a forma.tion which contain mobile water that appreciably affects reservoir performance; water in strata underlying an oil- or gas-bearing formation. Bourdon tube. A mechanical pressure-measuring instrument employing as its sensing element a curved or twisted metal tube, flattened in cross section and closed. Bubble point. In a solution oftwo or more components, the pressure at which the first bubbles of gas appear; same as saturation pressure. Bulk deusity. Density of the combined pore volume and rock volume; measured, for example, by a density log. Bulk volume. Total volume of a formation including the pore volume and the rock volume. Butanes. In addition to its normal scientific meaning of C.H IO (a mixture of two gaseous paraffins, normal butane and isobutane), a mixture mainly ofbutanes that ordinarily may contain some propane or pentanes. Capillarity. The effect of surface attraction forces among oil, gas, water, and rock in retaining fluid saturations within the pore structure of a porous medium. Refer to Capillary pressure. Capillary pressure. A force per unit area resulting from surface forces at the interface between two immiscible fluids. Carbonates. Sedimentary rocks primarily composed of calcium carbonate (limestone) or calcium magnesium carbonate (dolomite), which form many petroleum reservoirs. Carbon dioxide flooding. A recovery process in which carbon dioxide is injected into an oil reservoir to improve recovery.
333
...
Cementation. The process of precipitation or growth of a binding material around grains or fragments of rock. Chase gas. Gas used to displace another phase in an enhanced recovery process. Chemical flooding. A recovery process in which chemicals added to water are injected into an oil reservoir to improve recovery. Choke. An orifice installed in a line to restrict the flow and control the rate of production. Clastics. Sedimentary rocks composed of fragments of . pre-existing rocks; sandstone is a clastic rock. Clay lattice. A three-dimensional pattern of clay parts in space. Compaction. A decrease in volume of sediments as a result of compressive stress, usually resulting from continued depositional loading by accumulation of overlying sediments. Compressibility. The rate of change in volume of rock and fluids with decrease in pressure. Compressibility is a major contributor to recovery efficiency and a cornerstone of reservoir performance. Condensate. A mixture of pentanes and heavier hydrocarbons recovered as a liquid from field separators, scrubbers or other gathering facilities, or at the inlet of a processing plant before the gas is processed. Conductivity. A property of an electrical conductor defined as the electrical current per unit area divided by the voltage drop per unit length. Conformance efficiency. The fraction of total reservoir volume that is contacted by injected fluid as a result of discontinuities in the reservoir; also referred to as continuity factor. Conglomerate. A sedimentary rock composed of coarse-grained rock fragments, pebbles or cobbles cemented together in a fine-grained matrix. Coning. A cone of gas or water that forms in the reservoir due to pressure drawdown at the perforations. Connate water. The original water ofdeposition trapped in the interstices of the reservoir rock. Conventional crude oil. Crude oil that, at a particular time, can be technically and economically produced through a well using normal production practices and without altering the natural viscous state ofthe oil.
Conventional natural gas. Natural gas that occurs in a normal, porous, permeable reservoir rock and that, at a particular time, can be technically and economically produced using normal production practices. Cricondentherm. Maximum temperature at which two phases (for example, liquid and vapour) can exist. Critical gas saturation. Saturation at which free gas in a reservoir becomes mobile. Critical pressure. The pressure required to condense a gas at the critical temperature, above which regardless of pressure, the gas cannot be liquefied: Critical temperature. That temperature above which a substance can exist only in the gaseous state, no matter what pressure is exerted. Crude bitumen. A naturally occurring viscous mixture consisting mainly of pentanes and heavier hydrocarbons. Its viscosity is greater than 10 000 ml'a-s measured at original temperature in the reservoir and atmospheric pressure, on a gas-free basis. Crude bitumen may contain sulphur and other nonhydrocarbon compounds and in its natural viscous state is not normally recoverable at a commercial rate through a well. Crude oil. A mixture, consisting mainly of pentanes and heavier hydrocarbons, that exists in the liquid phase in reservoirs and remains liquid at atmospheric pressure and temperature. Crude oil may contain small amounts of sulphur and other nonhydrocarbon compounds, but does not include liquids obtained from the processing ofnatural gas. Classes ofcrude oil are often reported on the basis of density, sometimes with different meanings. Acceptable ranges are as follows: • Light:
less than 870 kg/m' (greater than 31.1 0 API)
• Medium:
870 to 920 kg/m'' (31.1 to 22.3° API)
• Heavy:
920 to 1000 kg/rrr' (22.3 to 10° API)
• Extra-heavy: greater than 1000 kg/rrr' (less than 10° API) Heavy or extra-heavy crude oils, as defined by the density ranges given, but with viscosities greater than 10 000 ml'a-s measured at original temperature in the reservoir and atmospheric pressure, on a gas-free basis, would generally be classified as crude bitumen.
334
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _1
GLOSSARY
D' Arcy's Law. The basic law of fluid flow through a porous medium that expresses how easily a fluid of a certain viscosity flows through a rock under a pressure gradient. Decision tree. A graphical summary of the possible outcomes and probabilities of the events that comprise a project. Density. The ratio of the mass of an object to its volume. Density log. A radioactivity log for open-hole surveying that responds to variations in the specific gravity of formations; an excellent porosity-measuring device, especially for shaly sands. It is a contact log (i.e., a detector held against the wall of the hole). The tool emits neutrons and then measures the secondary gamma radiation that is scattered back to the detector. Depositional environment. The conditions under which sediments were laid down. Differential liberation. The liberation of gas from oil as pressure is reduced wherein the evolved gas is separated from its associated oil; usually the physical model related to transport ofoil and gas through the formation during the majority of the primary depletion life. Dip. The angle at which a stratum is inclined from the horizontal. Discounted cash flow. Future cash converted to present conditions using an appropriate discount rate. Displacement bias. A shift of the whole frequency distribution curve to higher or lower values. Displacement efficiency. The fraction of initial oil saturation that is displaceable by a given injection fluid. Displacement process. The process by which oil is displaced by water, gas, or another fluid. Disposal well. A well used for the disposal of salt water. The water is pumped into a subsurface formation sealed offfrom other formations by impervious strata of rock. Dolomite (CaMg(CO J)2)' A common rock-forming mineral. Dolomitization. The process whereby limestone is altered to dolomite by the substitution of magnesium carbonate for a portion ofthe original calcium carbonate.
Drillstem test. The procedure used to gather data on a formation to determine its potential productivity before installing casing in a well. In the drillstem testing tool are a packer, valves or ports that may be opened and closed from the surface, a sample chamber and a pressure-recording device. The tool is lowered in the wellbore on a string of drill pipe and the packer set, isolating the formation to be tested from the formations above and below and supporting the fluid column above the packer. A port on the tool is opened to allow the trapped pressure below the packer to bleed off into the drill pipe, gradually exposing the formation to atroospheric pressure and allowing the well to produce to the surface, where the well fluids may be sampled and inspected. From a record ofthe pressure readings, a number offacts about the formation may be inferred. Efflux. Quantities of hydrocarbons, water or other fluids that leave a reservoir or zone of interest via permeable formation boundaries. Electrical conductivity. Used for estimating reservoir properties; reciprocal ofelectrical resistivity. Refer to Conductivity. Electrical resistivity. The reciprocal of electrical conductivity; used for estimating properties such as water saturation and fracture porosity. It is one of the most useful measurements in borehole geophysics. Enhanced oil recovery. Refer to Recoveryenhanced. Established reserves. Those reserves recoverable under current technology and present and anticipated economic conditions, specifically proved by drilling, testing or production, plus that judgement portion of contiguous recoverable reserves that is interpreted, from geological, geophysical or similar information, to exist with reasonable certainty. This is a term that has been used historically in Canada, particularly by regulatory agencies, and typically comprises proved reserves plus one-half probable reserves. Ethane. In addition to its normal scientific meaning of C2H6 (a colourless, odourless gas of the alkane series), a mixture mainly of ethane that may contain some methane or propane. Evaporite. Deposits of mineral salts from sea water or salt lakes due to evaporation of the water. Expectation. The mean of all possible outcomes of an event.
335
..
DETERMINATION OF OIL AND GAS RESERVES
Facies. Part of a bed of sedimentary rock of similar depositional environment, composition, appearance and properties. Fault. A break in subsurface strata. Often strata on one side of the fault line have been displaced (upward, downward, or laterally) relative to their original position. Fault plane. A surface along which faulting has occurred. Filtrate. A fluid that has been passed through a filter. Fines migration. The dislocation and movement of fine particles within a reservoir. Fines migration can cause damage or impair permeability by blocking pore throats. Flash liberation. The liberation of gas from oil as pressure is reduced wherein the evolved gas remains in contact with the liquid phase. Flow test. A test of the ability of a well to produce fluids usually at a constant rate. Fluid saturation. The fraction of the pore volume occupied by fluid. Fluid viscosity. Internal friction of a fluid, caused by molecular interactions, that makes it resist a tendency to flow. Fold. A flexure of rock strata into arches and troughs, produced by earth movements. Formation heterogeneity. Variation both laterally and vertically of properties such as porosity, permeability, and formation thickness. Formation imaging. Logs that generate images (or "pictures") of the borehole from various sources including sonic and resistivity devices. Formation pressure. The pressure in a formation at a defined depth. Formation temperature. The temperature at a given point within a formation. Temperature usually increases with depth. . Formation volume. The volume of fluid, at formation pressure and temperature, that results in one barrel of stock tank oil. Fractional flow. Phase flow rate as a fraction of total flow rate. Fracturing. A stimulation to increase productivity that results in the formation ofa fracture in the wellbore area; conducted mostly in clastics.
336
Free-water I~vel. The level or depth at which capillary pressure IS equal to zero and which, in rocks ofvari_ able pore structure, is the only truly level referenc line between hydrocarbons and water. e Friable. Describes a substance that is easily rubbed,. crumbled, or pulverized into powder. Gamma ray detector. A device that is capable of sensing and measuring the amount of gamma particles emitted by certain radioactive substances. Gas. Refer to Natural gas. Gas chromatography. The process of separating constituents of a mixture by permitting a solution of the mixture to flow through a column ofadsorbent on which the different substances arc selectively separated into distinct bands or spots. Gas compressibility factor. A factor used to correct the Ideal Gas Law (pv = nRT) to actual measurements. Gas-oil ratio. The ratio of gllll in solulion to the 011 volume in which it is dissolved, u.ually expretled in cubic feet of gas per barrel of liquid It 101.325 kPa (14.65 psia) and 15.6'C (60·F). Genetic sand unit. Formation consisling ofsand. from the same origin. Geostatistlcs. A specific statistical technique (based on the statistics of regionalized variables) that uses the position as well as the magnitude of a parameter; classical statistics' docs not generally use position, Other spatial statistics methods also exist. Gravity drainage. The movement of oil in a reservoir toward a wellbore resulting from the force of gravity. Gravity override. Preferential movement of one fluid over another due to density differences. Gross pay. The gross economically productive thickness of a formation containing hydrocarbons. Gross swept volume. The reservoir rock volume that is swept by injected fluid. Heterogeneity. A lack ofuniformity in formation properties such as permeability, porosity and thickness. Homogeneity. Uniformity ofreservoir properties in all directions. Horizontal sweep efficiency. The areal fraction of a pattern contacted by the injected fluid; also referred to as areal sweep efficiency.
GLOSSARY
Horizontal waterflood scheme. The injection ofwater in a pattern of wells with oil production from wells completed between injectors. Hybrid sand unit. A formation with sands from different origins. Hydrate. A hydrocarbon and water compound that forms under reduced pressure and temperature in gathering, compression, and transmission facilities for gas; flakes of hydrate resemble snow or ice and impede fluid flow. Hydrocarbon pore volume. The pore volume in a reservoir containing hydrocarbons; the product of hydrocarbon-filled thickness, porosity, and hydrocarbon saturation usually expressed for a unit area. May be represented on a contour map as a type of volumetric map. Hydrodynamic flow. The motion and action of water and other liquids in the subsurface. Hydrodynamic trap. An oil or gas reservoir trapped by surrounding water movement; usually leads to tilted water-oil contacts. Hydrodynamics. The study of the motion of a fluid and of the interactions of the fluid with its boundaries, especially in the incompressible ideal (frictionless) case. Hydrostatic head. The pressure exerted by a body of water at rest. Hysteresis. A change in process path in successive experimental tests. Ideal Gas Law. The volume occupied by an ideal gas depends only upon temperature, pressure, and the number of molecules (moles) present (pv = nRT). Imbibition. The increase in saturation of the wetting phase in a porous medium with time. Improved recovery. Refer to Recoveryimproved. Influx. Quantities of hydrocarbons, water or other fluids that enter a reservoir or a designated portion ofa reservoir through permeable formation boundanes. Initial reserves. A term often used to refer to reserves prior to deduction of any production. Alternatively, initial reserves can be described as the sum of remainingreserves and cumulative productionat the time of the estimate.
