BJ Frac Manual v1.0 June 2005(A4)

BJ Services’ Frac Manual Contents

BJ SERVICES COMPANY

HYDRAULIC FRACTURING MA N U A L

Version 1.0 (A4-Sized)

June 2005 Page i

Version 1.0 June 2005 Uncontrolled Copy – DO NOT DISTRIBUTE

Tony Martin Region Engineer Singapore

BJ Services’ Frac Manual Contents

Contents Contents .....................................................................................................................................ii List of Figures ............................................................................................................................ v 1.

Introduction ................................................................................................................... 1

2.

Basics of Hydraulic Fracturing...................................................................................... 4 2.1 2.2 2.3 2.4 2.5

3.

Types of Fracturing..................................................................................................... 12 3.1 3.2 3.3 3.4 3.5 3.6

4.

Stress........................................................................................................................................53 Strain ........................................................................................................................................53 Young’s Modulus.......................................................................................................................54 Poisson’s Ratio .........................................................................................................................55 Other Rock Mechanical Properties ............................................................................................56 In-Situ Stresses.........................................................................................................................58 Stresses Around a Wellbore......................................................................................................59 Fracture Orientation ..................................................................................................................60 Breakdown Pressure and Frac Gradient....................................................................................61 Rock Mechanical Properties from Wireline Logs........................................................................63

2-D Fracture Models ................................................................................................... 68 8.1 8.2 8.3

Page ii

Proppant Pack Permeability and Fracture Conductivity .............................................................45 Proppant Selection....................................................................................................................48 BJ Services FlexSand and LiteProp .........................................................................................50

Rock Mechanics ......................................................................................................... 53 7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8 7.9 7.10

8.

Water-Based Linear Systems....................................................................................................29 Water-Based Crosslinked Systems ...........................................................................................30 Oil-Based Systems....................................................................................................................33 Emulsions .................................................................................................................................35 Energised Fracturing Fluids.......................................................................................................35 Visco-Elastic Surfactant Fluids ..................................................................................................36 Additives ...................................................................................................................................40

Proppants ................................................................................................................... 45 6.1 6.2 6.3

7.

Fundamental Fluid Properties ...................................................................................................19 Shear Stress and Shear Rate....................................................................................................19 Types of Fluid ...........................................................................................................................20 Measuring Viscosity ..................................................................................................................23 Apparent Viscosity ....................................................................................................................25 Flow Regimes and Reynold’s Number.......................................................................................26 Friction Pressure .......................................................................................................................27

Fluid Systems ............................................................................................................. 29 5.1 5.2 5.3 5.4 5.5 5.6 5.7

6.

Low Permeability Fracturing ......................................................................................................12 High Permeability Fracturing .....................................................................................................12 Frac and Pack Treatments ........................................................................................................13 Skin Bypass Fracturing .............................................................................................................15 Coal Bed Methane Fracturing....................................................................................................16 Fracturing Through Coiled Tubing .............................................................................................16

Fluid Mechanics.......................................................................................................... 19 4.1 4.2 4.3 4.4 4.5 4.6 4.7

5.

The Basic Process ......................................................................................................................4 Pressure .....................................................................................................................................5 Basic Fracture Characteristics.....................................................................................................6 Fluid Leakoff ...............................................................................................................................8 Near Wellbore Damage and Skin Factor ....................................................................................9

Radial or Penny-Shaped ...........................................................................................................68 Kristianovich and Zheltov - Daneshy (KZD) ...............................................................................69 Perkins and Kern – Nordgren (PKN)..........................................................................................70


BJ Services’ Frac Manual Contents 9.

Fracture Mechanics .................................................................................................... 72 9.1 9.2 9.3

10.

Advanced Concepts.................................................................................................... 80 10.1 10.2 10.3 10.4 10.5 10.6 10.7 10.8 10.9

11.

Planning and Execution...........................................................................................................121 Anatomy of a Minifrac..............................................................................................................124 Decline Curve Analysis ...........................................................................................................125 Pressure Matching ..................................................................................................................131 Near Wellbore Effects and Multiple Fractures..........................................................................132 Minifrac Example 1 - 2D Minifrac Analysis...............................................................................134 Minifrac Example 2 - 3D Pressure Matching with FracProPT...................................................139 Minifrac Example 3 – Problems with Tortuosity .......................................................................147 Minifrac Example 4 – Perforation Problems.............................................................................153

Designing the Treatment .......................................................................................... 164 17.1 17.2 17.3 17.4 17.5 17.6 17.7

Page iii

The Step Up Test ....................................................................................................................115 The Step Down Test ...............................................................................................................116 Step Rate Test Example – Step Up/Step Down Test ...............................................................117

The Minifrac .............................................................................................................. 121 16.1 16.2 16.3 16.4 16.5 16.6 16.7 16.8 16.9

17.

Controlling Fracture Initiation...................................................................................................109 Controlling Tortuosity ..............................................................................................................111 Perforating for Skin Bypass Fracturing ....................................................................................112

The Step Rate Test .................................................................................................. 115 15.1 15.2 15.3

16.

Economic Justification for Fracturing .......................................................................................101 Completion Limitations ............................................................................................................104 Things to Look For ..................................................................................................................106

Perforating for Fracturing.......................................................................................... 109 14.1 14.2 14.3

15.

Steady State Production Increase .............................................................................................95 Pseudo-Steady State Production Increase ................................................................................96 Nodal Analysis ..........................................................................................................................99

Candidate Selection.................................................................................................. 101 13.1 13.2 13.3

14.

RES’s FracPro and Pinnacle Technology’s FracproPT..............................................................91 Meyers & Associates’ MFrac .....................................................................................................92 Other Simulators .......................................................................................................................93

Predicting Production Increase................................................................................... 95 12.1 12.2 12.3

13.

Tortuosity ..................................................................................................................................80 Nolte Analysis ...........................................................................................................................82 Dimensionless Fracture Conductivity.........................................................................................82 Tip Screen Out ..........................................................................................................................83 Multiple Fractures and Limited Entry .........................................................................................84 Proppant Convection and Settling .............................................................................................85 Proppant Flowback ...................................................................................................................86 Forced Closure..........................................................................................................................88 Non-Darcy Flow ........................................................................................................................88

3-D Fracture Simulators ............................................................................................. 91 11.1 11.2 11.3

12.

LEFM and Fracture Toughness .................................................................................................72 Non-Linear and Non-Elastic Effects...........................................................................................75 The Energy Balance..................................................................................................................77

General ...................................................................................................................................164 Designing for Skin Bypass.......................................................................................................165 Designing for Tip Screen Out ..................................................................................................166 Designing for Frac and Pack ...................................................................................................167 Designing for Tight Formations ...............................................................................................168 Designing for Injection Wells ...................................................................................................170 Designing for CBM Treatments ...............................................................................................170


BJ Services’ Frac Manual Contents 17.8 17.9 17.10

18.

Real-Time Monitoring and On-Site Re-Design......................................................... 176 18.1 18.2 18.3

19.

Horsepower Requirements......................................................................................................209 Flow Lines...............................................................................................................................210 High Pressure Pumps .............................................................................................................211 Intensifiers...............................................................................................................................214 Blenders, Gel Hydration and Liquid Additives..........................................................................216 Proppant Storage and Handling ..............................................................................................218 Treatment Monitoring ..............................................................................................................220 Wellhead Isolation Tool ...........................................................................................................221 The Frac Spread – How it Fits Together. .................................................................................224

Designing Wells for Fracturing ................................................................................. 228 21.1 21.2 21.3

22.

Pressure Matching ..................................................................................................................186 Well Testing for Fracture Evaluation........................................................................................193 Other Diagnostic Techniques .................................................................................................205

Equipment................................................................................................................. 209 20.1 20.2 20.3 20.4 20.5 20.6 20.7 20.8 20.9

21.

Real-Time Data Gathering.......................................................................................................176 On-Site Redesign....................................................................................................................181 Real-Time Fracture Modeling ..................................................................................................183

Post Treatment Evaluation ....................................................................................... 186 19.1 19.2 19.3

20.

Designing for Coiled Tubing Fracturing ...................................................................................172 Unified Fracture Theory and Proppant Number .......................................................................173 Net Present Value Analysis .....................................................................................................174

How Many Wells do I Need to Drill? ........................................................................................228 The Best Wells are the Best Candidates for Fracturing ...........................................................229 Designing Wells for Fracturing ................................................................................................229

The Fracture Treatment: From Start to Finish.......................................................... 232 22.1 22.2

Frac Job Flow Chart ................................................................................................................232 Example Treatment Schedules ...............................................................................................238

Nomenclature ........................................................................................................................ 241 Index

Page iv

.................................................................................................................................. 245


BJ Services’ Frac Manual List of Figures

List of Figures Section 2 2.1a 2.3a 2.5a

Typical hydraulic fracture treatment job plot. Diagram showing fracture half Length xf, fracture height H, and fracture width W. Illustration of the reduction in permeability around the wellbore.

Section 3 3.3a 3.3b

3.4a

Diagram illustrating the components of the frac-pack completion. Diagram illustrating two of the four positions in which a standard gravel pack or frac pack tool can be set. The left hand side shows the squeeze position, in which fluids flow down the tubing, through the crossover, out into the annulus below the GP packer and into the formation. The right hand side shows the lower circulating position. Fluid flows down to the perforations, as for the squeeze position. However, because the setting tool has been shifted upwards, the fluid can flow either into the formation, or back through the screens, up the washpipe (inside the screens) through the crossover, and out into the annulus above the tubing (shown in blue). By closing the annulus at surface, the fluid can be squeezed into the formation, whilst maintaining a dead string on the annulus, to monitor BHP. Diagram illustrating how the skin bypass fracture penetrates the skin to allow undamaged communication between the reservoir and the wellbore.

Section 4 4.2a 4.3a 4.3b 4.3c 4.4a 4.4b 4.4c 4.4d 4.5a 4.6a

Graph illustrating Newton’s law of fluids Relationship between shear rate and shear stress for a Bingham plastic fluid. Relationship between shear rate and shear stress for a power law fluid. Note that the graph shows the relationship in its most common form. However, in certain fluids the line can also curve upwards. Power law fluid log-log plot. Chandler 35 viscometer. The position of the rotor is indicated (A), whilst the bob is hidden inside this. The cup (B) holds the test fluid, and is mounted on a support (C) that can move up and down as required. Cross-section through the rotor and bob on a model 35 viscometer. Schematic diagram showing the model 35 viscometer bob assembly. Fann 50 high pressure, high temperature rheometer. This model is fully computer controlled, whereas earlier models had manual controls and were twice the size of the model shown. Graph illustrating the change in apparent viscosity for a power law fluid at two different shear rates. Diagram illustrating the three flow regimes.