Initial volumes in place. The gross volume of crude oil, natural gas and related substances estimated, at a particular time, to be initially contained in a reservoir before any volume has been produced and without regard for the extent to which such volumes will be recovered. Injection. The pumping of fluids into the reservoir via wellbores, for wellbore conditioning or stimulation or for improved recovery operations. In situ recovery. A term that is used, when referring to oil sands, for the process of recovering crude bitumen from oil sands other than by surface mmmg. Intercalation. Insertion of a bed or stratum of one material between layers of another material. Interfacial tension. The force per unit length existing at the interface between two immiscible fluids. Irreducible water saturation. The minimum water saturation that can be obtained in a reservoir under normal operations. Isochrone. A line on a chart connecting all points having the same time of occurrence of particular phenomena or of a particular value of a quantity. Isolating packers. Devices used for isolating an interval in a well. Isopach map. A geological map of subsurface strata showing contours of the thickness of a given formation underlying an area; one type of volumetric map. Isotherm. A line connecting points of equal temperature. Isothermal. Having constant temperature; at constant temperature. J function. A dimensionless grouping of the physical properties of a rock and its saturating fluids proposed by Leverett. Kerogen. A solid bituminous substance occurring in certain shales that decomposes to oil and natural gas when heated. Klinkenberg. Mathematical correction oflaboratory air permeability measurements (made on formation material) into equivalent liquid permeability values, necessitated by gas slippage in pores. Laterolog. A resistivity measuring device using electrodes in which a current is forced through the formation in a sheet ofpredetermined thickness, so that the measurement involves a limited vertical extent.
337
.
ii\f. .",~,~
DETERMINATION OFOIL AND GASRESERVES
Liquefied petroleum gases. A term commonly used to refer to hydrocarbon mixtures consisting predominantly of propane and butanes. In Canada, ethane is also frequently included. Lithification. The conversion of unconsolidated deposits into solid rock by compaction and cementing together of the individual rock grains. Lithology. The description ofthe physical character of a rock as determined by eye or with a low-power magnifier; based on color, structures, mineralogic components, and grain size. Mandrel. A cylindrical bar, spindle, or shaft around which other parts are arranged or attached, or that fits inside a cylinder or tube. Marketable natural gas. Natural gas that meets specifications for its end use, whether it occurs naturally or results from the processing of raw natural gas. Field and plant fuel and losses are excluded, excepting those related to downstream reprocessing plants. The heating value of marketable natural gas may vary considerably, depending upon its composition, and therefore quantities are usually expressed not only in volumes, but also in terms of energy content. Material balance method. An engineering method of defining project performance wherein expansion of in situ rock and fluids is related to influx-efflux and production-injection streams; may be arranged to determine fluids in place or production performance. Matrix. The continuous, fine-grained material in which large grains of a sediment or sedimentary rock are embedded. Mean. The most commonly used measure of central tendency; the average value of repeated trials. The mean represents the most probable value of an estimate of reserve volume or value. Median. A measure of central tendency; the middle value or the arithmetic mean of the two middle values of a list of numbers, for a list containing an odd or even number of members, respectively. Geometrically, the value that divides a histogram or frequency distribution into two parts of equal area; also the 50 percent probability level on a cumulative distribution function or expectation curve.
338
Methane. In addition to its normal scientific meaning of CH. (a light, odourless, colourless gaseous hydrocarbon), a mixture mainly of methane that ordinarily may contain some ethane, nitrogen helium or carbon dioxide. • Miscibility. The tendency or capacity of two or more liquids to form a uniform blend, that is, to dissolve in each other; degrees are total miscibility, partial miscibility, and immiscibility. Miscible flooding. A recovery process in which a fluid (a "solvent") that is capable of dissolving into the crude oil it contacts is injected into an oil reservoir to improve recovery. Micellar flooding. The addition of surfactants to injected water to reduce interfacial tension. Micro-fractures. Fractures not easily seen by the naked eye; might be seen in thin sections. They usually feed macro-fractures. Microlog. A wellbore resistivity log recorded with electrodes mounted at short distances from each other in the face ofa rubber-padded microresistivity sonde and with different depths of investigation. Comparison of the two curves identifies mudcake which indirectly identifies the presence of permeable formation. Microporosity. Porosity that is visible only at high magnification and that is generally not effective. Mobility. The ratio ofthe permeability ofa given phase to the viscosity of that phase. Phase mobility is an indication of how easily that phase moves in the reservoir. Mobility ratio. The ratio of the mobility of the displacing phase behind the flood front to the displaced phase ahead of the flood front. Mode. A measure of central tendency; the most commonly occurring value of a set of numbers. Mole. An amount of substance of a system which contains as many elementary units as there are atoms of carbon in 0.012 kilogram of the pure nuclide carbon-12; the elementary unit must be specified and may be an atom, a molecule, an ion, an electron, a photon, or even a specified group of such units. Morphology. The observation of the form oflands.
,"
GLOSSARY
Mudcake. The residue that forms on the wall of the borehole as the drilling mud loses filtrate into porous and permeable formations; also called well cake or filter cake.
Nuclear magnetism inject log. A tool that uses a pulsed nuclear magnetic resonance analyzed to determine fluid content, total and free fluid porosity, and permeability.
Mud-gas log. The recording of information derived from examination and analysis offonnation cuttings made by the bit and mud circulated out of the hole. A portion of the mud is diverted through a gasdetecting device and examined under ultraviolet light to detect the presence of oil or gas. Often carried out in a portable laboratory set up at the well.
Oil sands. Deposits of sand or sandstone or other sedimentary rocks that contain crude bitumen.
Natural fracture. A discontinuity in rock caused by diastrophism, deep erosion of the overburden, or volume shrinkage. Examples would include shales that lose water, the cooling of igneous rock, and the desiccation of sedimentary rock.
Pentanes plus. A mixture mainly of pentanes and heavier hydrocarbons, which ordinarily may contain some butanes, and which is obtained from the processing of raw gas condensate or crude oil.
Natural gas. A mixture of lighter hydrocarbons that exist either in the gaseous phase or in solution in crude oil in reservoirs but are gaseous at atmospheric conditions. Natural gas may contain sulphur or other nonhydrocarbon compounds. Natural gas liquids. Those hydrocarbon components that can be recovered from natural gas as liquids including, but not limited to, ethane, propane, butanes, pentanes plus, condensate, and small quantities of nonhydrocarbons. Net present value. The value obtained when all cash flow streams, including the investment, are discounted to the present and totalled. Neutron log. A radioactive device that emits high energy neutrons and records a curve which responds primarily to the amount of hydrogen in the formation. Thus, in clean formation where the pores are filled with water or oil, the neutron log measures the amount of liquid-filled porosity. Nonconventional crude oil. Crude oil that is not classified as conventional crude oil. An example would be kerogen contained in oil shale deposits. Bitumen is also generally included in the nonconventional crude oil category as a matter of practice, although some wells may produce at commercial rates without steam injection. Also referred to as unconventional crude oil. Nonconventional natural gas. Natural gas that is not classified as conventional natural gas. An example would be coal-bed methane. Also referred to as unconventional natural gas.
Oolite. A spherical to ellipsoidal body, 0.25 to 2.00 mm in diameter, which mayor may not have a nucleus, and has concentric or radial structure or both; usually calcareous, but may be hematitic or of other composition.
Permeameter. A device for measuring permeability by measuring the flow of fluid through a sample across which there is a pressure drop. Petroleum. A naturally occurring mixture consisting predominantly of hydrocarbons in the gaseous, liquid or solid phase. Petroleum reservoir (pool). A porous and penneable underground rock formation that contains a natural accumulation ofcrude oil or natural gas and related substances, or combinations ofthem, that is confined by impermeable rock or water barriers, and that is individual and separate from other reservoirs. Phase behaviour. The equilibrium relationships between water, liquid hydrocarbons, and dissolved or free gas, either in reservoirs or as separated aboveground in gas-oil production facilities. Polymer flooding. The addition ofpolymers to injected water to improve mobility ratios and increase oil recovery. Pore volume. The pores in a rock considered collectively; the product of porous thickness times porosity. May be represented on a contour map, a type of volumetric map. Porosimetry. The measurement of the porosity of reservoir rocks. Porosity. The volume of the pore space expressed as a percentage of the total volume of the rock mass. Pressure depletion. Pressure decline in a reservoir due to oil or gas production. Pressure transient analysis. The estimation ofreservoir properties from measurements offlow, buildup and drawdown pressures.
339
DETERMINATION OFOILANDGASRESERVES
Primary recovery.
Refer to Recovery - primary.
Production tests. Tests conducted to determine the productivity of a given reservoir. Propane. In addition to its normal scientific meaning of C,H. (a heavy, colourless hydrocarbon of the paraffin series), a mixture mainly of propane that ordinarily may contain some ethane or butanes. Pseudo-critical and pseudo-reduced properties (temperature and pressure). Properties of pure hydrocarbons are often the same when expressed in terms oftheir reduced properties. The same reducedstate relationships often apply to multicomponent systems if "pseudo" critical temperatures and pressures are used rather than the true critical properties of the systems. The ratios of the temperature and pressure of interest to the pseudo-critical temperature and pressure are called the pseudo-reduced temperature and pressure respectively. Pulsed neutron log. A special cased-hole logging tool that uses radioactivity reaction time to obtain measurements ofwater saturation, residual oil saturation, and fluid contents in the formation outside the casing of an oil well. PVT data. Information describing the physical inter-relationship ofpressure, volume, and temperature of reservoir fluids and various production and injection streams. Pyrobitumen. Any of various dark-colored, relatively hard, nonvolatile hydrocarbon substances often associated with mineral matter, which decompose upon heating to yield bitumens. Pyrolysis. The breaking apart of complex molecules into simpler units by the use ofheat, as in obtaining gasoline from heavy oil. Raw natural gas. Natural gas as it is produced from the reservoir prior to processing. It is gaseous at the conditions under which its volume is measured or estimated and may include varying amounts of heavier hydrocarbons (that may liquefy at atmospheric conditions) and water vapour. May also contain sulphur and other nonhydrocarbon compounds. Raw natural gas is generally not suitable for end use. Recovery - enhanced. A term that, in Canada, is equivalent to improved recovery.
Recovery - improved. The extraction of addif al '1 Ion cru de 01 , natural gas and related substances fr . hr h om reservoirs t oug a production process other th nat~ral depletion. Includes both secondary a : tertiary recovery processes such as pressure maintenance, cycling, waterflooding, thermal methods ~he~ic~1 flo?ding, and the use of miscible and immiscihls displacement fluids. Recovery - primary. The extraction of crude oil natural gas and related substances from reservoirs utilizing only the natural energy available in the reservoirs. Recovery - secondary. A term frequently used to describe the extraction of additional crude oil natural gas and related substances from reservoirs through pressure maintenance schemes such as waterflooding or gas injection. Recovery - tertiary. A term frequently used to describe the extraction of additional crude oil. natural gas and related substances from reservoirs using recovery methods other than natural depletion or pressure maintenance. A tertiary proeClllcan be implemented without a precedinll prinwy or secondary recovery scheme. Remaining reserves. Initial reserves less cumulative production at the time of the estimate. Reservoir.
Refer to Petroleum reservoir.
Reservoir continuity. No interruption of a reservoir by faults, facies changes. or any other type of heterogeneity. Residual 011 saturation. Following a recovery process, the oil saturation at which oil will no longer flow in a normal immiscible water-oil system. Resin. Any of a class of solid or semisolid organic products of natural or synthetic origin with no definite melting point, generally of high molecular weight; most resins are polymers. Resistivity. The electrical resistance offered to the passage of current; the inverse of conductivity. Resistivity log. The measurement of subsurface electrical resistivity accomplished either by sending current into the formation and measuring the ease of electrical flow or by inducing an electrical current into the formation and measuring how large it is. Risk. The probability ofloss or failure.
340
GLOSSARY
Salt dome intrusive. A subsurface mound or dome of salt. Sandwich loss. The volume of oil remaining unswept at the top of a reservoir after water flooding or at the bottom of the reservoir after gas or miscible flooding. Saturation. Refer to Fluid Saturation. Saturated oil. Oil that contains all the gas that is capable of dissolving given the compositions of that oil and gas at the particular temperature and pressure. Saturation pressure. Also known as bubble-point pressure; the pressure at which the first bubble of gas comes out of solution. Secondary recovery. Refer to Recoverysecondary. Seismic. The measurement of the response to energy waves travelling through rock layers. The energy waves may be created by earthquakes, explosives or by dropping or vibrating a heavy weight. Some energy is reflected whenever the waves cross an interface ofrock layers of distinctly different properties. Measurements can be made at the surface of travel time, which may be related to depth, and wave amplitude variations, which may relate to changes in rock properties (porosity, etc.). Separator. An oilfield vessel or series of vessels in which pressure is reduced so that the dissolved gas associated with reservoir oil is flashed off or removed as a separate phase. Also known as gas separator, oilfield separator, oil-gas separator, and oil separator. Shrinkage. The decrease in volume of a liquid phase caused by the release of solution gas or by the thermal contraction of the liquid; the reciprocal of formation volume factor. Shrinkage factor. The reciprocal of the formation volume factor expressed as barrels of stock tank oil per barrel of reservoir oil. Solution gas. Natural gas that is dissolved in crude oil in the reservoir at original reservoir conditions and that is normally produced with the crude oil; also known as dissolved gas. Solvent flooding. Refer to Miscible flooding. Sonde. A logging tool assembly, especially the device in the logging assembly, that senses and transmits formation data.
Sonic log. A device that measures the time required for a sound wave to travel through a definite length of formation. Refer to Acoustic log. Sour gas. Natural gas that contains corrosive, sulphurbearing compounds such as hydrogen sulphide, sulphur dioxide, and mercaptans. Specific gravity. The ratio of the density of a material to the density of some standard material, such as water at a specified temperature, 4'C or 60'F or (for gases) air at standard conditions ofpressure and temperature. Spontaneous potential. A recording of the difference between the electrical potential of a movable electrode in the borehole and the electrical potential of a fixed surface electrode. Stock tank cubic metre. One cubic metre of oil at standard temperature and atmospheric pressure. Static gradient. Pressure measured in a wellbore at various depths while a well is shut in. Stratification. A structure produced by deposition of sediments in beds or layers (strata), laminae, lenses, wedges, and other essentially tabular units. Stratigraphic trap. A type of reservoir capable of holding oil or gas, in which the trap is formed by a change in the characteristics of the formationwhich could be loss ofporosity and permeability or a break in its continuity. Stringer. A narrow vein or irregular filament of mineral traversing a rock mass of different materials. Structure map. A map showing contour lines drawn through points of equal elevation on a stratum, key bed, or horizon, in order to depict the attitude ofthe rocks. Structural trap. A type ofreservoir containing oil and! or gas, formed by deformation of the earth's crust that seals off the oil and gas accumulation in the reservoir, forming a trap. Anticlines, salt domes, and faulting of different kinds form structural traps. Sulphur. As used in the petroleum industry, the elemental sulphur recovered by conversion of hydrogen sulphide and other sulphur compounds extracted from crude oil, natural gas or crude bitumen. Surface loss. The quantity of natural gas removed at field processing plants as a result ofthe recovery of liquids and related products and the removal of nonhydrocarbon compounds, plus the gas used for fuel; also referred to as shrinkage.
341
-
DETERMINATION OFOILAND GASRESERVES
Surfactant. A soluble compound that reduces the surface tension of liquids, or reduces interfacial tension between two liquids or a liquid and a solid. Sweep efficiency. The volume swept by a displacing fluid divided by the total volume being flooded. Sweet gas. A petroleum natural gas containing no corrosive components, such as hydrogen sulphide, sulphur dioxide, and mercaptans. Synthetic crude oil. A mixture of hydrocarbons derived by upgrading crude bitumen from oil sands, and kerogen from oil shales or other substances such as coal. May contain sulphur or other nonhydrocarbon compounds and has many similarities to crude oil. Tertiary recovery. Refer to Recovery - tertiary. Thermal conductivity. The heat flow across a surface per unit area per unit time, divided by the negative of the rate of change of temperature with distance in a direction perpendicular to the surface. Tilts. Blocks that have received a marked tilt in regions of block faulting. Regional tilts occur on the margins of basins of subsidence in the earth's crust. Tool resolution. The precision of a tool to investigate a given property. Transition zone. The interval directly above the free water level in a reservoir where capillary effects result in significant changes in water and hydrocarbon saturations in response to pore structure variations and elevation. Transmissibility. The ability of a reservoir to conduct fluids spatially in response to pressure differentials. Depends upon permeability and formation flow geometry. Production potential depends heavily upon reservoir transmissibility. Trap. A mass of porous, permeable rock that is sealed on top and down both flanks by nonporous, impermeable rock that prevents the free migration of hydrocarbons and concentrates them in a limited space. Uncertainty. The spectrum ofpossible outcomes of an evaluation. Unconformity. Lack of continuity in deposition between rock strata in contact with one another corresponding to a gap in the stratigraphic record; the surface of contact between rock beds in which there is a discontinuity in the ages of the rocks.
342
Unconsolidated sand. A sand formation in who h individual grains are not cemented together. I/~ unconsolidated sandstone produces oil or g it will produce sand if not controlled or correcte~' Undersaturated oil. Oil that is capable of absorbin g more gas than an is IS nresenr present lin the reservoir. Undersaturated ~i~ typically displays relatively low compressibility and hence a rapid pressure decline with production. Undersaturated oil reservoir. A reservoir that is above the bubble-point pressure. Ultimate potential recovery. A term sometimes used to refer to an estimate at a particular time of the initial reserves that will have become developed in an area by the time all exploratory and development activity has ceased, having regard for the geological prospects of the area, the known technology, and the anticipated economic conditions. It includes cumulative production; remaining proved, probable and possible reserves; and future additions to reserves through extensions and revisions to existing pools and the discovery of new pools. It may also be described as initial reserves plus those other resources that may be recoverable in the future. Unitization. A term denoting the joint operation of separately owned producing leases in a pool or reservoir. Upgrading. The process ofconverting crude bitumen or heavy crude oil into synthetic crude oil. Utilization rate. In an enhanced oil recovery process, the amount of gas or fluid injected per incremental oil recovered. Variability bias. An alteration in the shape of a frequency distribution curve. Vertical sweep efficiency. The vertical fraction of reservoir swept by injected fluid. Vertical waterflood scheme. The injection of water at wells completed at the bottom of the formation; oil production is from wells completed at the top ofthe formation. Vesicle. A cavity in lava formed by entrapment of a gas bubble during solidification. Viscous fingering. Faster advance ofa displacing phase as compared to the displaced phase due to an unfavorable mobility ratio.
GLOSSARY
Volumetric estimation. An estimate of hydrocarbon or water volume based on a combination of volumetric maps and other data which in total must account for the reservoir area, thickness, porosity, and hydrocarbon and water saturation. Volumetric mapping. A contour map of a parameter or combination of parameters that relate to reservoir volume. Voidage. The reservoir volume of hydrocarbons and water removed from the formation via wellbores during a term of producing operations. Voidage replacement. The volume at reservoir conditions of fluids injected into a producing pool to offset fluid withdrawals during depletion. Voidage replacement ratio. The quotient of voidage replacement divided by reservoir voidage. Vugs. Pore spaces that are larger than would be expected from the normal fitting together of the grains that compose the rock framework. Vugs are often formed during dolomitization.
Water channelling. Preferential movement of water towards a wellbore due to unfavourable mobility ratio and pressure drawdown at the wellbore or due to the presence of higher permeability streaks. Waterflooding. An improved recovery process in which water is injected into a reservoir to increase oil recovery. Water injector. A well in which water has been injected into an underground stratum to increase reservoir pressure. Water saturation. Portion of the pore volume occupied by water. . Weighted-mean. The number obtained by multiplying each value of x by the probability (or probability density) ofx and then summing (orintegrating) over the range of x. Well density. The intensity of drilling in a given area. Wetting phase. The liquid phase (oil, gas or water) that
"wets" reservoir rock. Wire line. A rope composed of steel wires twisted into strands that are in turn twisted around a central core of hemp or other fiber to create a rope of great strength and considerable flexibility; used as drilling, coring, servicing, and winch lines.
343
BIBLIOGRAPHY
The following are additional recommended references that have not been cited in the text.
Determination of In-Place Resources Alexander, L.G. "Theory and Practice ofthe ClosedChamber Drillstem Test Method." 51st Annual SPE Fall Technical Conference, New Orleans, LA, SPE Paper No. 6024, 1976. AmericanPetroleum Institute. "Sampling Petroleum Reservoir Fluids." API RP 44, Washington, DC, 1966. Amyx, 1.W.,Bass, D.M., and Whiting, R.L. Petroleum Reservoir Engineering. McGraw-Hill, New York, NY, 1960. Archer,1.S. "Reservoir Volumetrics and Recovery Factors." In Developments in Petroleum Engineering, Elsevier Science PublishingCo., NewYork,NY,1985. Bankhead, C.C., Jr. Processing of Geological and EngineeringData in Multipay Fields for Evaluation. Trans., AIME, Reprint Series #3, 1970. Bujnowicz, R. "PVT Data Generation, Reportingand General Use." 39th Annual Tech. Meeting, Petroleum Society ofCIM, Calgary, AB, Paper No. 88-39-66, 1988. Calhoun,1.e., Jr. Fundamentals ofReservoir Engineering. University of OklahomaPress, Norman, OK, 1947. Campbell, J.M., and Holander, D.P. "The Effect of Pore Configuration,Pressure and Temperature on Rock Resistivity." Paper presented at 7th Annual SPWLA Logging Symposium, May 1966. Celbuliak, N., Hamp, T., Mayder, A., Shaw, 1., and Vokey, G. "Coring for Connate Water Saturation - Utikuma Keg River Sandstone." JCPT, Nov. 1990. Core Laboratories. "Applications of Core Data in Integrated Reservoir Description and Exploitation." Version 1.2, Calgary, AB, 1990. Craft, B.C., and Hawkins, M.F. Petroleum Reservoir Engineering. Prentice-Hall, Inc., New York, NY, 1959.
Dullien, F.A.L. Porous Media Fluid Transport and Pore Structure. Academic Press, New York, NY, 1979. Enderlin, M.B., Hansen, D.K.T., and Hoyt, B.R. ''The Role of Rock Volumes in Log to Core Integration." Paper presented at CWLS 12th Formation Evaluation Symposium, Calgary, AB, Sep. 1989. EnergyResources ConservationBoard. Gas Well Testing Theory and Practice. Guide G-3, Calgary, AB, 1979. - - - . Pressure and Deliverability Testing Oil and Gas Wells. Guide G-40, Calgary, AB, 1990.
Fatt, 1. "Effect of Overburden and Reservoir Pressure on Electrical Logging Formation Factor." AAPG Bulletin, Vol. 41, No. II, 1957. Gatlin, C. Petroleum Engineering - Drilling and Well Completions. Prentice-Hall, Inc., Englewood Cliffs, NJ, 1969. Hensel, W.M. Jr., Honarpour, M.M., Sprunt, E.S., and York, C.E. "Compilation of Electrical Resistivity Measurements Performed by Twenty-Five Laboratories." The Log Analyst, Jan.-Feb. 1988. Hitchon, B. "Geothermal Gradients, Hydrodynamics and Hydrocarbon Occurrences,Alberta, Canada." AAPG Bull., Vol. 68. 1984. International Human Resource DevelopmentCorp. Video Library for Exploration & Production Specialists: Cranquist, C. "Reserves Estimation," PE508; Fowler, P.T., Hepburn, J.R., and Morrill, D.C. "Subsurface Mapping," GLZ02; Bradley, M.E., and Anstey, H.A. "Seismic Contouring," GP502; Boston, MA. James, S.C. "A Rapid Accurate Unsteady State Klinkenberg Permeameter." SPEJ, No. 12, 1972. Keelan, D.K. "A Critical Review of Core Analysis Techniques." Paper presented at 22nd Annual TechnicalMeeting of Petroleum Society ofCIM, Banff, AB, Jun. 1971.