Section 5 5.1a 5.2a 5.2b 5.2c 5.3a 5.6a

Page v

Hydration of polymer gels in water. A shows a polymer molecule before hydration in water, whilst B shows a polymer molecule after hydration in water. A crosslinked polymer. A shows the hydrated polymer prior to addition of the crosslinker. B shows the crosslink chemical bonds between the polymer molecules. pH ranges for crosslinkers (after SPE 37359). Temperature range for crosslinkers (after SPE 37359). Aluminium phosphate association polymer. Proppant transport as a function of foam quality. This graph is a combination of the work performed by several individuals and organisations. It is intended as a qualitative illustration of the effect foam quality has on the ability of the fracturing foam to transport and suspend proppant.



Section 6 6.1a 6.1b 6.1c 6.1d

6.3a

The effect of uniform and natural grain size distribution on porosity. Diagram illustrating how larger grains have larger pore spaces and hence greater permeability. Diagram illustrating the difference between a proppant with good sphericity and roundness (left), and a proppant with poor sphericity and roundness (right). Three SEM micrographs showing the effects of frac fluid residue. The micrograph on the left shows undamaged proppant before the addition of the frac fluid. The center micrograph shows the residue left by a poorly designed crosslinked system. The final micrograph shows the same proppant pack after an enzyme breaker has been used. SEM micrograph of FlexSand grain clearly showing the indentations caused by the closure of the surrounding proppant grains.

Section 7 7.1a 7.2a 7.4a 7.5a 7.5b 7.7a 7.8a

A block of material subjected to a force F. Strain produced by the application of force F. Application of force F also produces a deformation in the y direction. Force F applied to produce a shear stress. Volume changes from V1 to V2 as pressure increases from P1 to P2. Three dimensional stresses around a wellbore. Changes in stress regime due to erosion.

Section 8 8.1a 8.2a 8.3a

Propagation of a radial or penny-shaped fracture. Schematic showing the general shape of the KZD fracture. The Perkins and Kern - Nordgren fracture.

Section 9 9.1a 9.1b 9.1c 9.2a 9.2b 9.2c 9.3a

The Griffith crack. Failure modes in Linear Elastic Fracture Mechanics. Coordinate system for stress intensity factor. The Cleary et al approach. Crack tip diameter and the plastic zone. Note that rp is the radius of the plastic zone. The shape of the plastic zone, for a Poisson’s ratio of 0.25. Sources of Energy Gains and Losses for the fracturing fluid. Energy Gains + Energy Losses = 0.

Section 10 10.1a. Diagram illustrating the effects of horizontal stress contrast on tortuosity (after GRIAST 1996). 10.2a The Nolte plot. 10.4a The Tip Screen Out. 10.6a Proppant convection. As the heavier slurry enters the fracture it sinks and displaces the lighter slurry upwards. 10.7a Illustration of the “Pipelining” effect.

Section 12 12.2a 12.2b 12.2c 12.3a

Page vi

Transient production. The red lines illustrate the variation of pressure with distance from the wellbore, as time increases. The radius of the disturbed formation is continually increasing. Pseudo-steady state production. The radius of the disturbed formation has reached the reservoir boundary, re, and now the reservoir pressure is decreasing. The McGuire-Sikora Curves. Nodal analysis IPR curves for a gas well with a fracture of varying propped fracture width.



Section 13 13.3a

The effect of skin factor upon production rate. Note that this Figure is based purely on skin factor effects. No fracture stimulation is included.

Section 14 14.1a 14.1b 14.2a 14.2b 14.3a 14.3b

The Effect of perforations on fracture initiation. Perforating for zonal coverage. Perforation strategy for vertical wells. Perforation strategy for horizontal wells. The Effect of fracture initiation point on skin bypass fracs. Multiple skin bypass fracs over a long interval.

Section 15 15.1a 15.2a 15.3a 15.3b 15.3c

The step up test. The step down test. Step up pressure-rate crossplot using the example data. This plot shows the fracture extension pressure to be at +/- 6570 psi. Step down pressure-rate crossplot for the example data. The convex shape of the curve indicates near wellbore friction dominated by tortuosity. Step down pressure-rate crossplot for the example data, using surface treating pressure (STP). This graph illustrates the danger of using STP for step rate test analysis, as in this case, the near wellbore friction would have been incorrectly diagnosed as being perforation dominated.

Section 16 16.2a 16.2b 16.3a 16.3b 16.3c 16.3d 16.4e 16.4f 16.6a 16.6b 16.6c 16.6d 16.6e 16.7a 16.7b 16.7c 16.7d

16.7e 16.7f 16.7g 16.7h

Page vii

Typical minifrac job plot, showing BHTP, STP and rate. Expanded plot showing BHTP. Typical minifrac pressure decline curve. Use of a square root time plot to determine closure pressure. Typical minifrac pressure decline Horner plot. Graph showing the variation of g(∆tD) with ∆tD. Typical Nolte G time pressure decline plot. Example derivative plot based on a Horner Plot. Minifrac example 1 job plot. BH gauge pressure decline against elapsed time. Possible closure pressure at +/2770 psi (where the two red lines cross, marking a change in gradient). Note the sudden drop of about 50 psi as the pumps shut down at t = +/- 13 mins. BH gauge pressure decline against the square root of elapsed time. Possible closure pressure at +/- 2790 psi (where the two red lines cross, marking a change from straight line to curve). G function plot. The “true” ISIP is at +/- 3150 psi, whilst the closure pressure appears to be at +/- 2780 psi (where the two red lines cross). This gives a Gc of 1.30. Horner plot. The results from this plot are ambiguous and do not help in the analysis. Minifrac example 2 step rate test job plot. Step rate test crossplot for minifrac example 2, step rate test, showing fracture extension at +/- 8700 psi. Minifrac example 2 job plot. Comparison between gauge and calculated BHTP for minifrac example 2. Note that whilst the calculated BHTP follows the same general trend as the gauge BHTP, the actual value is quite different. Short term variations in the trend of the calculated BHTP are caused by the variations in rate. The general offset of the data is probably caused by incorrect input data in the fracture monitoring package (in this case FracRT). Minifrac example 2 pressure decline with derivative. Minifrac example 2 pressure decline square root time plot, with derivative. Initial pressure match for minifrac example 2. Interim pressure match after the stresses have had a first approximate adjustment. In this case, the stress gradient for the sandstone was increased from 0.62 to 0.68 psi/ft, Version 1.0 June 2005 Uncontrolled Copy – DO NOT DISTRIBUTE

BJ Services’ Frac Manual List of Figures and then 1300 psi was added to each stress. Note that the pressures are on a larger vertical scale than in Figure 16.7g. 16.7i Minifrac example 2 final pressure match. 16.7j FracProPT estimated fracture dimensions for minifrac example 2. 16.8a Minifrac example 3 treatment plot. 16.8b Minifrac example 3, detail of post-treatment pressure decline. 16.8c Minifrac example 3, square root time pressure decline plot. 16.8d Horner plot for minifrac example 3. Note that several lines may be fitted to the final slope on the LHS of this plot. In fact, the reservoir pressure is substantially lower than that indicated on the plot (as the well is produced by ESP’s), so all of these lines may be unreliable. 16.8e G Function plot for minifrac example 3. Note the true ISIP of +/- 2730 psi, and the closure pressure of +/- 2320. These values are in agreement with the value obtained from other plots, such as the pressure decline and the square root time plots. 16.8f MFrac output showing the initial pressure match before any adjustments were made. There is very little agreement between the predicted and actual BHTP’s. 16.8g Final MFrac output, after the model has been adjusted. 16.9a Job plot for Minifrac Example 4, Step Rate Test 1 16.9b Step up crossplot for Step Rate Test 1. Fracture extension seems to be at approximately 9100 psi. 16.9c Step down crossplot. Note the concave shape of the best fit curve, indicating that the near wellbore friction is dominated by the perforations. 16.9d Minifrac Example 4 job plot. 16.9e Detail of job plot showing bottom hole proppant concentration, gauge BHTP and slurry rate, as the proppant slug enters the formation. Note the +/- 400 psi rise in pressure. 16.9f Minifrac pressure decline, showing +/- 650 psi near wellbore friction and a closure pressure of +/- 8350 psi. 16.9g Square root of time plot for the minifrac pressure decline. This gives a slightly lower closure pressure than Figure 16.9f, at +/- 8230 psi. 16.9h Job plot for second step rate test. 16.9i Step down crossplot for the second step rate test. 16.9j Minifrac Example 4 BHTP plot before pressure matching. 16.9k Minifrac Example 4 pressure match using MFrac. 16.9l Job plot for the main treatment for Minifrac Example 4. Note the proppant concentration is measured at the surface. 16.9m Detail of the main treatment for Minifrac Example 4, showing the formation’s response to the proppant slugs. Proppant concentration is bottom hole.

Section 17 17.4a

17.9a

The diagram on the LHS illustrates the position of the slurry and the ‘pack’ at screenout – with the top of the ‘packed’ proppant at the top of perforations, and the annular space between the completion and the wellbore full of slurry, up until the crossover ports. The RHS shows the position of the pack after all the proppant has been allowed to settle. Optimum dimensionless fracture conductivity against dimensionless proppant number (after Economides et al, 2002).

Section 18 18.1a 18.1b 18.1c 18.2a

Process loop for real-time fracture modeling and redesign. Inside of a typical frac control van, showing the numerical display and some of the displays being run by JobMaster. Remote data transmission schematic. On-site redesign process flowchart.

Section 19 19.1a 19.2a

Pressure matching. The variables in the simulator are adjusted to make the calculated net pressure match the actual net pressure. Anatomy of a drawdown / build-up well test (after Agarwal, 1980)

Page viii


BJ Services’ Frac Manual List of Figures 19.2b 19.2c

19.2d 19.2e 19.2f 19.2g 19.2h 19.2i 19.2j 19.2k 19.2l 19.3a 19.3b

Graphs illustrating the deviation from transient flow caused by a reservoir boundary (i.e. pseudo-steady state flow). Constant rate drawdown semi-log plot. The straight line section can be used to evaluate the permeability. The deviation from the straight line at late time, is due to boundary effects of the reservoir, as the transient flow changes to pseudo-steady state flow. Example Horner plot, showing the extrapolation of the straight line portion to obtain the reservoir pressure, Pi. Once again, deviation from the straight line is caused by a change from transient flow to pseudo-steady state flow. Log-log diagnostic plot with derivative for the pressure build-up of an infinite-acting reservoir (i.e. no boundaries and no pseudo-steady state flow). Log-log diagnostic plot with derivative for the pressure build-up of reservoir with a partial boundary (e.g. a sealing fault). Log-log diagnostic plot with derivative for the pressure build-up of an infinite conductivity fracture. Log-log diagnostic plot with derivative for the pressure build-up of a finite conductivity fracture. Type curves for a single well in an infinite reservoir, with wellbore storage and skin damage (after Agarwal, Al-Hussainy and Ramey, 1970). Example of a log-log plot of ∆t against ∆P, used for type curve matching. Post-treatment log-log plot of well test data for a gas well. Type curves for a well with a finite conductivity, vertical fracture (after Agarwal et al, 1979 and Economides et al, 1987). The principle of tiltmeter fracture diagnostics (after Cipolla and Wright, 2000). Generic temperature log illustrating that the treating fluid has entered only a small portion of the perforated interval. The fracture will have initiated in the smaller interval. However, this does not necessarily mean that this is the center of the fracture.