345
q DETERMINATION OF Oil AND GASRESERVES
Link, P.K. Basic Petroleum Geology. OGCI Publications, Tulsa, OK, 1982. Maier, L.F. "Recent Developments in the Interpretation and Application of DST Data." JPT, Nov. 1962. Majorowicz, J.A., Jones, F.W., and Jessop, A.M. "Preliminary Geothermics ofthe Sedimentary Basins in the Yukon and Northwest TerritoriesEstimates from Petroleum Bottom-hole Temperature Data." Bull. ofCan. Pet. Geol., Vol. 36,1988. Majorowicz, J.A., Jones, F.W., Lam, H.L. and Jessop, A.M. "The Variability of Heat Flow Both Regional and With Depth in Southern Alberta, Canada: Effect of Groundwater Flow?" Tectonophysics, Vol. 106, 1984. Mitchell-Tapping, H.J. "Porosity and Permeability Relationship to Cleaning Effectiveness in Whole Core Analyses." Log Analyst, May-Jun. 1982. Murphy, R.P., and Owens, W.W. "The Use of Special Coring and Logging Procedures for Defining Reservoir Residual Oil Saturations." JPT, 1973. Ruth, D.W., and Kenny, J. "The Unsteady State Gas Permeameter." 38th Annual Technical Meeting, Petroleum Society of CIM, Calgary, AB, Paper 87-38-52, Jun. 1987. Scheidegger, A.E. The Physics ofFlow Through Porous Media (3rd ed.). University of Toronto, Toronto, ON, 1960. Schlumberger. "Roundtable: Strategies for Thin-Bed Formation Evaluation." Oilfield Review, Jul. 1991. Scholle, P.A., Bebout, D.G., and Moore, C.H. (ed.). "Carbonate Depositional Environments." AAPG, Memoir 33, 1983. Schowalter, T.T. "Mechanics of Secondary Hydrocarbon Migration and Entrapment." AAPG Bulletin, Vol. 63, No.5, May 1979. Scientific Software. Reservoir Engineering Manual. NTIS PB - 247-806. 1975. Walker, R.G. (ed.). "Facies Models." Geoscience Canada. Reprint Series I, 1979. Worthington, P.F. "Effective Integration of Core and Log Data." SCA Conference, Paper No. 9102, 1991.
Carbon Dioxide Flooding Holm, L.W. "CO, Flooding: Its Time Has Come." JPT, Dec. 1982. Klins, M.A. Carbon Dioxide Flooding - Basic Mechanisms and Project Design. International Human Resource Development Corporation, Boston, MA, 1984. Mungan, N. "Carbon Dioxide Flooding _ Fundamentals." JCPT, Jan. - Mar., 1981. - - - . "Carbon Dioxide Flooding - Applications." JCPT, Nov. - Dec. 1982. Stalkup, F.1. "Carbon Dioxide Miscible Flooding: Past, Present, and Outlook for the Future." JPT, Aug. 1978.
Crude Oil Markets Canadian Petroleum Association. 1991 Statistical Handbook. Calgary, AB, 1992. Energy Resources Conservation Board. Alberta Energy Resource Industries. Monthly Statistics, Calgary, AB.
- - - . Selected Statistics and Forecasts. Calgary, AB,1991. Petroleum Monitoring Agency Canada. Canadian Petroleum Industry. 1991 Monitoring Report. Minister of Supply and Services Canada, Cat. No. M2722/1991-2E, ISBN 0-662-19847-6, 1992. Statistics Canada. The Crude Petroleum and Natural Gas Industry. Minister of Industry, Science and Technology, Cat. No.26-213 Annual, ISSN 0068-7103,1992.
Hydrocarbon Miscible Flooding Blackwell, R.J., Wall, T., Rayna, J.R., Lindley, D.C., and Anderson, J.R. "Recovery of Oil by Displacement with Water Solvent Mixtures." Trans., AIME, Vol. 21,1960. Brigham, W.E., Reed, P.W., and Dew, J.N. "Experiments on Mixing During Miscible Displacement in Porous Media." Trans., AIME, Vol. 222; SPEJ, Mar. 1961. Claridge, E.L. "Prediction of Recovery in Unstable Miscible Flooding." SPEJ, Apr. 1972. - - - . "Design of Graded Viscosity Banks for Enhanced Recovery Processes." Ph. D. Dissertation, University of Houston, Houston, TX, Jul. 1979.
346
r
BIBLIOGRAPHY
Craig, F.E. Jr. "A Laboratory Study of Gravity Segregation in Frontal Drives." Trans., AIME. Vol. 210,1957.
Pozzi, A.C., and Blackwell, R.J. "Design of Laboratory Models for Study of Miscible Displacement." SPEJ, Mar. 1963.
Gardner, A.a., Jr., Peaceman, D.W. and Pozzi, A.L. Jr. "Numerical Calculation of Multiple Dimensional Miscible Displacement by the Method of Characteristics." SPEJ, Mar. 1964. Koval, E.J. "A Method for Predicting the Performance of Unstable Miscible Displacement in Heterogeneous Media." Trans., AIME, Vol. 228; SPEJ, Jun. 1963.
Shelton, J.L. and Yarborough, L. "Multiple Phase Behaviour in Porous Media During CO, or Rich Gas Flooding." JPT, Sep. 1977.
Peaceman, D.W., and Rachford, H.H., Jr. "Numerical Calculation of Multi-dimensional Miscible Displacement." Trans., AIME, Vol. 225; SPEJ, Dec. 1962.
Thermal Stimulation Edmunds, N.R. "An Analytical Model of the Steam Drag Affecting Oil Sands." Paper presented at the 34th Annual CIM Technical Meeting, Banff, AB, May 1983.
347
-
AUTHOR INDEX
Abdassah,D., 234, 235 Abramowitz, M., t93, 196 Adamache, I., 179, 180 Adams, R.H., 188, 196 Agbi, B., 227, 235 Aguilera, R. 214, 220 Aitcheson, 1,107,119,232,235 Alberta Energy,257-259, 265 Alberta Petroleum Marketing Commission, 291,293,296 Alexander, L.G., 345 Ali, S.M., 190, 196 Amaefule, J.O., 48, 52 American Petroleum Institute, 53, 54, 55, 64,78,80,237,249,345 Amyx, J.W., 96, 100, 345 Anderson,J.H., 179, 180 Anderson,J.R., 346 Anderson,W.G., 161, 169 Archer,J.S., 345 Archie,G.E., 66, 74 Aronofsky,lS., 167-169 Arps, J,J., 224, 225, 235, 238, 249 Arshi, A.A., 294, 296 Asgarpour, S.S., 177-180,206,213 Attanasi, E.D., 279, 280 Au, A.D.K., 215, 220 Aziz, K., 92, 95,187,188, t92, 197,214, 215,220 Babu, D.K., 207, 212 Balay, R.H., 277, 279 Bankhead, C.C., Jr., 345 Barclay,J.E., 232, 236, 243, 246, 250, 278, 280 Bass, D.M., 96, 100, 345 Bebout, D.G., 346 Beeler, P.F., 179, 180 Behie, A., 215, 220 Belvins,T.R., 188, 190, 197, 198 Belyea, H.R., 188, 197 Bennett, F.. 179, 180 Bielecki,1, 205, 209, 212 Bilozer, D.E., 179, 180 Bird, K.I., 279 Blackwell,R.I., 235, 236, 346, 347 Bloy, G., 179,180 Boberg, T.C., 188, 197 Bobrowski, F.P., 96, 100 Bodily, S.E., 108, 119 Bokhari,S.W., 179, 181 Bournazel, C., 139, 144 Bousaid, 1., 195, 198
Bowers, B., 205, 209,212 Boyd, W.E., 153 Bozac, P.G., 176, 181 Brigham, W.E., 346 British Columbia Ministry of Energy, Mines andPetroleum Resources, 257.265 Broomhall,R.W., 179, 181 Brown, J.A.C., 107, 119,232,235 Brown, S.L., 188, 199 Bruskotter, J.F. 152, 153 Buckles, R.S., 102,105, 188, 197 Bujnowicz, R., 345 Burger,D.H., 179,181 Burns, lA., 188, 197 Bursell, C.G., 190, 197 Buskirk, D.L., t79, 180 Butler,R.M., t88, 189,191, 194, 197,207, 208,212 Butler, S., 179, t80 Calhoun, i.c, Jr., 345 Campbell, A.D., 271, 279 Campbell,J.M., 268, 278, 279, 345 Canadian Petroleum Association, 346 Canadian Well Logging Society,79, 80 Cao, S., 85 Capen,E.C., 108, It9, 274, 279 Card, C., t78, 180 Cardwell, W.T., 188, 197 Carter, R.D., 125, 127, 138, 144, 153, 164, 169,170,176,181,185 Celbuliak,N., 345 Chaney, P.E., 139, 144 Chaperon.L, 207, 212 Chapman, D.S., 82, 83, 85 Chase, C.A., 214, 220, 221 Chen, S.M., 178, 180 ChevronCanada Resources, 94, 96, 97 Chierici, G.L., 139, t44 Chin, T., 96, 100 Chinna, H., 190, 198 Christian,L.D., 153 Christie, D.S., 212, 213 Christie,J.A., 243, 246, 249 Chu, C., 190, 197 Ciucci, G.M., 139, 144 Claridge,E.L., 176, 180,346 Clark, I., 275, 279 Clark, N.I., 145, 153 Closmann,P.J., 188, 190, 199 Coats, K.H., 188, 198,214,217,220 Collins, R.E., 168, 169 ComputalogGearhartLtd., 49, 52
Conn, R.F., 243, 246, 249 Cook, A.B., 96, 100 Coonibie, D., 188-190, 198 Cooper, H.E., 176, 181 CoreLaboratories, 345 Cornelius, A.J., 139, 144 Cornell, D., 146, 153 Cornish, R.G., 179, 182 Craft, B.C., 97, '100, 125, 127, 147, 153, 345 Craig, F.E., Jr., 347 Craig, F.F., 138, 144, 155-161, 164, 169, 176,180,185 Crawford, P.B., 168, 169, 188, 197 Crawly, A., 177, 180 Crichlow, H.B., 214, 220 Cronquist, C., 83, 85 Crovelli, R.A., 277, 279 Curran, R., 255, 265 Dalton, R.L., 214, 221 Danielsen, C.L., 153 Dardaganian, S.G., 161, 169 Davis, J.C., 275, 279 Davis, M.I., 228, 236 Davis, R., 179, 180 Dawson,A.G., 179, 180 de Swaan, A., 167, 169 Deans, H.A., 175, 181 Deeds, C.T., 188, 190, 199 Deming, D., 82, 83, 85 Dempsey,J.R., 217, 220 Denbina, E.S., 188, 197 Derocco, M., 188, 189, 198 Des Brisay, e.L., 139, 144 Dew, J.N., 346 Dietz, D.N., 138, 140, 144 Dillabough,lA., 189, 197 Dolton, G.L., 279 Dore, T.L., 195, 198 Doscher, T.M., 188, 197 Dranchuk, P.M., 92, 95 Drew, L.J., 268, 278, 279 Duerkson,J.H., 188, 189, 198 Dugdale, P.I., 194, 197 Duggan, J.O., 148, 153 Dullien, F.A.L., 345 Durrant, A.I., 187, 197 Dyes, A.B., 168, 170, 176, 181, 185 Dykstra, H., 188, 197 Edmunds, N.R., 347 Elenbass, r.a., 146, 153 Elfrink, E.B., 137, 144
349
DETERMINATION OF OIL AND GAS RESERVES Ellis, H.E., 137, 144 Enderlin, M.B., 345
Energy Resources Conservation Board, 82, 85,102,105,152,153,161,170,237, 240,241,249,253,345,346
Energy, MinesandResources Canada, 290, 291,296 Enger, S.R., 179, 181 Erikson, R.A., 176, 181, 185 Ershagi, 1., 234, 235 Espiritu, R., 208, 212 Fabes, L., 194, 197 Farouq Ali, S.M., 187, 188, 197 Farquharson, R.G., 194, 197 ' Fatt, 1., 345 Fetkovich, MJ., 230, 235 Flach, P.O., 209, 212, 213 Fang, O.K., 179, 181 Fontanilla, J.P., 187, 197 Freeborn, R., 207, 212 Frydl, P.M., 179, 180 Garb, FA, 266, 269, 277, 279 Gardner, A.a., Jr., 347
Gas Processors Suppliers Association, 91, 95,151,153 Gates, C.F., 194, 197 Gatlin, C., 345 Geffen, T.M., 185 Gentry, R.W., 224, 225, 227, 235 Geoghegan, J.G., 179, 180
Geotechnical Resources Ltd., 56, 64 Gilman, J.R., 215, 221 Gontijo, J.E., 188, 192, 197 Goodrich, J.H., 152, 153 Gould, T.L., 160, 170 Griffith, J.D., 179, 181 Guerrero, E.T., 150, 153,238,249 Habermann, B., 176, 181 Hamblin, A.P., 232, 236, 243, 246, 250, 278,280 Hamp, T., 345 Hamilton, J.M., 62, 64 Hankinson, R.W., 92, 95 Hanna, M., 194, 198 Hanson, O.K.T., 345 Haun, J.D., 278, 279 Havlena, D., 126. 127 Hawkins, M.F., 97,100,125,127,147, 153,345 Henderson, J.H., 217,220 Henley, D., 138, 144 Hensel, W.M., Jr., 345 Henson, W.L., 139, 144 Hermanrud, C., 85 Herring, T.R., 207, 212 Heseldin, G.M., 68, 74 Heyse1, M., 208, 212 Higgins, R.V., 164, 170
350
Hitchon, B., 345 Hahn, M.E., 275, 279 Holander, D.P., 345 Holm, L.W., 204, 346 Honarpour, M.M., 345 Home, A.L., 179, 181 Home, J.S., 195, 198 Homer, D.R., 87, 90 Houghton, J.e., 279 Howe, G.R., 176, 181 Hoyt, B.P., 345 Hurst, W., 125, 127, 138, 144 Hutchinson, C.A., 124, 127 Hutchinson, T.S. 138, 144
International Human Resource Development Corp., 345 Isaaks, E.H., 275, 279 Ivory, J., 188, 189, 198 Jack, H.H., 215, 221 Jackson, D.D., 179, 180 James, S.G., 345 Jeanson, B., 139, 144 Jenkins, G.R., 194, 196, 198 Jenkins, M.K., 176, 181 Jessop, A.M., 346 Johnson, W.M., 188, 198 Jones, F.W., 82-85, 346 Joshi, S.D., 206-209, 212 Jung, K.D., 194, 197 Kadane, K.B., 274, 279 Kahneman, D., 266, 280 Katz, D.L., 96, 100, 146, 153 Kazemi, H., 215, 221 Keefer, D.L., 108, 119 Keelan, D.K., 48, 52, 345 Kemp, C.E., 138, 144, 168, 170 Kennedy,P., 179, 180 Kenny, J., 346 Kersey, D.G., 48, 52 Khallad, A., 188, 194, 197, 199 Khan, A.M., 188, 196 Khan, A.R.• 138, 144 Kimbler, O.K., 176, 181 Kirkpatrick, J.W., 194, 196, 198 Klins, M.A., 204, 346 Kobayashi, R., 146, 153 Koval, EJ., 176, 181,347 Krukowski, J.V., 261, 265 Kuhme, A., 179, 180 Kular, G. 188-190, 198 Kuo, M.C.T., 139, 144, 161, 170,208,212 Kyte,J.R., 167,170 Lai, F.S.Y., 179, 181 Lake, J.W., 214, 221 La1, F.S., 179, 182 Lam, H.L., 82-85, 346 Langenheim, R.W., 191, 198
Lantz, R.B., 188, 197 Laurie, R.A., 179, 180 Lee, P.J., 232, 236, 243, 246, 250, 278, 279 280 ' Leighton, AJ., 164, 170 Lerche, I., 85 Leverett, M.C., 69, 74 Lewis, D., 179, 180 Lindley, D.C., 346 Link, P.K., 346 La, H.Y., 188, 191, 194, 197, 198 Loder, W.R., 179, 180 Lohec, R.E., 234, 236 Long, D.R., 228,236 Longstaff, WJ., 178, 182,214,221 Lowe, K., 188-190, 198 Lusztig, P.A., 264, 265 Ma, T.D., 179, 181 MacDonald, A.J., 207, 212 Mahaffey, J.L., 176, 181 Maier, L.F., 346 Mainland, G.G., 188, 198 Majoros,S., 175, 181 Majorowicz, J.A., 82, 83, 85, 346 Mallimes, R.M., 179, 181 Marschall, D.M., 48, 52 Marshal, D., 138, 144 Martinsen, R., 207, 212 Marx, J.W., 191, 198 Masse, L., 167-169 Mast, R.F., 279, 280 Masters, C., 278, 279 Mathews, C.W., 176, 181 Mattax, c.c., 167, 170,214,215,221 Mayder, A., 345 McCord, D.R., 96, 100 McCray, AW., 266, 274-278, 280 McCrossan, R.G., 239, 249 Mcintyre, FJ., 171, 180, 181 McKibbon, J.H., 126, 127, 194, 196, 198 McNab, G.S., 188, 191, 194, 197 Mead, N.H., 224, 235 Megill, R.E., 275, 278, 280 Mehra, R.K., 194, 198 Meisingset, K.K., 85 Meldau, R.F., 190, 196, 198 Merrick, RJ., 153 Metwally, M., 194, 198 Miller, B.M., 278, 280 Miller, C., 257, 265 Mitchel-Tapping, HJ., 346 Mollins, L.D., 137, 144
Moore, C.H., 346 Moore, D.W., 185 Morse, RA, 185 Mosteller, F., 274, 280 Mukherjee, D., 180, 181 Mungan, N., 204, 346 Murphy, R.P., 346
AUTHOR INDEX Muskat, M., 137, 139, 144 Mutalik, P., 206, 207, 213 Myhill, N.A., 190, 191, 198 Nagel, R.G., 179, 181 Natanson, S.G., 167-169 National Energy Board, 289, 290, 294-296 National Energy Board Act, 290, 296 Newendorp, P., 266, 274-278, 280 Ng, M.C., 227, 235 Nieman, R.E., 179, 181 Noble, M.D., 139, 144 Nolen, J.S., 214, 221 Odeh, A.S., 126, 127,207,212 Oglesby, KD.,,188, 198 Okazawa, T., 176, 179, 181 Olynyk, J., 178, 180 Osadetz, KG., 232, 236, 243, 246, 250, 278,280 Owens, W.W., 138, 144,346 O'Dell, S., 257, 265 PanCanadian Petroleum Ltd., 59, 60, 68, 70 Papatzcos, P., 207, 212 Papst, W., 178, 180 Parker, J.R., 278, 280 Parson, R.L., 188, 197 Patel, R.S., 179, 181 Patzek, T.W., 189, 198 Paxman, D.S., 126, 127 Pearse, J., 257, 265 Peggs, J.K., 179, 181 Perry, G.E., 152, 153 Petroleum Communication Foundation, 287,296
Petroleum Monitoring AgencyCanada, 346 Phillips, K.A., 92, 95 Pirson, S.J., 137, 144, 185 Pittman, G.M., 190, 197 Pizzi, G., 139, 144 Ploeg,J.F., 188, 189, 198 Podruski, J.A., 232, 236, 243, 246, 250, 278,280 Poettman,F.H., 146, 153 Poon, D.C., 207, 212 Pope, J.A., 178, 180 Porter, KE., 212, 213 Pow, M., 179, 180 Powell, J.D., 48,52 Pozzi, A.L., 347 Prats, M.A., 188, 189, 193, 197, 198 Pritchard, D.W.L., 179, 181 Procter, R.M., 232, 236, 243, 246, 250, 278, 280 Province of Alberta, 257, 258, 265 Pursley, S.A., 188, 198 Purvis, R.A., 92, 95, 229, 231, 236 Rachford, H.H., Jr., 347
Railroad Commission of Texas, 152, 153
Ramey, H.J., 188, 199 Rawlins, E.L., 151, 153 Rayna, J.R., 346 Reed, P.W., 346 Reinbold, E.W., 179, 181 Reisz, M.R., 211, 212 Renke, S.M., 179, 181 Rice, D.D., 278, 280 Rice, T.D., 139, 144 Richardson, J.G., 235, 236 Roberts, T.G., 137, 144 Robertson, S., 228, 236 Robinson, D.B., 92, 95 Robinson, J.G., 266, 280 Rock, N.M.S., 275, 280 Roebuck, LF., 185 Rogers, E.E., 188, 198 Romney, G.A., 194, 196, 198 Root, D.H., 279, 280 Root, P.J., 214, 221 Rotter, M.B., 188, 197 Rubin, 8., 215, 220 Russell, B., 207, 208, 213 Russell, D.G., 152, 153 Ruth, D.W., 346 Rutherford, W.M., 176, 181 Saizew, H., 194, 197 Sander, P.R., 189, 190, 198 Sanderlin,J.L., 185 Sarern, A.M.S., 160, 170 Saskatchewan Energy and Mines, 257, 265 Scheidegger, A.E., 346 Schellhardt, M.A., 151, 153 Schilthuis, R.J., 125, 127, 137, 138, 144 Schlumberger of Canada, 44, 52 Schlumberger, 64, 67, 72, 74, 346 Schoemaker, R.P., 225, 226, 235 Schoeppel, R.I., 165, 170 Scholle, P.A., 346 Schowalter, T.T., 346
Schuenemeyer, J.H., 279 Schwab, B., 264, 265 Scientific Software, 346 Scott, G.C., 212, 213 Scott, G.R., 292, 296 Scott, K., 188, 189, 198 See,D.L., 179, 181 Sepehrnoori, K., 214, 221 Seto, A.C., 176, 181 Settari, A., 214, 215, 220 Shaw, J., 345 Shell Development Company, 51, 52, 71, 74 Shepherd, D.W., 188, 198 Shelton, J.L., 347 Shipley, R.G., 188, 198 Shreve, D.R., 185, 186 Silberberg, LH., 138, 144 Singhal, A.K., 178, 180,206,213 Skjaeveland, S.M., 207, 212
Slider, H.C., 158, 160, 170, 226, 236 Sluijk, D., 278, 280 Smith, L.B., 195, 198 Smith, L.R., 152, 153 Smith, O.J.E., 215, 221 Smits, L.J.M., 72, 74 Sobocinski, D.P., 139, 144 Sorensen, L.E., 179, 181 Spencer, G.B., 96,100 Spetzeler, c., 269, 280 Springer, S.J., 178, 180,206,209,213 Sprinkle, T.L., 153 Sprunt, E.S., 345 Srivastava, R.M., 275, 279 Stael von Holstein, c., 269, 280 Stalkup, F.L, 175, 181,204,346 Standing, M.B., 92, 93, 95, 96, 100
Statistics Canada. 346 Stegeimeier, G.L., 190, 191, 198 Stegun, LA., 193, 196 Stewart, J.M., 62, 64 Stiff, H.A., 79, 80 Stiles, W.E., 186 Stoian, E., 247, 250 Stokes, D.D., 190, 198 Stone, H,L., 176, 181,216,221 Suffridge, F.E., 189, 196 Suprunowicz, R., 208, 212 Surface, R.A., 194, 197 Tarvydas, R., 257, 265 Taylor, o.c, 232, 236, 243, 246, 250, 278, 280 Taylor, H.G., 179, 182 Telford, A.S., 247, 250 Thambynayagam, C., 187, 197 Thele, K.J., 214, 221 Thomas, H.K., 92, 95 Thornton, R.W., 194, 197 Tiffin, D.L., 178, 181 Todd, M.R., 177, 178, 180, 182,214,220, 221 Towson, D., 186, 199 Tracy, G.W., 125, 127, 137, 144 Trimble, A.E., 190, 197 Tsang, P.W., 179, 181 Tversky, A., 266, 280
University of Calgary and Canadian Petroleum Tax Society, 262, 265
Valencia, L.E., 48, 52 Van Dijk, c., 190, 199 Van Everdingen, A. F., 125, 127 Van Regan, N., 179, 180 Vary, J.A., 146, 153 Vinsome, P.K.W., 215, 220 Vogel, J.V., 190, 191, 199 Vokey, G., 345 Wahl, W.L., 137, 144
351
c
Walker, R,G" 346 Wall, T., 346 Wang, P.C.C., 278, 279 Warren, A., 248, 250 Warren, I.E., 107, 116, 119,214,221 Wattenburger, R.A., 207, 208, 213 Waxman, M.H., 72, 74,188,190,199 Weinaug, C.F., 146, 153 Weinmeister, M., 42, 43 Welch, L.W., 185, 186
Welge, H,J., 138, 140, 144, 185, 186 Whiting, R.L., 96, 100, 345 Wichert, E., 92, 95 Willhite, G.P., ISS, 158, 170, 187, 199 Williams, R.L., 188, 199 Witte, M.D., 164, 169 Wong, A., 194, 196, 198 Wong, F.Y., 179, 181 Wong, T., 178, 180
Wood, K.N., 179, 182 Woodford, R.B., 179, 182 Worthington, P.F., 346 Wuckoff, R.D., 139, 144 Yang, W., 207, 208, 213 Yarborough, L., 152, 153,347 York, C.E., 345 Youngren, G.K., 214, 221 Youtz, c, 274, 280
SUBJECT INDEX
abandonment, 264
absolute minimum/maximum value approach, 107 absolnteopen flow test, 77 absolute permeability, 216
approach absolute minimum/maximum value, 107 single-value, 106 aquifers, 124 Archie equation, 66, 69
acceleration project, 310
area-weighted average, 90
abandonmentpressure, 148, 247
accounting full-cost, 316 successful-efforts, 317 accounting requirements, 316 acoustic log, 57 acquisition of data, 46, 55, 96 acquisition of data, 35 actual value, 268 additives, 189 adjustedattributedCanadianroyalties and taxes, 260 age, 258 aggregators, 297 Alberta agencies, 281 Albertaroyalty tax credit, 260 allocationfactors, 223 analogous fields, 36 pool, 249 pools, 163 reservoir, 22 analogous reservoirs,
17
analogy, 132, 244 geologic, 278 analysis compositional, 77. 80 decline, 132, 136, 137, 139, 162 Homer, 88 of data, 36, 82, 96 pressure transient, 36 PVT, 77 statistical,
107
volumetric, 158 Warren's probability, 107 analytical methods, 137, 138, 140, 156 performance predictions, 164 transient type curves, 230 water influx models, apparatus
Dean Stark, 65, 66 rising bubble, 174
124
areal extent, 112 areal sweep efficiencies,
167-168
areal yield methods, 278 arithmeticaverage, 90 arithmetic method, 8
asphaltenes in carbon dioxide flooding, 203 associated gas, 29, 145 associated gas reserves,
150
assumptions, 120 constant reservoir volume, 121 constant temperature, 121 material balance, 121 pressure equilibrium, 121 reliableproduction data, 121 representative PVT data, 121 attributedCanadianroyaltyincome, 259 auditing evaluations, 314 average Albertamarketprice, 257 area-weighted, 90 arithmetic, 90 porosity, 216 recovery factors, 240 reservoir pressure, 89 volume-weighted, 90 backpressure testing, 152 balancesheet, 263 barrelsof oil equivalent, 318 best estimate, 11 bias, 268 central, 269 cognitive, 271 displacement, 269 motivational, 269 variability, 269 bitumen, 46, 72, 103, 187 black oil simulation, 177 simulators, 214 borehole environments, 49 borrowing, 314 bottom water, 190