Section 20 20.1a 20.2a 20.3a 20.3b 20.3c 20.3d 20.3e 20.3f 20.3g 20.4a 20.4b 20.4c

20.4d 20.5a 20.5b 20.5c 20.5d 20.5e

Page ix

Typical pump curves. This set is for a 30-16-6 frac skid, with a 16V92TA engine, a CLBT8962 transmission and a pacemaker pump with a 4.5 inch fluid end. Nominal rating of the pump skid is 700 HHP. Chart showing fluid velocity against fluid rate for various nominal diameters of Figure 1502 high pressure iron. Schematic diagram of a generic frac pump. Generic frac pump, suction stroke. Generic frac pump, discharge stroke. Skid mounted 16V 92T pump unit (700 HHP). Skid splits into two parts. Two views of a trailer-mounted Gorilla pump unit (2700 HHP). Body-load Kodiak pump unit (2200 HHP). Skid-mounted 1300 HHP pump unit. Schematic diagram of a generic intensifier. Schematic diagram of the intensifier hook-up. Intensifier worksite. Each intensifier (A) is hooked up to three frac pumpers (B), which are pumping the power fluid. Power fluid is handled by the power fluid unit (C). Intensifiers are rigged into a manifold (D). Note that whilst there are three intensifiers and 9 power fluid pumpers on location, there are also an additional two frac pumpers (E) rigged up to the downhole line to provide extra horsepower. Detail of an intensifier. In the foreground, on the RHS, is the downhole fluid end. In the background, on the LHS, is the power end, complete with high pressure iron rigging it to the frac pumpers. Generic flow diagram for a frac blender. Note that on a blender fitted with a Condor tub (such as BJ’s Cyclone blenders), the functions of the blender tub and the discharge pump are combined. 125D Frac blender, capable of 125 bpm and 35,000 lbs/min proppant rate. Body-load mounted Cyclone II blender, capable of 25 bpm. Skid mounted Cyclone blender. LFC hydration unit.


BJ Services’ Frac Manual List of Figures 20.6a

Frac sand being delivered from a Sand King to the hopper of a blender. Note that there are two blenders in this picture – one is on standby as a backup in case of equipment failure. 20.6b Vertically mounted, gravity feed proppant bins. 20.6c Trailer mounted sand dumper. 20.6d BJ Services Sand King. 20.6e Sand belt conveyor. 20.7a External view of BJ’s Stimulation Van 1800. 20.7b External view of a treatment monitoring container. 20.7c Two internal views of a treatment monitoring van. 20.8a Generic wellhead isolation tool rigged up to wellhead. The WIT is connected to the wellhead via the wellhead’s top flange. At this point the wellhead master valve and sub master valves are closed, maintaining control of the well and allowing the frac lines and WIT to be pressure tested. 20.8b+c Once the WIT has been connected to the wellhead and pressure tested (Fig 20.8a), the next stage is to close the valves of the frac lines (not shown – note that some WIT’s have their own master valves) and open the master and sub master valves on the wellhead. One the wellhead is open, the stinger is stroked down into the top of the tubing by pumping hydraulic fluid into the master cylinder. 20.8d Wellhead isolation tool rigged up on location. Note the two 3” frac lines connected to either side, plus the remote actuated 4” plug valve. 20.9a Schematic diagram of a frac spread. 20.9b Large scale treatment, carried out on several low permeability zones simultaneously. Note the number of Sand Kings and frac tanks on location, as well as the use of two blenders (one for backup in case of equipment failure). This frac spread features a separate mobile field lab (bottom left) and a third blender, just for gelling up the tanks and for pumping fluid from the tanks that are located a significant distance from the blender (located just above the bottom left hand row of frac tanks). 20.9c The MV Blue Ray, a Gulf of Mexico frac boat, designed primarily for high permeability, frac and pack treatments. 20.9d Skin Bypass Frac spread, using the “batch” frac method. The two frac pumps are positioned opposite each other, just below the wireline mast (the small read and yellow derrick). A third pump (with “BJ” painted on its roof) is being used as an annulus pump. The two vertical stainless steel tanks on the RHS are for fluid storage. The two batch mixers (each with two round batch tanks - the blue batch mixer is 2 x 50 bbls, whilst the red one is 2 x 40 bbls), used to batch mix the proppant into the gel, are located at the bottom of the picture. 20.9e Coiled tubing frac spread. The wellhead is positioned directly below the CT injector (center of picture), with the reel on the RHS. On the LHS are two nitrogen tankers. The main part of the frac spread is positioned behind the injector, with the sand dump truck being the most prominent feature. 20.9f The MV Thanh Long. This was a boat put together for a single fracturing treatment, for a customer operating offshore Vietnam. The aft deck holds the following equipment:- 4 x 1200 HHP frac pumps, Cyclone II blender, 2 x 640 cu ft proppant bins, treatment monitoring container c/w field lab, 4 x 165 bbls tanks and a 100 bbl vertical tank.

Section 22 22.1a

Page x

Frac job process flow diagram.


BJ Services’ Frac Manual 1. Introduction

1.

Introduction

History The first attempts at fracturing formations were not hydraulic in nature – they involved the use of high explosives to break the formation apart and provide “flow channels” from the reservoir to the wellbore. There are records indicating that this took place as early as 1890. Indeed, one of the predecessor companies of BJ Services, the Independent Torpedo Company (founded in 1905), used nitroglycerine to explosively stimulate formations in Ohio. This type of reservoir stimulation reached its ultimate conclusion with the experimental use of nuclear devices to fracture relatively shallow, low permeability formations in the late 1950’s and early 1960’s. In the late 1930’s, acidising had become an accepted well development technique. Several practitioners observed that above a certain “breakdown” pressure, injectivity would increase dramatically. It is probable that many of these early acid treatments were in fact acid fractures. In 1940, Torrey recognized the pressureinduced fracturing of formations for what it was. His observations were based on squeeze cementing operations. He presented data to show that the pressures generated during these operations could part the rocks along bedding planes or other lines of “sedimentary weakness”. Similar observations were made for water injection wells by Yuster and Calhoun in 1945. The first intentional hydraulic fracturing process for stimulation was performed in the Hugoton gas field in western Kansas, in 1947. The Klepper No 1 well was completed with 4 gas producing limestone intervals, one of which had been previously treated with acid. Four separate treatments were pumped, one for each zone, with a primitive packer being employed for isolation. The fluid used for the treatment was war-surplus napalm, surely an extremely hazardous operation. However, 3000 gals of fluid were pumped into each formation. Although post treatment tests showed that the gas injectivity of some zones had been increased relative to others, the overall deliverability from the well was not increased. It was therefore concluded that fracturing would not replace acidising for limestone formations. However, by the mid-1960’s, propped hydraulic fracturing had replaced acidising as the preferred stimulation method in the Hugoton field. Early treatments were pumped at 1 to 2 bpm with sand concentrations of 1 to 2 ppa. Today, thousands of these treatments are pumped every year, ranging from small skin bypass fracs at $20,000, to massive fracturing treatments that end up costing well over $1 million. Many fields only produce because of the hydraulic fracturing process. In spite of this, many industry practitioners remain ignorant of the processes involved and of what can be achieved.

Page 1



The Process Hydraulic fracturing occurs as a result of the phenomenon described by Darcy’s law for radial flow:kh∆P ...................................................................... (1.1) q = µ ln(re/rw) Where q is the flow rate, k the formation permeability, h the net height, ∆P the pressure differential (or drawdown), µ the fluid viscosity, re the drainage radius and rw the wellbore radius. This Equation describes the flow rate for a given reservoir-wellbore configuration, for an applied pressure differential. Re-arranging this Equation gives a different emphasis: ∆P

=

q µ ln(re/rw) ................................................................... (1.2) kh

This Equation describes the pressure differential produced by a given flow rate. Remembering that Darcy’s Equation applies equally to injection and to production, Equation 1.2 tells us the pressure differential needed to pump a fluid of viscosity µ into a given formation at a given rate q. As the flow rate increases, the pressure differential also increases. Pressure and stress are essentially the same thing (see Section 2.2), so that as the fluid flow generates a pressure differential, it also creates a stress in the formation. As flow rate (or viscosity) increases, so does the stress. If we are able to keep increasing the rate, eventually a point will be reached were the stress becomes greater than maximum stress that can be sustained by the formation – and the rock physically splits apart. This is how we frac, by pumping a fluid into a formation at high rate and – consequently – high pressure. However, it is important to remember that it is pressure – not rate – that creates fractures (although we often use rate to create the pressure). Pressure – and stress – is stored energy, or more accurately stored energy per unit volume. Energy is what hydraulic fracturing is all about. In order to create and propagate a fracture to useful proportions, we have to transfer energy to the formation. Producing width and physically tearing the rock apart both require energy. Overcoming the often highly viscous frac fluid’s resistance to being pumped also takes energy. So the key to understanding the hydraulic fracturing process is to understand the sources of energy gain, such as the frac pumps and the well’s hydrostatic head, and the sources of energy loss and use. The sum of these is always equal to zero. As pressure is energy, a great deal can be learned about a formation by studying the pressures produced by a treatment. The product of the pressure and the flow rate gives us the rate at which energy is being used, i.e. work. This is usually expressed as hydraulic horsepower. The analysis of the behaviour of fracturing pressures is probably the most complex aspect of the process that most Frac Engineers will become involved in. Once a fracture has been created, proppant is placed inside it. If the treatment has been designed effectively and pumped without any problems, then this proppant should form a highly conductive path from the reservoir to the wellbore. This is what makes the well produce more.

Page 2



Using this Manual This manual is not intended as an all-inclusive work on the science of hydraulic fracturing. Instead, it is intended to be a practical introduction to the science and art involved in these processes. It is intended to be used by junior Engineers who wish to gain some knowledge of the fracturing process, and by experienced Engineers who wish to gain a deeper insight into specific areas. This manual has been written with the intent that anyone with a technical background can come to understand fracturing. Readers are invited to consult the references at the end of each section for more detailed information on any specific subject. The author of this manual welcomes any comments that the reader may have – whether it is about something which is unclear, an omission or something that is just simply incorrect. I welcome any constructive comments that the reader may have. Throughout this manual, the author has used United Kingdom English, rather than American English. Consequently, some readers may find the occasional word that seems to be spelled in a manner somewhat different from that which they are used to. Examples include programme (instead of program), acidise (instead of acidize), grey (instead of gray), aluminium (instead of aluminum) and sulphate (instead of sulfate). The author makes no apologies for this.

Acknowledgements This manual has taken five years to complete, on and off (two to write and three to get proof read.....). Over this period, I have received assistance from a number of persons who deserve my thanks. Todd Gilmore, for continually reviewing each section as it was written; Antonio Moreira for correcting the mistakes and omissions in the equipment section; Phil Rae for his continuing help, support and encouragement; and finally Dave Cramer, Ron Matson, Harold Hudson and Kieran O’Driscoll, for the vital but tedious and time consuming process of proof reading. Thanks to you all. Tony Martin, Singapore, June 2005.

References Torrey, P.D.: “Progress in Squeeze Cementing Applications and Technique”, Oil Weekly, July 29, 1940. Yuster, S.T. and Calhoun, J.C., Jr.: “Pressure Parting of Formations in Water Flood Operations – Part I”, Oil Weekly, March 12, 1945. Yuster, S.T. and Calhoun, J.C., Jr.: “Pressure Parting of Formations in Water Flood Operations – Part II”, Oil Weekly, March 19, 1945. Farris, R.F. : “Hydraulic fracturing, a method for increasing well productivity by fracturing the producing formation and thus increasing the well drainage area”, US Patent reissued Nov 10, 1953. Re. 23733. Howard, G.C., and Fast, C.R.: Hydraulic Fracturing, Monograph Series Vol 2, SPE, Dallas, Texas, USA (1970).

Page 3


BJ Services’ Frac Manual 2. Basics of Hydraulic Fracturing

2.

Basics of Hydraulic Fracturing

Hydraulic fracturing is the process of providing a conductive path from the reservoir to the wellbore. How this is achieved depends upon the objectives, the reservoir and the well.

2.1

The Basic Process

As fluid is pumped into a permeable formation, a pressure differential is generated that is proportional to the permeability of the formation, kf . As the rate increases, this pressure differential between the wellbore pressure and the original reservoir pressure also increases. This pressure differential causes additional stress around the wellbore. Eventually, as the rate is increased, this pressure differential will cause stresses that will exceed the stress needed to break the rock apart, and a fracture is formed. At this point, if the pumps are shut down or the pressure is bleed off, the fracture will close again. Eventually, depending on how hard the rock is and the magnitude of the force acting to close the fracture, it will be as if the rock had never been fractured. By itself, this would not necessarily produce any increase in production. However, if we pump some propping agent, or proppant, into the fracture and then release the pressure, the fracture will stay propped open, providing the proppant is stronger than the forces trying to close the fracture. If this proppant also has significant porosity, then under the right circumstances a path of increased permeability has been created from the reservoir to the wellbore. If the treatment has been designed correctly, this will produce an increase in production. Generally, the process requires that a highly viscous fluid is pumped into the well at high rate and pressure, although this is not always the case (see Skin Bypass Fracturing, below). High rate and high pressure mean horsepower, and this is why the process generally involves large trucks or skids with huge diesel engines and massive pumps. A typical frac pump will be rated at 700 to 2700 hydraulic horsepower (HHP) – to put this in perspective, the average car engine (outside North America, that is) has a maximum power output of 80 to 100 HP.

Pressure, Rate, Proppant Concentration

In order to create the fracture, a fluid stage known as the pad is generally pumped first. This is then followed by several stages of proppant-laden fluid, which actually caries the proppant into the fracture. Finally, the whole treatment is displaced to the perforations. These stages are pumped consecutively, without any pauses. Once the displacement has finished, the pumps are shut down and the fracture is allowed to close on the proppant. The Frac Engineer can vary the pad size, proppant stage sizes, number of proppant stages, proppant concentration within the stages, the overall pump rate and the fluid type in order to produce the required fracture characteristics. Typically, the treatment will look like Figure 2.1a:-

BHTP

Rate STP

Prop Conc

Time

Figure 2.1a – Typical hydraulic fracture treatment job plot

Page 4



2.2

Pressure

Everybody understands what pressure is. Or at least, everyone thinks they understand what pressure is. If you ask someone to define pressure, then they will usually say “force divided by area”, or something similar. This is not what pressure is - it is merely how we measure, create and use pressure. The simple fact is that pressure is stored energy, and we use that energy to perform work on the formation during the fracturing process. Everything we do in fracturing can be thought of in terms of energy. For instance, when we pump a fluid into a fracture we start out with chemical energy – in the form of diesel fuel. This is converted to mechanical energy by the diesel engine. The high pressure pump then transfers this mechanical energy into pressure in the fracturing fluid. As the fluid moves into the formation, the pressure is transformed into stress in the formation (see below), which is another form of stored energy, and so the walls of the fracture are pushed back, creating fracture width and forcing the fracture to propagate. Work is defined as the rate at which energy is used – in the SI system, one watt is defined as a joule per second. Therefore, by observing the way the pressure is changing, or not changing, with respect to time, we can tell how much work we are performing on the formation (see Section 10.2 – Nolte Analysis). Pressure and stress are essentially the same thing. The only difference is that stresses act in solids and pressures act in liquids and gases. Because liquids and gases easily deform away from any applied force, pressures tend to act equally in all directions. Stresses, however, tend to act along planes, so that a solid experiencing a stress will always have a plane where the stresses are a maximum, and a plane perpendicular to this where the stresses are at a minimum. In fracturing, we refer to several different pressures. These names merely refer to where and when we are measuring (or calculating) the pressure; Surface Treating Pressure, STP – also referred to as wellhead pressure, injection pressure, tubing pressure (if we are pumping down the tubing), PSTP, Pwellhead, Ptubing and so on. The name speaks for itself – it is the pressure that the pumps have to act against at the surface. Hydrostatic Pressure – also referred to as hydrostatic head, PH, HH and Phydro. This is the pressure downhole due to the weight of the column of fluid in the well. This pressure is a function of the density of the fluid and the vertical depth: HH

= 0.433 γ TVD .................................................................. (2.1)

where HH is the hydrostatic head in psi, γ is the specific gravity of the fluid and TVD is the true vertical depth at which the pressure is acting. This looks relatively easy to calculate, but can get quite complicated in a dynamic system in a deviated well with fluids of several different densities actually in the well – which is the usual situation during a frac job. We use computers to keep track of this. Tubing Friction Pressure – also known simply as friction pressure, Pfrict or ∆Pfrict. This pressure will be covered in more detail in later sections of this manual (see Section 4). For now, we can define it qualitatively as the pressure caused by the resistance of the fluid to flow down the tubing. Friction pressure decreases with increasing tubular diameter and increases with rate. Bottom Hole Treating Pressure – BHTP or PBHT. This is the pressure inside the well, by the formation being treated. Generally, at is calculated at the center of the perforated interval. At this point, the fluid has not passed through the perforations or into the fracture. Unless there are gauges in the well, or there is a static column, this pressure is usually calculated:-

Page 5


BJ Services’ Frac Manual 2. Basics of Hydraulic Fracturing BHTP

= STP + HH - ∆Pfrict ........................................................ (2.2)

As there are always uncertainties with the calculation of ∆Pfrict (unless fluid rate is zero), there will always be uncertainties in calculated BHTP. Perforation Friction Pressure – also known as perforation friction or ∆Pperf. This is the pressure drop experienced by the fluid as it passes through narrow restrictions generally referred to as perforations:∆Pperf

2

=

2.93 SG (q/n) 4 .............................................................. (2.3) d

where ∆Pperf is in psi, SG is the specific gravity of the fluid, q is the slurry rate in bpm, d is the perforation diameter in inches and n is the number of perforations. Near Wellbore Friction Pressure – a.k.a. near wellbore friction or ∆Pnwb. This is the sum of the perforation friction and any pressure losses caused by tortuosity, which will be covered in greater detail in Section 10. Closure Pressure – Pc or Pclosure. This is the force acting to close the fracture. Below this pressure the fracture is closed, above this pressure the fracture is open. This value is very important in fracturing and is usually determined from a minifrac, by careful examination of the pressure decline after the pumps have been shut down. Extension Pressure – or Pext. This is the pressure required in the frac fluid in the fracture in order to make the fracture propagate. It is usually 100 to 200 psi greater than the closure pressure, and this pressure differential represents the energy required to actually make the fracture propagate, as opposed to merely keeping it open (i.e. Pclosure). In hard formations, fracture extension pressure is close to the closure pressure. In softer formations, where significant quantities of energy can be absorbed by plastic deformation at the fracture tip, extension pressure can be significantly higher than closure pressure (see Section 9). The fracture extension pressure can be obtained from a step rate test. Net Pressure – or Pnet. This is a fundamental value used in fracturing and the analysis of this variable forms a whole branch of frac theory by itself. This will be discussed in detail later on in this manual. For now, Pnet is the difference between the fluid pressure in the fracture and the closure pressure, such that:Pnet

= BHTP – ∆Pnwb - Pclosure .................................................. (2.4) = STP + HH – ∆Pfrcit – ∆Pnwb - Pclosure ............................... (2.5)

Pnet is a measure of how much work is being performed on the formation. By analysing the trends in Pnet a great deal can be determined about how the fracture is growing – or shrinking. Instantaneous Shut in Pressure – or ISIP or ISDP. This is the pressure, which can be determined either at surface or bottom hole, which is obtained just after the pumps are shut down, at the start of a pressure decline. If measured at bottom hole, the ISIP should be equal to the BHTP, provided Pnwb is zero. One of the methods for determining if the Pnwb is significant is to compare the ISIP and the BHTP from a minifrac (provided the BHTP is reliable).

2.3

Basic Fracture Characteristics

Every fracture, regardless of how it was pumped or what it is designed to achieve, has certain basic characteristics, as shown in Figure 2.3a (below).

Page 6


BJ Services’ Frac Manual 2. Basics of Hydraulic Fracturing All fracture modeling is designed around determining these three characteristics, height H, half length xf and width W. Once these three characteristics have been determined, other quantities such as proppant volume, fracture conductivity and ultimately production increase can be determined. It is usually assumed that the two wings of the fracture are identical and 180 º apart (i.e. on opposite sides of the wellbore. This is not necessarily the case. It is also normal to model the fracture wings as being elipitcal in shape - however, the reality is that the geometry is probably quite a bit more complex. However, based on the three characteristics of width, half length and height, we can define a few simple parameters, which will be used frequently in this manual:-

W

xf

H

Figure 2.3a – Diagram showing fracture half Length xf, fracture height H, and fracture width W.

Aspect ratio; AR

H = x .................................................................................. (2.6) f

So a radial frac, which is perfectly circular and has a height equal to twice the fracture half length, has an AR of 0.5 Fracture conductivity; Fc

=w ¯ .kp .............................................................................. (2.7)

where w ¯ is the average fracture width and kp is the permeability of the proppant pack. Remember that the width in Equation 2.7 is the propped width, which is usually less than the width actually created during the treatment. The propped width is a function of the volume of proppant pumped into the fracture, expressed in terms of the mass of proppant per unit area of the fracture face. This areal proppant concentration is expressed in terms of lbs/sq ft, and is not to be confused with the slurry proppant concentration, that is expressed in lbs/gal (or ppg). This is a measure of how much proppant is added by the surface mixing equipment to a gallon of frac fluid. Another way of expressing slurry proppant concentration, which is used less often but is clearer and easier to understand, is ppa, or lbs of proppant added. This clearly illustrates the quantity of proppant being added to a gallon of clean fluid.