bottom-hole pressure, 49 temperature, 49, 81 bottom-water drive, 239 break-through carbon dioxide flooding, 203 ratio, 177 bubblepoint, 97 Butler model, 194 by-product. See related product. calculation of initial solution gas in place, 29 of oil in place, 28 calendar-day rate definition of, 223 caliper method, 57 call option, 295 Canadian crude oil exports, 291 development expense, 261 exploration anddevelopment overhead expense, 263 exploration expense, 261 oil andgas property expense, 261 refining, 290 capillaryeffects, 37 capillarypressure, 27, 67 data, 216 capital cost allowance, 263 capital costs, 256 carbon dioxide availability and cost, 202 flooding, 200, 202 carbonate plugging in carbon dioxide flooding, 203 pools, 242 carried interest, 254 cash flow, 253, 255 causesof failure in situcombustion process, 196 ceiling tests, 316 central tendency, 240 characteristics flow, 75 hydrocarbon, 131 of natural gas, 145 reservoir, 131 chasegas slug size, 178 chemicalmethod, 65-66
353
• DETERMINATION OF OIL AND GASRESERVES
classification empirical, 273 of cumulative production, 7 of miscible hydrocarbon reserves, of reserves, 4 of resources, 4 clastic pools, 242 clay Dual Water Model, 72 presence of, 72 swelling, 54 closed-chamberdrillstem tests, 75 combination drive, 134, 135, 140 of forward combustion and waterflooding, 195 combined methods, 278 thermal drive, 195 combustion process enriched air, 195 in situ, 194 commodity pricing, 253 compaction drive, 134 company grossremaining reserves. 7 netremaining reserves, 7 compositional analysis, 77, 80 simulator, 177, 214 compressibility factor, 28 fornatural gases. 93 computer mapping, 42 computer simulation Monte Carlo, 107, 206 computer solutions, 127 conditions in situ, 63 confidence level. 9 conformance efficiency, 160, 163 conical steam zone model, 191 coningand cresting, 207 connate water saturation, 65 conservation controls, 283 Conservation of Mass Law of, 149 Conservation of Matter Law of, 120 constant reservoir volume. 121 temperature, 121 constant andvariable rate test, 77 contacts fluid, 48 tilted oil-water, 38 conventional crude oil, 287 core, 45, 47. 50 analysis. 55 markets, 297 oil-base, 102
354
179
permeabilityfrom, 101 porosity data. 58 porosity from. 101 saturations from, 102 coring oil-base, 65 correlation of log and coreporosity, 62 corrosion in carbon dioxide flooding, 203 costs capital, 256 facility operating, 256 field, 256 finding, 317 general andadministrative, 256 intangible, 263 operating. 256 replacement, 317 sunk. 309 tangible, 263 well abandonment. 256 Craig-Geffen-Morse Method, 164 cresting coning and, 207 loss, 207 cross-contouring, 39 cross-plot neutron-density. 58 Crown interest, 254 royalty, 7-8 crude oil conventional, 287 density of conventional, 241 exports Canadian, 291 light, 287 markets, 253, 287 nonconventional, 288 oxidation of, 195 prices, 255 production, 288 royalty, 258 cumulative production. 7 curve relative frequency, 30 curves analyticaltransienttype, 230 empirical depletiontype, 230 cutcurves definition of, 223 cutoff values, 28 cutoffs permeability, 46 porosity, 45 cutoffs, 102 cyclic steam stimulation, 187 cycling of gas condensate reservoirs, 152
daily rates, 222 D' Arcy's Law, 53 data acquisition of. 35, 46 analysis of, 36, 58, 82 capillary pressure. 216 coreporosity, 58 interpretation of, 48 pressure buildup, 86 production and well, 216-217 PVT, 120 quality of, 85 reliability of, 31, 63, 101, 121, 136, 157, 237 rock compressibility, 216 seismic, 35 source of, 31, 81 data acquisition. 58 database, 57, 168, 237 Energy Resources Conservation Board, 58 Dean Stark apparatus, 56, 65, 66 technique, 65 decision matrices, 276 decision trees, 277 decline analysis, 136, 137, 139, 162 exponential, 20, 225-226 harmonic, 20, 229 hyperbolic, 226-229 declinecurve analysis, 18, 132 definition of, 223 methods for a group of wells, 231 methods for a single well, 224 source of data for, 222 deficiencies, 308 defining net pay, 45 degree of uncertainty, 266 Delphi Method, 274, 278 density log, 57 of conventional crude oil. 241 depletion calculations, 317 immature, 157 mechanism, 206 depletion strategy, 131 planning, 132 purpose, 131 depositional environments, 36, 242 deregulation, 291, 298 Canadian, 299 United States, 299 deterministic procedure, 8, J(}-II, 30 developed nonproducing reserves, 6 producing reserves, 6 reserves, 6
SUBJECT INDEX
dew-point reservoirs,
146
diagenesis, 37 diagram multi-contact ternary,
174
P-X, 173 ternary, 173 Dietz Method, 138, 140 differentialliberation, 97 process, 97 test, 79 dimensionless solutions,
230
dippingfaults, 39 discount rate, 264 discounted net profit before investment distribution, 116 return on investment, 308 discovered resources, 5
displacement efficiency, 167, 175, 201 process, 154 domestic needs, 284 drainage volume, ·72 drawdown test, 77
drilling incentives, 285 rotary, 75 drillstemtest, 75, 104 closed-chamber, 75 open-hole, 72
drive bottom-water, 239 combination, 134, 135, 140 combinedthermal, 195 134
compaction,
dissolvedgas, 201 edge water, 239 expansion, 239 gas cap, 134, 140, 239 gravitysegregation, 238 solutiongas, 133, 207, 238 steam, 195 water, 134, 137 drive mechanism,
147
natural, 247 drive mechanisms natural, 238 primary, 238 dry gas, 152 reservoirs, 145 Dual Water Model, 72 dual-porosity formulation, 214 dynamic implicitmethod, 215 earned depletion, 263 economic development, 282 evaluations,
limit, 142
119
sensitivity cases, 8 uncertainty, 266 economic conditions, 8 economically recoverable reserves, 253
economics, 206 edge water drive, 239 effect Forcheimer, 53
K1inkenberg, 53 effective net pay, 102 permeability, 53 porous zone, 102 porous zones, 103 efficiencies areal sweep, 167-168 displacement, 167, 201 vertical sweep, 167-168 efficiency areal sweep, 202 conformance, 160, 163 displacement, 175, 201 recovery, 203 vertical sweep, 202 volumetric sweep, 177, 202 electrical conductivity, 27 method, 65-66 electromagnetic heating, 196 empiricalclassification, 273 empiricaldepletion type curves, 230 energy equivalence, 9 energyequivalent, 318 Energy Resources Conservation Board, 58, 281 coreand cuttings storage, 58 reserve database, 237 engineering uncertainty, 266 enhanced gas recovery, 153 oil recovery, 131, 171 oil recovery simulators, 214 enriched aircombustion process, 195 environment regulatory, 253 environments borehole, 49 depositional, 36, 242 equation Archie, 66, 69 material balance, 17, 120, 122 of state, 178 volumetric, 108, 158, 175 WylieTime-Average, 58 equivalent barrels of oil, 318 energy, 318 gas, 9 oil, 9
errors in logs andcore analysis, 64 in material balance method, 123 sources of, 123, 268 estimate best, 11 single, 11 single-value, 107 volumetric, 106 estimated value, 268 estimates material balance, 30 refinement of volumetric, 43 uses of resource, 31 volumetric 27 estimation of uncertainties, 178 evaluation of explorationwells, 313 of undeveloped lands, 278 of unexploredlands, 313 evaluators, 11, 31, 34, 315 expansion drive, 239 expectation, 266 experimental methods miscibility, 173 exponential decline, 20, 225-226 exports to the USA, 286 extraction, 201
facility operating costs, 256 sizing, 142 factor average recovery, 240 compressibility, 28 formation, 66 formation volume, 28, 105, 122 gas compressibility, 28, 91, 105 gas deviation, 91, 113 gas formation volume, 91, 94 oil formation volume, 96 oil recovery, 239 recovery, 17, 31, 168, 210 shrinkage, 96 failure cyclic steam stimulation andsteam flood process, 190, 196 fair market value, 312 faults dipping, 39 FederalCompetition Act, 285 federal government, 281 fence option, 295 field costs, 256 examples, 188 field performance of miscible floods, 179 financial statements, 263
355
a DETERMINATION OFOILAND GASRESERVES
finding costs, 317 fines migration, 53
future
finite-difference method, 215 first-contact miscible process, 172 fiscal policies, 285 flash liberation process, 97 flash liberation test, 79
flood oil saturation at the start of, 160 flooding carbon dioxide, 200, 202 hydrocarbon miscible, 171 micellar, 154, 168 polyroer, 154, 168 floods horizontal, 162 horizontal miscible,
172
vertical miscible, flow
uncertainty,
294-295
gas analyses,
78
associated, 29, 145 compressibility factor, 28, 91, 105 cost allowance, 258 cresting, 208 deliverability forecasting, 151 deviation factor, 91, 113
dry, 145, 152 equivalent, 9 expansion method,
reserves and production,
reservoir performance, 219 formation factor, 66 fluid saturation, 216
heterogeneity, 50 67
gravity segregation drive, 238 grid block orientation, 166
gross
overridingroyalty, 254 pay, 44 7
half-life reserves,
314
harmonicdecline, 20, 229 heavy oil, 187, 287 heterogeneous reservoir, 209
samples, 78
Higgins-Leighton Method,
saturation, 113 secondary recovery of,
historical performance, 278
sweet, 132
281
gravity, 258, 287 gravity override, 195
swept volume, 160
153
history-matched simulation,
145
horizontal floods schemes, horizontalmiscible floods,
wet, 145 gas cap, 124, 145, 190 drive, 134, 140, 239 secondary, 145
horizontalwells, 205 critical rates for, 207 performance projection,
dissolved, 201 29
thin, 190 volume factor, 28, 105
Horner analysis,
volume factors,
gas-oil interface, 37 gas-water injection
HumbleMethod, 137 hydrocarbon
alternate, 202 gas-water interface, 37 gauge
characteristics, 131 miscible flooding, 171 pore volume maps, 39 hydrocarbons
water resistivities, 79, 103 forward combustion andwaterflooding combination of, 195 fractional flow, 155 fracture permeability, 63 fractured reservoirs,
63
fractures, 55 Free Trade Agreement, 286 freehold interest,
254
royalty, 7-8, 260 frontal displacement model, 191 frontieroil, 288 fully implicitmethod, 215
356
electronic, 86 mechanical, 86 general and administrativecosts, geochemical material balance
methods, 278 geologicanalogy, 278 geological mapping, 39 model, 177 period, 243
206
producibility, 206 uses of, 205 yield from, 205
gas in place, 113 gas-oilcontact, 38
122
22
162 172
horizontal sweep efficiency, 159 horizontal waterflood schemes, 156, 159
gas drive gas in place calculation of initial solution,
164
history match, 166 history matching, 219
solution, 28, 145 sour, 92, 145 151
federal, 281 " provincial,
remaining reserves, reservoir, 44
recovery, 147 reserves, 148 reservoirs, 123
forecasting gas de1iverability, models, 191
going concern value, 311 government agencies, 32 governments, 306
block sizing, 166 design, 217