Page 7



2.4

Fluid Leakoff

Hydraulic fracture treatments are pumped into permeable formations – there is little point in carrying out the process in a formation with zero permeability. This means that as the fracturing fluid is being pumped into the formation, a certain proportion of this fluid is being lost into the formation as fluid leakoff. The leakoff coefficient is a function of the formation permeability kf, the fracture area A, the pressure differential between the fracturing fluid and the formation ∆P, the formation compressibility, viscosity and the fluid characteristics. Often, this coefficient is set as a constant throughout the treatment, which means that the fluid loss rate varies with time and fracture area only, and does not vary with pressure differential or fluid type. The effect of the formation permeability and the fluid characteristics are often combined together into a single leakoff coefficient, variously called CT, CL or Ceff. We shall use Ceff. This coefficient defines the volume of fluid leaked off into the formation VL, as follows:VL

= π Ceff A

t ................................................................... (2.8) ½

where t is the time that the fracture has been open. The units of Ceff are generally ft/min , so in Equation 2.8 if the area is in square feet, the leakoff volume is in cubic feet. Remember that the area A is the surface area of the whole fracture, including both sides of both wings of the fracture. A fracture geometry model must be used to determine the value for A. In a multilayer reservoir, with different values of Ceff for each zone, the total leakoff will be the sum of the leakoff for each zone. The leakoff coefficient is usually determined from minifrac tests and from analysis of previous treatments. A more accurate method for calculating fluid loss is to use a dynamic leakoff model, in which variations in the pressure differential and the fluid composition are taken into account. In dynamic leakoff, the overall leakoff coefficient is generally assumed to have three components; the viscosity controlled coefficient CV or CI, the compressibility controlled coefficient CC or CII and the wall-building coefficient Cw or CIII. The viscosity controlled coefficient is the effect of the fracture fluid filtrate moving into the formation under Darcy linear flow conditions, and is defined as (in field units):CI

= 0.0469

kf φ ∆P ...................................................... (2.9) 2µf

where kf is the permeability of the formation to the frac fluid filtrate, φ is the formation porosity and µf is the frac fluid filtrate viscosity in cp. The compressibility controlled coefficient defines the leakoff which is due to the formation compressing, and allowing volume into which the frac fluid filtrate can move. It is defined, in field units, as:CII

= 0.0374 ∆P

kr cf φ

µr

................................................. (2.10)

where kr is the permeability of the formation to the reservoir fluid, cf is the compressibility of -1 the formation in psi and µr is the reservoir fluid viscosity in cp. The wall building coefficient is usually determined experimentally using a standard fluid loss test. The volume of filtrate is plotted against the square root of time, to give a slope m. The wall building coefficient is then defined as (in field units):-

Page 8


BJ Services’ Frac Manual 2. Basics of Hydraulic Fracturing CIII

=

0.0164 m .................................................................... (2.11) Af

where Af is the area of the filter cake in the fluid loss cell. Generally, modern fracture simulator will have wall-building coefficients for a wide range of fracturing fluids, so that all the Engineer has to do is select the fluid type. The three components can then be combined to produce Ceff as follows:Ceff

=

1+

2 ClClICllI ............................ (2.12) 2 2 2 (ClClll) +(4 CII ( CI + CIII )) 2

This is for dynamic fluid leakoff. The components can be arranged in a different form for harmonic fluid leakoff:Ceff

(ClCllClll) = (C C + C C + C C ) .................................................. (2.13) l ll ll lll l lll

This process of deducing the theoretical leakoff coefficient looks to be rather intimidating, and in practice is only used in fracture simulators. During minifrac analysis, the permeability of the formation and the wall building coefficient are varied to produce the required leakoff rate. Generally, the dynamic model is better than the harmonic, although under most circumstances there will not be much difference between the two. This is especially true for a non-wall-building fluid, or for gas reservoirs. Another form of fluid loss into the formation is called spurt loss. This is the fluid loss which occurs on “new” parts of the fracture, before the fluid has a chance to build up a filter cake. Usually, the fracture models take a simplistic approach to spurt loss and use a spurt loss coefficient, Sp , such that:Vs

= A Sp ............................................................................. (2.14)

where Vs is the volume of fluid lost due to spurt loss and A is the total area of the fracture (both wings). A more detailed approach to spurt loss (and fluid loss in general) can be found in SPE Monograph Volume 12, Recent Advances in Hydraulic Fracturing, Chapter 8 (see references).

2.5

Near Wellbore Damage and Skin Factor

Darcy’s Equation for radial flow defines the rate at which oil is produced from the reservoir into the wellbore, under steady state flow conditions. In field units for an oil well, Darcy’s Equation becomes:q

=

0.00708 k h ∆P .......................................................... (2.15) µ ln (re/rw)

where q is the downhole flow rate in bbls/day. We can see that the wellbore radius, rw has a huge impact on the flow rate. This is easily visualised, as the closer the fluid comes to the wellbore, the more congested the flow paths become and the faster the fluid has to move. Therefore, the final few inches by the wellbore are the most critical part of the reservoir. Unfortunately, this is also the part of the reservoir most susceptible to damage. This damage can come from a variety of sources, but most often comes from the process of drilling the well in the first place. A full discussion on sources of formation damage is beyond the scope of this manual. However, the major sources are; particulates in the drilling fluid (barite, calcium carbonate Page 9


BJ Services’ Frac Manual 2. Basics of Hydraulic Fracturing etc), filtrate invasion, whole fluid invasion, pH of drilling fluid and surfactants in the drilling fluid. What this results in, is a region around the wellbore of reduced permeability, as illustrated in Figure 2.5a. This reduction in permeability around the wellbore is generally referred to as the Skin, which was first rationalised by van Everdingen and Hurst (1949). The skin factor, S, is a variable that is used to describe the difference between the ideal production given in Equation 2.15, and the actual production through the damaged area. Generally, the skin is measured using a pressure build up test. The API has defined the skin factor for an oil well as follows (see Section 19):S

= 1.151

P1hr - Pwf k - log10 + 3.23 ....................... (2.16) m φµcrw2

where Pwf is the bottom hole stabilised flowing pressure (psi), P1hr is the bottom hole pressure after one hour of static pressure build up (psi), k is the formation permeability, m is the slope of the graph of P against log10[(t + ∆t)/∆t ] (in psi per log10 cycle), φ is the porosity (fraction), µ -1 is the fluid viscosity (cp), c is the average reservoir compressibility (psi ) and rw is the wellbore radius (feet).

Wellbore

Damage

Permeability low

high

Figure 2.5a – Illustration of the reduction in permeability around the wellbore

To help matters, m can be found from the following (in field units):m

=

162.6 q µ .................................................................... (2.17) kh

Note that both q and µ are at bottom hole conditions. A completely undamaged reservoir will have a skin factor of zero. Damaged reservoirs will have skins in the ranging from 0 to 50 or even higher. Under certain circumstances, stimulation can result in a negative skin factor, which means that the well is producing more than predicted by ideal Darcy flow. Once the skin factor has been obtained, it can be used in Darcy’s Equation to give the modified flow from a skin damaged reservoir:q

Page 10

=

0.00708 k h ∆P ......................................................... (2.18) µ [ln (re/rw) + S]


BJ Services’ Frac Manual 2. Basics of Hydraulic Fracturing This means that as S increases, flow rate decreases, and vice versa. Another way of employing the skin factor is to use an effective wellbore radius, as given in Equation 2.19:rw ’

-S

= rw e ............................................................................ (2.19)

This means that in a damaged wellbore, the well is behaving as if it had a smaller wellbore radius, whilst a stimulated reservoir behaves as if it had a larger wellbore radius.

References Howard, G.C., and Fast, C.R.: Hydraulic Fracturing, Monograph Series Vol 2, SPE, Dallas, Texas (1970). Gidley , J.L., et al.: Recent Advances in Hydraulic Fracturing, Monograph Series Vol 12, SPE, Richardson, Texas (1989). Archer, J.S. and Wall, C.G.: Petroleum Engineering – Principles and Practices, Graham and Trotman, London (1986). van Everdingen, A.F. and Hurst, W.: “The Application of the Laplace Transformation to Flow Problems in Reservoirs”, 1949, Trans., AIME, 186, 305-324. Meyer and Associates, MFrac version 5.10 on-line Help section, 2003.

Page 11


BJ Services’ Frac Manual 3. Types of Hydraulic Fracturing

3.

Types of Hydraulic Fracturing

There are various different types of hydraulic fracturing, which have evolved around the basic process of creating a fracture and then propping it open. The type of treatment selected depends upon the formation characteristics (permeability, skin damage, fluid sensitivity, formation strength), the objectives of the treatment (stimulation, sand control, skin bypass or a combination) and the constraints we have to work within (cost, logistics, equipment etc).

3.1

Low Permeability Fracturing

There are various different types of hydraulic fracturing, which have evolved around the basic process of creating a fracture and then propping it open. The type of treatment selected depends upon the formation characteristics (permeability, skin damage, fluid sensitivity, formation strength), the objectives of the treatment (stimulation, sand control, skin bypass or a combination) and the constraints we have to work within (cost, logistics, equipment etc). This type of fracturing is often carried out in tight gas formations, found in areas such as the Rocky Mountains, Algeria, Western Germany, parts of Australia and many other places worldwide. Permeabilities for such formations range 1 md right down to 1 µd and less. This type of treatment is also applicable to low permeability oil formations, although permeabilities tend to be 1 or 2 orders of magnitude greater. In order for hydrocarbons to flow down the fracture, rather than through the adjacent formation, the fracture must be more conductive than the formation. Given that the kp for 20/40 Colorado Silica frac sand is 275 darcies (provided closure pressure is below 3,000 psi), we can see that even a very narrow fracture will have a much higher conductivity than the formation itself. This does not allow for the effects of non-Darcy flow (see Section 10). Therefore, the limiting factor defining how much the reservoir production has increased is not how conductive the fracture is (as any propped fracture will be significantly more conductive than the formation), but instead is how fast the formation can get the hydrocarbon to the fracture. Therefore, when treating low permeability reservoirs, fractures should be designed with a specific minimum fracture conductivity, but a large surface area - which means, because formations are usually limited in height, designing for maximum fracture half length, xf. See Section 17.9 for a detailed discussion of how to determine the required fracture conductivity. Because formation permeability is low, fluid leakoff also tends to be low. This has two consequences. First, pad volumes tend to be very low, relative to the rest of the job volumes. In some cases, a pad is hardly needed at all – the proppant-laden fluid can be used to create the fracture. The second consequence is that fracture closure time – the length of time taken for the fracture to close on the proppant after the treatment has finished – tends to be long. This means that the fracturing fluid has to suspend the proppant for a relatively long period of time at bottom hole temperature. Therefore, hydraulic fracture treatments in low permeability formations tend to have fairly large fluid and proppant volumes, although the overall proppant concentration in the fluid is relatively low. Pad volumes are small. Treatment fluids are usually fairly robust, capable of maintaining viscosity for extended periods of time. The process of designing for low permeability formations is discussed in greater detail in Section 17.5.

Page 12



3.2

High Permeability Fracturing

High permeability fracturing is, not unexpectedly, the opposite of low permeability fracturing. In high permeability formations, moving the fluid through the rock to the fracture is easy. The hardest part is creating a fracture that is more conductive than the formation in the near wellbore region. In Equation 2.7, the concept of fracture conductivity was introduced. The next step is to define relative or dimensionless conductivity, CfD (often referred to as FcD in many previous publications):

CfD =

Fc xf kf ................................................................................. (3.1)

where xf is the fracture half length and kf is the permeability of the formation. CfD is a measure of how conductive a fracture is compared to the formation and compares the ability of the fracture to deliver fluids to the wellbore with the ability of the formation to deliver fluids to the fracture. A CfD of greater than one means that the fracture is more conductive than the formation, whereas a CfD of less than one means that the fracture is less conductive than the formation and the reservoir fluids flow more easily through the formation. This does not account for the effects of the skin factor – in reality all the fracture needs to be in order to increase production, is more conductive than the skin (see Section 3.4 – Skin Bypass Fracturing). From Equation 2.7, which stated that Fc = w ¯ .kp, we can see that two parts of the definition of CfD are fixed; kf and kp (although kp can be increased to a certain extent by using a better quality proppant). Therefore, in order to increase dimensionless conductivity, we have to maximise w ¯ and minimise xf. This means that we need a very short, wide fracture. In order to achieve this, a technique known as the Tip Screen Out (TSO) is often used. This will be discussed in more detail in Section 17.3. Because the formations have high permeability, fluid leakoff tends to be very high. Therefore, pad volumes tend to be a significant part of the treatment. This high leakoff is used by the technique of TSO fracturing. Young’s modulus tends to be very low, which means that creating fracture width is relatively easy. Formations with very high permeability also tend to have two other characteristics. First, they are often weak or unconsolidated, so that the fracturing process is often combined with gravel packing techniques to produce a frac pack treatment (see below, Section 3.3). Second, the formations also tend to have large skin factors, so that a significant production increase can be obtained simply by providing a conductive path through the skin (see Section 3.4, below). The processes involved in designing treatments for high permeability are discussed in greater detail in Section 17.3

3.3

Frac and Pack Treatments

The frac and pack (or simply frac-pack) treatment is a combination of a high permeability fracture treatment and a gravel pack treatment. Technically, the process of designing the actual treatment is the same as for a high permeability frac. Operationally, however, the process is much more complex, due the presence in the wellbore of the gravel pack completion. Figure 3.3a illustrates this.

Page 13



GP/Prod. Packer Fluid Control Valve

Blank Pipe Screen Sump Packer Figure 3.3a – Diagram illustrating the components of the frac-pack completion. Setting tool is shown in the squeeze position.

Figure 3.3b – Diagram illustrating two of the three positions in which a standard gravel pack or frac pack tool can be set. The left hand side shows the squeeze position, in which fluids flow down the tubing, through the crossover, out into the annulus below the GP packer and into the formation. The right hand side shows the lower circulating position. Fluid flows down to the perforations, as for the squeeze position. However, because the setting tool has been shifted upwards, the fluid can flow either into the formation, or back through the screens, up the washpipe (inside the screens) through the crossover, and out into the annulus above the tubing (shown in blue). By closing the annulus at surface, the fluid can be squeezed into the formation, whilst maintaining a dead string on the annulus, to monitor BHP.

Page 14


BJ Services’ Frac Manual 3. Types of Hydraulic Fracturing The treatment is normally pumped with the setting tool in the squeeze position, although sometimes the tool is in the lower circulating position (see Figure 3.3b). In either case, fracturing fluids are pumped down the tubing, through the setting tools, through the crossover, out into the annulus and into the perforations. As stated before, the pumping schedule is designed as if the completion did not exist, and a normal high permeability fracture treatment was being performed. With one exception – extra proppant (or gravel) is pumped on the final stage, in order to fill the annulus space between the screen and the casing, producing the gravel pack. The process of designing a frac and pack treatment will be discussed in more detail in Sections 17.3 and 17.4.

3.4

Skin Bypass Treatments

Skin bypass treatments are designed to do exactly what the name describes – bypass skin damage. These treatments are not necessarily designed to be the absolute optimum stimulation treatment for the well. Instead, these treatments are designed to be small, cost effective and easy to run operationally. Often these treatments are pumped in places where space or equipment weight is a limiting factor – such as offshore. In many cases, if the frac engineer was given a technical free hand to design the optimum treatment, the job itself would be much larger. However, given the restraints of cost and space that are often placed upon frac engineers, the skin bypass frac is an attempt (often highly successful) to produce effective stimulation. The skin bypass frac can also be considered as a more effective alternative to matrix acidising, when factors such as mineralogy, temperature, logistics and cost prevent the use of acid.

Figure 3.4a – Diagram illustrating how the skin bypass fracture penetrates the skin to allow undamaged communication between the reservoir and the wellbore.

Figure 3.4a shows the basic concept behind the skin bypass frac. Although the formation has considerable damage (dark-shaded area), this is effectively bypassed by the more conductive path created by the fracture. In order for the fracture to produce a production increase, it does not have to be more conductive than the formation (i.e. CfD > 1.0). It merely has to be more conductive than the damaged area. Of course, usually we are usually aiming for considerably Page 15


BJ Services’ Frac Manual 3. Types of Hydraulic Fracturing more than just the production increase due to skin bypass. Given that Skin Bypass Fracs are normally carried out on marginal wells (wells that cannot justify the expense of a major stimulation treatment), often the economics dictates that significant production increase must be obtained. Equation 3.1 gave the definition of dimensionless conductivity, which has to be greater than 1.0 for the fracture to provide stimulation of the formation. Equation 3.2 shows the condition, for a fracture which has HD ≤ 1.0, under which the skin bypass fracture is more conductive than the formation: Fc H kf

>

ln (re/rw) ln(re/rw + S) .................................................. (3.2)

Where Fc is the fracture conductivity (mdft), H is the fracture height (ft), re is the radial extent (ft), rw is the wellbore radius and S is the skin factor. So if S = 0, the RHS of Equation 3.2 goes to 1, so that then Fc has to be greater than H.kf , which is another way of saying that the CfD has to be greater than one. This Equation takes into account the fact that the fracture does not cover the entire zone vertically. However, it is an approximation, as it does not account for vertical flow or non-Darcy effects (Section 10). HD is the dimensionless height and is equal to the fracture height divided by the formation height.

3.5

Coal Bed Methane Fracturing

It is estimated that for every tonne of coal that is generated underground - by the process of coalification - up to 45 mscf of gas (mostly methane) is generated. In areas such as the Southern North Sea, this gas migrates upwards until it reaches an impermeable layer, so that the coal itself contains very little gas. In other cases, nearly all the gas remains in place, waiting to be produced. Coal itself usually has very low matrix permeability, with the gas being produced through natural fractures (called cleats) and through desorption from the coal itself. The objective of coal bed methane fracturing is to connect up the cleats with a propped fracture, allowing the gas to be produced both from the cleats and from the coal CBM fracturing is more of an art than a science. Because of the unusual characteristics of the formations, most fracture simulators are unable to accurately model these treatments. Engineers usually have to rely on experience and trial and error. These treatments usually consist of large volumes of proppant, pumped at low concentrations, at high rates. Various fluid systems have been used, but recent work has demonstrated that crosslinked fluids, especially guar-based gels, can be very damaging to the formation. The trend has been towards HEC, foams and even just water as the carrier fluid. Proppant concentrations tend to be in the 3 to 4 ppg range. Because wells are relatively low rate, large fracture conductivities are not required – what is needed is a conductive path from cleat to cleat. As formations are usually shallow, sand is generally selected as the proppant. CBM wells often tend to be marginal. They will not produce economically without a frac treatment, but even after a frac can be very low rate. Therefore, fracturing treatments tend to be fairly low tech, no frills operations, using minimal fluids technology and often eliminating the need for modern, sophisticated, computerised blending and pumping equipment. CBM fracturing will be covered in more detail in Section 17.7. Gas production from a CBM reservoir relies on different mechanisms than production from conventional reservoirs. The main production mechanism is not expansion of gas in pore spaces - coals generally have little or no primary porosity. Instead, as stated above, the gas is adsorbed into the coal itself. In order to produce the gas, the pressure has to be reduced below a specific critical pressure, at which point the gas starts to desorb. Some CBM Page 16


BJ Services’ Frac Manual 3. Types of Hydraulic Fracturing reservoirs are naturally below this critical pressure. Most, however, are significantly above this pressure. In such cases, considerable quantities of water have to be rapidly produced in order to get the reservoir pressure low enough to initiate gas desorption. Often, a propped fracture plays a critical role in this de-watering process.

3.6

Fracturing Through Coiled Tubing

Fracturing through coiled tubing has been around since the early 1990’s, and was first carried out through a string of coiled tubing that was left in the well after the treatment, becoming the production tubing. However, as the industry began to perceive the advantages of this process – and as Engineers began to leave their preconceived coiled tubing ideas behind – the concept has become more widely accepted. The advantage of coiled tubing fracturing does not lie with the design or type of fracture that is placed in the ground, as most types of fracture can be performed this way. The benefits of CT fracturing lie in the operational aspects of how the treatments are placed. The obvious limitation for coiled tubing fracturing is the diameter of the coil and the maximum pressure it can be taken to. However, this restriction is not nearly as bad as it initially seems. With modern fluid systems, friction pressure down the coiled tubing can be dramatically reduced, allowing treatments to be pumped at quite high rates. Also, as the coiled tubing is static during the treatment (i.e. the tubing is not being plastically deformed on a continuous basis), the maximum allowable pressure is far higher than is normal for CT operations.

Advantages 1.

2.

3. 4.

The coiled tubing can be used to isolate the completion from the fracturing process. By setting a squeeze packer at the end of the tubing, the hole tubing string is protected from the pressure and temperature changes normally experienced by the completion. This means that completions that are pressure-limited (due to sliding sleeves, packer ratings, poor quality tubing, wellhead size etc) can be fractured. Completions which cannot be cooled down too much (due to risk of stinging the tubing out if the PBR on the packer), can also be fractured. Coiled tubing fracturing is particularly effective when working on monobore completions, or on wells that have not yet been completed. By using an opposing cup tool, the coiled tubing can be used to easily isolate one zone from another. An extension of this, is that the tool can be very easily moved from one zone to another, allowing multiple fracs to be performed in rapid succession. If required, the coiled tubing can be used to gas lift the well on to production after the treatment(s). Coiled tubing can often be used as an alternative to a workover. This can mean significant cost saving, especially offshore.

Disadvantages 1.

2. 3.

4.