57
marketers, 306, 307 new, 257 nonassociated, 29, 145 old, 257 permeabilities, 53 permeability, 53 prices, 255
53
266
Geological Surveyof Canada, 282 geophysicallog method, 65 glaze, 101
formation volume factor, 91, 94
distribution, 206 regimes, 72 tests, 36, 46, 70 fluid contacts, 48 expansion, 133 interface, 37 invasion, 247 properties, 216 saturation, 27 fluid expansion, gas pools, 247
resistivity index,
futures,
5 5
flow rate, 151
171
characteristics, 75
Forcheimer effect,
play, 243
initial volumes in place, unrecoverable volumes, future initial reserves, 5
88
in place, 28, 118 256
presence of,
14
hydrodynamic flow trapping, 38 hydrodynamics, 84 hyperbolicdecline, 226-229 Ideal Gas Law, 28, 91 imbibition, 167 immature depletion,
157
SUBJECT INDEX
immiscible carbon dioxide flooding, 200 gas injection, 183 Implicit Pressure Explicit Saturation Method, 215 in situ combustion process, in situ conditions, 63 income tax, 261
194
solution gas in place, 29 volumes in place, 5
resistivity, 202
wireline, 45 sandwich, 161 low productivityallowance,
257
mapping,
interface fluid, 37 gas-oil, 37 gas-water, 37 oil-water, 37, 39 interfacial tension, 69, 154 reduction in. 201 interference test, 77 internal rate of return, 264
International EnergyProgram, 286 inventory. 7 isnpach maps, 13, 38
J function, 69 Klinkenberg effect, 53 laboratory analysis of fluidproperties, 96 285
of Conservation of Mass,
69
loss
injection schemes dispersed gas, 183, 185 external, 183, 185 injectivity lack of, 190 intangible costs, 263 interest carried, 254 Crown, 254 expense, 256 freehold, 254 net profits, 255 overriding royalty, 254 production payment, 254 royalty, 254 working, 254
Latin Hypercube Method, Law D'Arcy's, 53 Ideal Gas, 28, 91
MercuryArchimedes Method, 57 Method Craig-Geffen-Morse, 164 Delphi, 274, 278 Dietz, 138, 140 Higgins-Leighton, 164 Humble, 137
57 density, 57 neutron, 57 resistivity, 63 log analysis, 57 logs, 46
petrophysical well, 16
injection alternate gas-water, rates, 189
maximum royalty, 257 mean, 11 mechanisms recovery, 188 median, 11 medium oil, 287
acoustic,
incremental economics, 310 industry databases, 57, 168, 237 initial reserves, 5
landcompensation,
lessor royalty. 7
light crude, 287 liquidpermeability, 53 liquids naturalgas, 22, 29 lithology, 242 log
277
38 computer, 42 geological, 39 volumetric, 35 maps hydrocarbon pore volume, 39 isopach, 13, 38 mechanically contoured, 39 porosity-thickness, 38 structure, 38 thermal gradient, 81 market demand forces, 297, 302 mechanisms, 300 value, 312 marketing options, 301 markets, 290 core, 297, 300 crude oil, 287 crude oil, 253 naturalgas, 297 naturalgas, 253 noncore, 300 noncore, 300 types of, 300 Marshal Method, 138 Marxand Langenheim model, 191 material halance, 35, 105, 126,136, 139, 140 assumptions, 121
determination of hydrocarbons in place, 120 equation, 17, 120, 122 estimates,
30 17, 149 materialbalance equation special cases of, 122
method,
149
of Conservation of Malter, 120 leasing, 282 lending, 314
material halance method, 121, 123 mathematical functions, 215 marrix permeability, 63 maturewaterflood, 158
Latin Hypercube, 277 Marshal, 138 Mercury Archimedes, 57 ModifiedHurst, 138 MonteCarlo, 11, 30, 212 Musknt, 137 Pirson, 137 Robertsand Ellis, 137 Schilthuis, 138 Tracy or Tamer, 137 VVarren, 107, 118 VVedge, 138 VVelge, 140 method. See procedure.
arithmetic, 8 caliper,
57
chemical, 65-66 dynamic implicit, 215 electrical, 65-66 finite-difference,
215
fully implicit, 215 gas expansion,
57
geophysical log, 65 historical performance, 278 implicit pressure explicit saturation, 215 material balance, 17, 121, 123, 149 reservoir simulation, 22 semi-implicit, 215 statistical, 9, 231-233 straight-line, 126, 127, 136 summation-of-fluids, 57 volumerric, 12, 148, 163, 175, 205, 209
methods analytical, 137 areal yield, 278 combined, 278 experimental, 173 geochemical materialbalance, 278 of achieving miscibility, 172 of analysis, 275 theoretical, 234-235 volumetric yield, 278 waterflood prediction, 165 micellar flooding, 154, 168 mineral rights
357
c DETERMINATION OFOILANDGASRESERVES
ownership of, 254, 282 subsurface, 254 mineral tax, 260 minimum royalty, 257 miscibility methods of achieving, 172 vapourizing multiple-contact, 173 miscible flooding carbon dioxide, 200 residual oil saturation after, 175 miscible floods areal sweep efficiency for horizontal, 175 field performanceof, 179 vertical sweep efficiency for horizontal, 176 miscibleprocess first-contact, 172 multiple-contact, 172 mobility ratio, 154, 175 mode, 11, 240 model Butler, 194 conical steam zone, 191 dimensions, 166 frontal displacement, 191 geological, 177 Marx andLangenheim, 191 Myhill and Stegeimeier, 193 phases, 166 sensitivityanalysis, 218 steam overlay, 191 uncertainty, 268 Vogel, 194 models !D, 217 2D areal, 217 2D radial, 217 2D vertical, 217 3D, 218 forecasting, 191 Modified Hurst Method, 138 Monte Carlo computer simulation, 30,107, 206, 212, 277 multi-well pools, 13 multiple-contact miscible process, 172 Muskat Method, 137 Myhill and Stegeimeiermodel, 193
National Energy Board, 281 natural depletion mechanisms, 131, 133 natural drive mechanisms, 238, 247 natural gas, 28 characteristics of, 145 liquids, 22, 29, 151, 255 markets, 253, 297 royalty, 257 naturally fracturedreservoirs, 167
358
fracture system, 214 matrix system, 214 near-miscible carbon dioxide flooding, 200 negotiatingtool, 305 net overridingroyalty, 254 pay, 44, 103, 112 effective, 102 present value, 264, 309 profits interest, 255 reservoir, 44 net-back calculation, 320 neutron log, 57 neutron-density cross-plot, 58 new gas, 257 new oil, 258 New York Mercantile Exchange, 292 nonassociatedgas, 29, 145 nonassociated gas reserves, 148 nonconventional crude, 288 North American Free Trade Agreement, 286 numerical simulation, 132, 137, 166 NYMEX, 292, 294
oil age, 258 equivalent, 9 formation volume factor, frontier, 288 heavy, 287 medium, 287 new, 258 old, 258 quality, 141 rate, 206 recovery factors, samples, 79 synthetic, 288 third tier, 258 viscosity,
239
141
oil in place calculationof, 28 oil recovery factors affecting, 140 primary, 237 oil sands deposits, 189 royalty, 260 oil saturation at the start of flood, 160 oil-base cores, 102 coring, 65 oil-steamratio, 193 oil-water interface, 37, 39 oil-wet, 74 old gas, 257 old oil, 258
96
Ontario Securities Commission. 315 open-hole drillstem test, 72 open-hole wireline tools, 86 operated-dayrate definition of, 223 operating costs, 256 Operator, 259 options, 295 call, 295 fence, 295 put, 295 over-pressured reservoirs,
148
overhead fee, 256 overriding royalty, 7-8 royalty interest, 254 ownership mineral rights, 254 of reserves. 7 oxidation of crude, 195 P-X diagram, 173 par price, 258 parameter uncertainty. 268 parameters reservoir, 16 pay defining net, 45 gross, 44 net, 44 payout period, 307 pentanes plus royalty, 258 performance prediction 184 analytical, 164 performance prediction, 183 permeability, 27, 53, 72 absolute, 216 effective, 53 fracture, 63 from core, 53 gas, 53 horizontal, 101 liquid, 53 low, 190 matrix, 63 relative, 53, 54, 216 specific, 53 permeability cutoffs, 46 permeability from cores, 101 permeameter. 54 petrophysicalwell logs, 16 phase behaviour, 148, 201 diagrams, 146 pipeline companies, 306 gas reserves, 150 major interprovincial systems, pipelines feeder, 288
288
SUBJECTINDEX
PirsonMethod, 137 political uncertainty, 266 polymer flooding, 154, 168 pool area, 12 discovery, 157 parameter,
112
size, 240 pooling, 255 pools analogous, 163 carbonate, 242 clastic, 242 multi-well, 13 single-well, 12 porevolume compressibility tests, 63 porosity, 27, 48, 55, 72, 113, 159 average, 216 correlation of log andcore, 62 cutoffs, 45 from cores, 101 from well logs, 103 low, 190 primary, 37, 55 secondary, 55 porosity-thickness maps, 38 porpoising, 210 possible developedreserves, 6 reserves, 5, 17, 179 undeveloped reserves, 6 post-injection startup, 158 post-waterflood response, 158 presence of hydrocarbons, 14 pressure, 75, 113 abandonment, 148 average reservoir, 89 bottom-hole, 49 buildup data, 86 buildup test, 77 capillary, 67 equilibrium, 121 gradients, 86 pseudo-critical, 91-92 recorders, 75, 86 reservoir, 86, 104 stabilized bottom-hole, 86 threshold, 67 transient analysis, 36 pressure-depth plots, 37,38 pressure-volume test, 79 price average Alberta market, 257 select, 257 wellhead, 255 pricedifferentials for light and heavy crude, 294 price risk, 294 prices crude oil, 255 gas, 255
outlook for world oil, 295 pricing, 291 primary depletion, 131 drive mechanisms, 238 oil recovery, 237 porosity, 37, 55 probabilistic procedure, 8, 11, 30 simulation, 277 probability, 11, 17 analysis, 166 degree of, 10 increase, 11 probable developed reserves, 5 reserves, 5, 17, 180 undeveloped reserves, 5 procedure. See method. deterministic, 8, 10-11, 30 probabilistic, 8, II, 30 stochastic, 11 process differential liberation, 97 displacement, 154 enriched air combustion, 195 flash liberation, 97 steam flood, 189 thermal wave, 195 variations, 195 processes in situ combustion, 194 thermal recovery, 187 processing fees, 255 producers, 306 production cumulative, 7 data, 216 forecasting, 132, 304 paymentinterest, 254 rate, 140 revenue, 255 royalty, 260 tests, 75, 104 production performance charts definition of, 223 production rates carbon dioxide flooding, 203 productivity allowance, 257 products related, 29 profitability indices, 307 properties reservoir rock, 101 rock and fluid, 27 proved reserves, 17, 180 provincial governments, 281 pseudo-critical pressure, 91 properties, 91-92 temperature, 91
put option, 295 PVT
analysis, 77 data, 120 samples, 79 pyrobitumen, 72, 103
quantitative estimation,
274
108 rate of return, 308 rate-cumulative graphs, 223 rate-time graphs, 223
ranges,
ratio
break-through, 177 mobility, 154, 175 oil-steam,
193
viscous-gravity,
176
ratio curves definition of, 223 ratioyardstick reserves-to-production, 305 recombination sampling, 77 recover,
189