The extra cost of the coiled tubing unit, over and above the cost of the frac spread. However, often this extra cost can produce savings in other areas (rig time, frac crew time etc). The operating company must also be prepared to pay for some or all of the cost of the coiled tubing string. The extra space needed, due to the extra equipment required as compared to the frac spread by itself. Of course, if the CT unit is being used as an alternative to a workover rig, this may not be as significant. Rate limitations. In general, for a given fluid system, higher rates can be achieved through completions than through coiled tubing. However, it should be remembered that it is usually possible to take the static coiled tubing to higher pressures than the completion/wellhead assembly. Although it is possible to frac through coiled tubing with standard fluid systems, as the depth increases and/or the coiled tubing diameter decreases, it may be necessary to use more exotic and expensive fluid systems.

Page 17



References Product Catalogue, Colorado Silica Sand, 1994 Economides, M.J., and Nolte, K.G.: Reservoir Stimulation, Schlumberger Educational Services, 1987. Gidley, J.L., et al: Recent Advances in Hydraulic Fracturing, Monograph Series Vol 12, SPE, Richardson, Texas (1989). Bradley, H.B. (Ed): Petroleum Engineers Handbook, SPE, Richardson, Texas (1987) Rae, P., Martin, A.N., and Sinanan, B.: “Skin Bypass Fracs: Proof that Size is Not Important”, SPE 56473, presented at the SPE Annual Technical Conference and Exhibition, Houston, October 1999. O’Driscoll, K.: Middle-East Region Coal Bed Methane Fracturing Manual, BJ Services, 1995. Gavin, W.G.: “Fracturing Through Coiled Tubing – Recent Developments and Case Histories”, SPE 60690, presented at the 2000 SPE/ICoTA Coiled Tubing Roundtable, Houston, April 2000. Wong, G.K., Fors, R.R., Casassa, J.S., Hite, R.H., and Shlyapobersky, J.: “Design, Execution and Evaluation of Frac and Pack (F and P) Treatments in Unconsolidated Sand Formations in the Gulf of Mexico”, SPE 26563, presented at the SPE Annual Technical Conference and Exhibition, Houston TX, Oct 1993. Tiner, R.L., Ely, J.W. and Schraufnagel, R.: “Frac Packs – State of the Art”, SPE 36456, presented at the SPE Annual Technical Conference and Exhibition, Denver CO, Oct 1996.

Page 18


BJ Services’ Frac Manual 4. Fluid Mechanics

4.

Fluid Mechanics

Fluid Mechanics is the study of the behaviour of fluids. In the oil field, this means that fluid mechanics is used to predict fluid friction pressures and the forces due to the dynamics of fluid flow. Rheology is the study of the deformation and flow of matter, and in the oil field is used to predict the resistance of a fluid to the application of a force or pressure.

4.1

Fundamental Fluid Properties

Density (ρ) -

A measure of how much matter a material contains within a unit of volume. The denser a material is, the heavier a given volume. Provided the liquid composition remains constant, we can think of fluid density (especially for water-based fluid systems) as a constant – although it will actually decrease slightly with increasing temperature and increase slightly with increasing pressure. Hydrocarbon-based fluid systems are significantly more compressible, and assuming a constant density can result in inaccuracies (see references for diesel data).

Viscosity (µ) -

Viscosity is a measure of how much a fluid resists deformation as a result of an applied force or pressure. It is a measure of how “thick” the fluid is. Viscosity is only very rarely a constant value, as it can change dramatically with temperature, applied shear stress and fluid composition. Viscosity is defined as the relationship between shear stress and shear rate.

Temperature (T) -

A measure of how much energy a material contains – the hotter the material, the more energy. Although strictly speaking temperature is not a fundamental property, in the oil field it an important parameter that needs to quantified. Most fluid properties are affected to a greater or lesser extent by temperature.

4.2

Shear Stress and Shear Rate

Shear Rate (γ). In fluid mechanics, shear rate is a measure of how fast a fluid is flowing past a fixed surface. Shear rate can be thought of as a measure of how much agitation a fluid is receiving. Causes of Shear Rate:-

Spinning centrifugal pump Flow through a pipe Fann 35 Test Jet mixer Tank agitators

Shear Stress (τ). Shear stress is the resistance the fluid produces to an applied shear rate. For instance, it requires more force (pressure) to pump water at 20 bpm than at 10 bpm. Viscosity (µ). The fluid property that defines how much shear stress is produced by a shear rate, is called viscosity. The greater the viscosity, the greater the resistance of a fluid to shear agitation.

Newton’s Law of Fluids µ Page 19

=

τ γ ........................................................................................... (4.1)



Shear Stress, τ

This is known as Newton’s law of fluids, and is illustrated in Figure 4.2a:-

0

Slope = µ

0

Shear Rate, γ

Figure 4.2a – Graph illustrating Newton’s law of fluids

In oil field units, Newton’s law can be expressed as follows:-

µ

= 47,879

τ γ .............................................................................. (4.2) 2

-1

with µ measured in cp (centipoise), τ in lbf/ft and γ in sec . Newton was the first to realise the relationship in fluids between an applied force and the resistance to that force. His experiments were carried out on simple fluids such as water and brine, and not on more complex fluids, such as those used in stimulation activities.

4.3

Types of Fluid

In the oil field, we generally deal with three different types of fluids, according to how the relationship between shear stress and shear rate develops. These fluid types are defined below.

Newtonian Fluids As illustrated in Figure 4.2a, these are fluids for which Newton’s law is valid. Newtonian fluids have a straight line (linear) relationship between shear rate and shear stress until turbulence occurs. Equations 4.1 and 4.2 are valid. Examples of Newtonian fluids include:Fresh Water Sea Water Most Acids (ungelled) Diesel Alcohols Gases

Bingham Plastic Fluids Bingham plastic fluids require an initial shear stress to be induced before they will deform. Put another way, they have a yield point or gel strength that must be broken before the fluid can move (although some fluids have a gel strength that is nothing to do with yielding). This type of fluid is not Newtonian, although they usually have a constant viscosity once the initial gel strength has been overcome. Page 20



τ

= Yp + Pvγ ......................................................................... (4.3) 2

Yp is the yield point, and in the oil field has units of lbf/100 ft (note that in the oil field, τ has 2 the units lbf/ft , so the value for Yp has to be converted before it is used), whilst Pv is the plastic viscosity, with cp as its units.

Shear Stress, τ

Figure 4.3a illustrates the behaviour of a Bingham plastic fluid Examples of Bingham plastic fluids include some cement slurries and some drilling muds.

Slope = Pv

Yp

0

0

Shear Rate, γ

Figure 4.3a – Relationship between shear rate and shear stress for a Bingham plastic fluid.

Power Law Fluids The third group of fluids is generally referred to as power law fluids, although there are other names which have been used to describe them. In general, there is no linear relationship between shear rate and shear stress, so that apparent viscosity (the viscosity which the fluid appears to have, at a specific shear rate) changes with shear rate. The following Equation describes the behaviour of the power law fluid, and this is illustrated in Figure 4.3b. n’

= K’γ .............................................................................. (4.4)

Shear Stress, τ

τ

0

0

Shear Rate, γ

Figure 4.3b – Relationship between shear rate and shear stress for a power law fluid. Note that the graph shows the relationship in its most common form - “shear thinning”. However, in certain fluids the line can also curve upward - “shear thickening”.

Page 21


BJ Services’ Frac Manual 4. Fluid Mechanics K’ is referred to as the power law consistency index, and in order to be coherent has the n’ 2 rather awkward units of lbf sec /ft . n’ is the power law index and is dimensionless. In order to determine n’ and K’, the log of Equation 4.4 is taken; log τ

= log K’ + n’ log γ .......................................................... (4.5)

On a plot of logτ against logγ, the intercept of the vertical axis is log K’ and the gradient of the line is n’, as shown in Figure 4.3c;

log τ

Slope = n’

log K’

0

0

log γ Figure 4.3c – Power law fluid log-log plot

Power law fluids can be divided into 3 major categories; Shear-thinning fluids. In these fluids, n’ is less than 1, so that the fluids experience a decrease in apparent viscosity as the shear rate increases. Most of the fluids used for fracturing fall within this category. Newtonian fluids. Newtonian fluids are a special case of power law fluids in which n’ is equal to one, i.e. the viscosity is constant and equal to K’. Shear-thickening fluids. These fluids have an n’ greater than one, and so exhibit an increase in apparent viscosity as shear rate increases. Extreme examples of these fluids can behave as it they were solids when exposed to even moderate shear forces. Another example of a power law fluid is the Herschel-Buckley fluid, which is often used to model the flow behaviour of foams;

τ

= τ’o + K’’ γ

n’’

.................................................................... (4.6)

where τ’o is the threshold shear stress, K’’ is the Herschel-Buckley consistency index and n’’ the Herschel-Buckley exponent. Herschel-Buckley fluids are basically a combination of the Bingham plastic fluid and the power law fluid. An initial threshold shear stress has to be overcome before the fluid will flow. Once this has happened, the viscosity is not constant, and will vary according to the shear rate.

Page 22



4.4

Measuring Viscosity

In order to measure viscosity, two variables need to be determined. First, the shear rate of some moving device within the fluid, needs to be determined. Second, the resistance to this shear rate needs to be evaluated. This can be done either by measuring the amount of force required to move the source of shear rate, or by measuring the deflection on an object placed in the fluid, close to the source of shear rate. If the fluid being analysed is not Newtonian, then the apparatus will have to perform these tasks at several different shear rates. Once the resistance to the shear rate (i.e. the shear stress) has been determined at one or more known shear rates, the viscosity (or the components required to determine the apparent viscosity) can be derived.

Model 35 Viscometer The model 35 viscometer, produced either by Fann or Chandler, is the most common device used in the oil industry for determining viscosity and rheological properties. It is robust, easy to use and reliable. It can also be fairly easily calibrated, provided the user is familiar with the process. Figure 4.4a shows a photograph of a model 35 viscometer, whilst Figures 4.4b and 4.4c illustrate how it works;

Figure 4.4a – Chandler 35 viscometer. The position of the rotor is indicated (A), whilst the bob is hidden inside this. The cup (B) holds the test fluid, and is mounted on a support (C) that can move up and down as required.

A

B

C

Torsion Spring

Bob Shaft Rotor Fluid Bob Bob Shaft

Bob

Figure 4.4b – Cross-section through the rotor and bob on a model 35 viscometer

Page 23

Figure 4.4c – Schematic diagram showing the model 35 viscometer bob assembly


BJ Services’ Frac Manual 4. Fluid Mechanics The model 35 viscometer works by rotating the rotor (see Figure 4.4b) around the bob. The fluid is positioned in a narrow gap between the rotor and the bob. As the rotor spins, it produces a shear on the fluid, which in turn produces a drag force on the bob. The bob is mounted on a spring loaded bob shaft (see Figure 4.4c), so that as it experiences a drag force, it will rotate slightly. The greater the drag force, then more the bob rotates. Attached to the top end of the bob shaft is a dial indicator, allowing the operator to read how much the bob has rotated. As the bob deflection is directly related to the shear stress being experienced by the fluid, it is possible to use the dial reading as a measure of viscosity. Generally, the model 35 viscometer can spin the rotor at the following speeds, although these vary slightly from model to model. The speeds are 1, 2, 3, 6, 12, 20, 30, 60, 100, 200, 300 and 600 rpm. By plotting the rpm’s of the rotor (shear rate) against the dial reading (shear stress) it is possible to determine what type of fluid is being measured, by analysing the shape of the curve.