recoverable gas, 148 recovery carbon dioxide flooding efficiency, 203 estimates, 135 factor, 17, 31, 168, 210 carbon dioxide flooding, 203 distributions, 249 for large pools, 238 statistics, 237 mechanisms, 188, 195 parameters, 247 secondary, 171 tertiary, 154, 171 refinement of volumetric estimates, 43 refineries, 290 refining Canadian, 290 regulations, 283,284 regulatory approvals changes to, 298 regulatory constraints, 143 regulatory environment, 253 related products, 22-23,29, 151 relative frequency curve, 30 relative permeability, 53, 54, 216 measurement of, 54 reliability of data, 31, 63, 101, 121, 136, 157, 237 reliability of results, 162, 163, 164, 166 remaining proveddevelopedreserves. 5 proved reserves, 5 provedundeveloped reserves. 5 reserves, 5 replacement costs, 317 reserve
359
• DETERMINATION OFOIL ANDGASRESERVES
distribution, 116 gas losses, 150 parameters, 211 reserves, 3, 253 associated gas, 150 company grossremaining, 7 companynet remaining, 7 confldencelevelo~ 9 determinations, steps involvedin, 211 developed, 6 developed nonproducing, 6 developed producing, 6 economically recoverable, forecasting, 132
148
ownership, 7 company gross remaining reserves, 7 pipelinegas, 150 possible,S, 17 possible developed, 6 possible undeveloped, 6 probable,S, 17 probabledeveloped, 5 probable undeveloped, 5 proved, 17 remaining, 5 remaining proved. 5 5
7
characteristics. continuity,
35
131
37
deep, 190 expansion terms,
360
22
data, 216
levels of, 282 royalty Crown, 7-8 crude oil, 258 factor, 259 freehold, 7-8, 260
101
gross overriding,
simulation data requirements, simulation method, 22 simulators, 214
216
retrograde gas condensate, residual gas saturation, 148
141
254
intent, 259 interest, 254 lessor, 7 maximum, 257 minimum, 257
single-phase gas, 146 temperature, 81, 104 two-phase, 146 undersaturated, 80 undersaturated oil, 122, 133 voidage terms, 121
natural gas, 257 net ovetriding, 254 oil sands,
260
overriding, 7-8 pentanes plus,
258
production, 260 rate, 259 152
resource,
260
sulphur, 258 tax deduction,
rugosity, sales,
259
49
7
samples oil, 79
146
PVT, 79 reservoir fluid, 104 surface, 79 surface crude oil, 79
161, 175
resistivities
water, 78 sampling, 77
formation water, 79, 103 resistivity logs, 63, 69 resource allowance, 261 assessments, 281
subsurface, 77 surface recombination,
78
sandwicheffect, 135 sandwichloss, 161, 207
properties, 312 royalty, 260
saturatedoil reservoir, saturation, 65 connate water, 65
uses of estimates, 31 resources, 3, 4, 253 discovered, 5 undiscovered, 5 results retrograde gas condensate reservoirs, return on investment, 308 revenue, 255 production, 255
right-of-entry, 285 risingbubbleapparatus, risk, 266
137
63
rock properties, 216 rotary drilling, 75 royalties
saturated oil, 123 shallow, 190 simulation, 137, 140
oil saturation,
8
rock compressibility,
174
123
formation fluid, 216 gas, 113 residual oil, 161,175 water, 69
reliability of, 162, 163, 164, 166 122
fluid samples, 104 fluids, 77 forecasts of performance, 220 geometry, 141, 216 gross, 44 heterogeneities,
aggregation,
dew-point, 146 dry gas, 145 fractured, 63 gas, 123 naturally fractured, 167, 214 over-pressured, 148
reserves-to-production ratio, 305 reservoir analogous, 22 area and volume,
Robertsand Ellis Method,
volume, 27 reservoirs analogous, 17 cycling of gas condensate.
remaining proved undeveloped, 5 risk-weighting of estimates, 8 role of, 304 solutiongas, 150 sulphur, 23 undeveloped, 6 users of volumes, 306 uses of estimates, 311 uses of evaluations, 253 working interest share of,
model initialization, 218 net, 44 parameters, 16, 247 performance charts definition of, 223 performance forecasting, 219
rock properties,
253
remaining proved developed,
risk premium, 264 risk-weighting of reserves estimates
poor, 190 pressure, 86, 104 quality, 238
from improved recovery projects, future initial, 5 gas, 148 grossremaining, 7 half-life, 314 initial, 5 nonassociated gas,
heterogeneity, 63, 74 limits, 39 modelgrid design, 217
146
saturations from cores, schemes
102
dispersed gas injection, external injection,
183, 185
183, 185
horizontalwaterllood, 156, 159 vertical waterllood, 156, 161, 163 SchilthuisMethod, 138
-I SUBJECT INDEX
screening and feasibility studies, screening guidelines, 188 secondary gas cap, 145 porosity, 55 recovery, 171 recovery of gas, 153 securities commissions, 8 securities reporting, 315 security, 305 seismic 3-D,
174
36
data, 35 select price, 257,258 semi-implicit method,
215
sensitivity cases, 8 separator tests, 79 shale, 190 content. 63 presenceof, 72 shalysand interpretation process, 73 shrinkage factor, 96 shut-in period, 104 simulation black oil, 177 history-matched, 22 Monte Carlo, 277 numerical, 137, 166 probabilistic, 277 reservoir, 137, 140 studies, 177 simulators blackoil, 214 compositional, 214 enhanced oil recovery. 214 reservoir, 214 types of, 178 single estimate, 11 single-phase gas reservoir, 146 single-value approach, 106 estimate, 107 single-well pools, 12
siterestoration and reclamation, 256 slimtube test, 174 slug size chasegas, 178 solvent, 178 slug sizing, 202 solution gas, 28, 145 drive, 133, 207, 238 reserves, 150 solvent slug size, 178 sour gas, 92, 145 sourceof data, 31, 81 sources of errors, 64, 123, 268 sparecapacity, 256 specific permeability, 53 stabilized bottom-hole pressure, 86 stand-off, 207
standardconditions, 30 Standing Committee on Reserves Definitions, 3, 10 startup
post-injection, 158 state steady, 54, 125 unsteady, 54, 125 statement of changes in cashposition, 263 of income, 263 statistical analysis, 107 statistical method, 9, 231-233 statistical techniques, 58 statistics recovery factor, 237 steady state, 54, 125 steady-state conditions, 206 steam distillation, 195 drive, 195 floodprocess, 189 overlaymodel, 191 steam confinement lack of, 190 steam stimulation followed by wet combustion, 195 Stiff diagram, 79 stochastic evaluation, 275 stochastic procedure, 11 stock tank cubicmetre, 96 straight-line method, 126, 127, 136 stratigraphic traps, 39 structure maps, 38 subjective estimate, 274 subsurface sampling, 77 successorrules, 263 sulphur, 23, 29, 151 royalty, 258 summation-of-fluids method, 57 sunkcosts, 309 superposition theorem, 125 supplemental calculations aquifers, 124 gas caps, 124 water influx, 124 supplyindicator, 305 surface crude oil samples, 79 surface recombination sampling, 78 surface samples, 79 surfactants, 189 sweep efficiency, 162 areal, 202 horizontal, 159 total, 159 vertical, 160,202 volumetric, 202 sweet gas, 145 swelling, 201 synthetic oil, 288
tangible costs, 263 tariffs, 285, 290 tax credit, 285 taxation levels, 285 technical assistance, 286 technical uncertainty, 266 techniques statistical, 58 temperature, 113 bottom-hole, 49, 81 pseudo-critical, 91 reservoir, 81, 104 temperature gradient, 82, 85 ternary diagram, 173 multi-contact, 174 tertiaryrecovery, 154, 171 test absolute open flow, 77 constant andvariable rate, 77 differential liberation, 79 drawdown, 77 drillstem, 75, 104 flash liberation, 79 interference, 77 pressure buildup, 77 pressure-volume, 79 production, 75, 104 separator, 79 slim tube, 174 vapourization, 79 testing back pressure, 152 theoretical methods, 234,235 thermal conductivity, 84 cracking, 195 efficiency term, 193 expansion, 195 gradient maps, 81 recovery processes, 187 wave process, 195 third tier oil, 258 threshold pressure, 67 tiltedoil-water contacts, 38 timestep sizing, 166 tolls, 285, 290 tool resolution, 63 topgas, 256-257 Tracy or TamerMethod, 137 transition zones, 48 trap,
37
trapping hydrodynamic flow, 38 traps stratigraphic, 39 two-phase reservoir, 146 type-curve matching, 230 uncertainties, 206, 212 estimation of, 178
361
2Q
DETERMINATION OF OILAND GASRESERVES
uncertainty, 10, 30, 253, 266 causes of, 268 degreeof, 266 economic, 266 engineering, 266 estimation of, 273 geological, 266 magnitude of, 271 model, 268 parameter, 268 political, 266 technical, 266 use of, 271 undersaturated oil reservoir, 80, 122, 133 undeveloped reserves, 6
undiscovered hydrocarbon volumes,
278
resources, 5 unitization, 255 unrecoverable volumes, 5 currently uneconomic volumes, 5 residual unrecoverable volumes, 5
gross swept, 160 volume-weighted average, 90 volumes currently uneconomic, 5 future unrecoverable, 5 residual unrecoverable, 5
Welge Method, well
abandonment costs, conditioning, 78
unrecoverable, 5 unswept, 161 volumes in place future initial, 5 initial, 5 volumetric analysis, 158 calculation, 28 eqnation, 108, 158, 175 estimate, 106 estimates, 27 mapping, 35 method, 12, 148, 163, 175, 205, 209 sweepefficiency, 177, 202 yield methods, 278 vugs, 55
unsteadystate, 54, 125 unswept volumes,
161
US Securities and Exchange Commission, 315
useage fees, 285 uses of resource estimates,
31
valuing oil and gas companies,
311
vapourization, 201 test, 79 vapourizing multiple-contact miscibility, 173 vertical miscible floods, 171 resolution, 64
sweep efficiencies, 167-168 sweep efficiency, 160 waterflood schemes, 156, 161, 163 viscosity. 79 oil, 141 reduction, 195, 20 I viscous-gravity ratio, 176 Vogel model, 194 volume drainage, 72
362
WarrenMethod, 107, 118 theory, 107 nse of, 118 water, 190 analyses, 79 channelling, 209 drive, 134, 137 influx, 121, 124 samples, 78 saturation, 48, 69 saturations from well logs, water-wet, 74 waterflood horizontal schemes, 156
mature, 158 prediction methods,
103
165
verticalschemes, 156, 163 waterflooding, 154 combination of forward combustion
and,
195 39
weightedaverage cost of capital, 264 weighted-mean, 240
256
data, 216 horizontal, 205 patterns, 188, 195-196 porosity from logs, 103 spacing, 142, 152, 196 test analysis,
70
testing, 48 tests, 222 water saturations from logs,
103
wellhead price, 255 wet combustion
steam stimulation followed by, 195 wet gas, 145 wettability, 69, 74, 102 wireline logs, wireline tools
45
open-hole, 86 working capital, 257 working interest, 254 working interest share, 7 Wylie Time-Average Equation, 58 zones transition, 48
reserves drainedfrom, 208
wedge zones,
138, 140
-I I
I