τ

= 0.01066 N θ .................................................................. (4.7)

γ

= 1.703 ω ......................................................................... (4.8)

where N is the spring factor of the torsion spring fitted to the model 35 viscometer (usually equal to 1), θ is the dial reading and ω is the speed of the rotor in rpm’s. It should be noted that Equation 4.8 is valid only for the R1 rotor and B1 bob combination – for other combinations refer to the manufacturer’s manual. By using Equations 4.7 and 4.8, a plot of shear rate against shear stress can be produced, or if necessary, a log-log plot. From these, the viscosity defining parameters can be derived.

Other Methods for Measuring Viscosity Various other methods for measuring viscosity are available; i)

Helical Screw Rheometer. Uses a helical screw inside a sleeve. The screw rotates and fluid flows up the inside of the sleeve and out of the top. The amount of force taken to rotate the screw is measured to produce the shear stress. The shear rate is derived from the speed of the screw. Used by some service companies for in-line real-time viscosity measurement during frac jobs.

ii)

Fann 50 HPHT Viscometer. Works on the same principle as the model 35 viscometer, but is designed so that the analysis can be carried out at high temperature and pressure. These viscometers are also usually remote controlled by a PC, allowing shear rate and temperature schedules to be used, as well as the recording of all data. Although quite expensive, these machines are commonly used for designing frac fluid systems. Figure 4.4d shows a Fann 50.

Figure 4.4d – Fann 50 high pressure, high temperature rheometer. This model is fully computer controlled, whereas earlier models had manual controls and were twice the size of the model shown.

Page 24


BJ Services’ Frac Manual 4. Fluid Mechanics iii)

Brookfield In-Line Viscometer. Viscometer designed to provide real time viscosity measurement for fluids flowing down a process line. This viscometer works on a similar principle to the model 35, although the rotor and bob are of a different size and shape.

iv)

Funnel Viscometer. A simple device for determining apparent viscosity. It consists of a funnel with a hole in the end. A specific volume of the fluid is placed in the funnel, and the time taken for it to drain out of the small hole in the bottom of the funnel is measured. A chart then provides a quick conversion from time to apparent viscosity.

The above are the most commonly used varieties in the oil industry, although it should be remembered that a wide variety of devices and methods are available.

4.5

Apparent Viscosity

The apparent viscosity of a fluid is the viscosity of the fluid at a specific shear rate. For a Newtonian fluid, the apparent viscosity is the same as the actual viscosity. For all other fluids, the apparent viscosity is the slope of a line on a shear rate vs shear stress curve, from the origin to the line, at a specific shear rate, as shown in Figure 4.5a:-

Shear Stress, τ

2 1

0

a

0

Shear Rate, γ

b

Figure 4.5a – Graph illustrating the change in apparent viscosity for a power law fluid at two different shear rates.

As can been seen in Figure 4.5a, for a shear thinning power law fluid, the apparent viscosity of the fluid (the slope of the two lines) decreases as the shear rate increases. At shear rate "a" the slope of line 1 (and hence the apparent viscosity) is greater than the slope of line 2 at the greater shear rate "b". Hence the fluid is said to be shear thinning. In practice, it is the apparent viscosity that is usually measured. The model 35 viscometer is set up so that at 300 rpm (with an R1 rotor, B1 bob and spring factor = 1), the apparatus reads apparent viscosity directly – no additional calculations are required. The apparent viscosity can be calculated as follows, for a power law fluid:-

µapp

Page 25

=

47879 K' ...................................................................... (4.9) 1-n'

γ



4.6

Flow Regimes and Reynold’s Number

Figure 4.6a illustrates the three different flow regimes that a fluid can experience, with plug flow being at the lowest fluid velocity, and turbulent flow being at the highest.

Plug

Laminar

Turbulent

Figure 4.6a – Diagram illustrating the three flow regimes

i)

Plug Flow. At low flow rates, the fluid flows with an almost uniform velocity profile. The fluid moves with a uniform front across almost the entire flow area.

ii)

Laminar Flow. As the flow rate increases, the velocity profile begins to change. Fluid close to the walls of the pipe (or duct, or fracture) flows slowest, whilst fluid in the center of the pipe flows fastest. Fluid velocity is a function of distance from the pipe wall. Also known as streamline flow.

iii)

Turbulent Flow. As the flow rate continues to increase, the contrast in velocity across the flow area becomes unsustainable, and the fluid breaks down into turbulent flow. This is characterised by a series of small scale eddies and whirls, all moving in the same overall direction.

The friction pressure produced by the fluid flow is highly dependent upon the flow regime. Therefore, it is important to be able to determine the flow regime.

Reynold’s Number The flow regime is found by using the Reynold’s number (NRe), as follows; 100 <

NRe NRe NRe

< 100 < 2000 > 2000

Plug Flow Laminar Flow Turbulent Flow

It should be remembered that these are very generalised numbers. The actual numbers can vary significantly, depending upon the circumstances. The Reynold’s number itself can be found from the following formula:NRe

Page 26

=

ρdv ........................................................................... (4.10) µ


BJ Services’ Frac Manual 4. Fluid Mechanics where ρ is the fluid density, d is the inside diameter of the pipe, v is the “bulk” fluid velocity along the pipe and µ is the viscosity. Equation 4.10 is for SI units, whilst Equation 4.11 is for field units; NRe

= 132,624

SG q ........................................................... (4.11) dµ

where SG is the specific gravity, q is the flow rate in bpm, d is the inside diameter in inches and µ is the viscosity in cp. Obviously Equations 4.10 and 4.11 only apply to Newtonian fluids, i.e. fluids with a constant viscosity. As stated before, Frac Engineers only rarely deal with Newtonian fluids, so below is Equation 4.11 converted for power law fluids; 2-n'

NRe

SG v = 15.49 K'(96/d)n' ......................................................... (4.12)

where v is the velocity in ft/sec. To make things easier, v can be easily found from the flow rate, q:v

q = 17.157 d2 .................................................................. (4.13)

with q in bpm and d in inches. Usually, when fracturing, it is best to keep abrasive fluids at flow rates below that needed for turbulent flow. This is to prevent the erosion of flow lines and the washing out of seals, caused by the action of the proppant. BJ Services'Standard Practices states that for abrasive fluids, the fluid velocity must be kept below 40 ft/sec.

4.7

Friction Pressure

One of the ultimate objectives of fluid mechanics - as far as the Frac Engineer is concerned, anyway – is to be able to predict the friction pressure (∆Pfrict) of the fluids that are being pumped. This is often very difficult, as fluid composition and temperature is constantly changing as the treatment progresses. In addition, two-phase (liquid and proppant) and even three-phase (liquid, proppant and gas) flow is common. Predicting fluid friction pressure is therefore an unreliable process and there really is no substitute for reliable bottom hole pressure data. Failing that, the next best option is to use friction pressure tables, such as BJ’s Fracturing Fluids – Friction Pressure Data. These tables are usually based on data generated by actually pumping the fluid around a flow loop, and so are based on a situation similar to the actual treatment process. Most modern fracture simulators incorporate data from these tests in their fluid models, so friction pressures predicted by these are also reasonably reliable (although not perfect, as the temperature of the wellbore is constantly changing) unless there is proppant in the fluid. Finally, when the three methods outlined above are not possible, the friction pressure may be calculated from fluid data, using the one of several available methods. The method outlined below, based on the use of Fanning friction factors, is fairly reliable (i.e. it is just as good as the data used as inputs), but is not intended for use in narrow diameter pipes at higher than normal flow rates (such as for coiled tubing treatments). ∆Pfrict

2

= 0.325

SG L v f ......................................................... (4.14) d

where L is the length of pipe in feet and f is the friction factor (dimensionless).

Page 27


BJ Services’ Frac Manual 4. Fluid Mechanics The friction factor is determined by using the Reynold’s number. For plug and laminar flow:f

16 = N ............................................................................. (4.15) Re

whilst for turbulent flow:f

0.0303 = N 0.1612 .................................................................... (4.16) Re

So the first step in the process of finding ∆Pfrict is to determine the Reynold’s number. Once that has been found, the friction factor can be determined, which in turn leads to the friction pressure.

References Howard, G.C., and Fast, C.R.: Hydraulic Fracturing, Monograph Series Vol 2, SPE, Dallas, Texas (1970). Gidley , J.L., et al.: Recent Advances in Hydraulic Fracturing, Monograph Series Vol 12, SPE, Richardson, Texas (1989). Economides, M.J., and Nolte, K.G.: Reservoir Stimulation, Schlumberger Educational Services, 1987. Economides, M.J.: A Practical Companion to Reservoir Stimulation, Elsevier, 1992 FracRT Version 4.6 User’s Manual, BJ Services, 1995 onwards Friction Pressure Manual, The Western Company, 1989 onwards Fracturing Fluids – Friction Pressure Data, BJ Services, 1983 onwards API Recommended Practice 39, Measuring the Properties of a Cross-Linked Water-Based rd Fracturing Fluid, 3 Edition, American Petroleum Institute, May 1998 Stimulation Engineering Support Manual, BJ Services, October 1996 onwards Standard Practices, BJ Services, 2000 onwards

Page 28


BJ Services’ Frac Manual 5. Fluid Systems

5.

Fluid Systems

The fracturing fluid is a vital part of the fracturing process. It is used to create the fracture, to carry the proppant into the fracture, and to suspend the proppant until the fracture closes. On a more basic level, the fluid system is the vehicle that allows us to transfer mechanical energy (in the frac pumps) into work performed on the formation. In order to carry out these tasks efficiently, the ideal fluid must have a combination of the following properties. i) ii) iii) iv) v) vi) vii) viii)

Low cost. Ease of use. Low tubing friction pressure. High viscosity in the fracture, to suspend the proppant. Low viscosity after the treatment, to allow easy recovery. Compatibility with the formation, the reservoir fluids and the proppant. Safe to use. Environmentally friendly.

Some of these properties are not easy to combine in the same fluid. Usually, the process of selecting a fracturing fluid is a trade off. It is up to the Engineers to decide which properties are most important and which properties can be sacrificed. In order to make this choice easier, there are a number of fluid systems available for fracturing.

5.1

Water-Based Linear Systems

The first fracturing fluid, used in Kansas in 1947, was gasoline gelled with war surplus napalm. Obviously this was a highly dangerous fluid, and it wasn’t long before water based systems were available. The first of these systems used starch as the gelling agent, but by the early 1960’s guar was introduced and soon became the most common polymer for fracturing. Today, polymers derived from the guar bean are used in most fracturing treatments - the other main source of polymers being cellulose and it' s derivatives. Before the dry polymer is added to the water, the individual molecules are tightly curled up on themselves. As the polymer molecule hydrates in water, it straightens out – which is why these fluids are referred to as linear gels – as illustrated in Figure 5.1a:-

A

B

Figure 5.1a – Hydration of polymer gels in water. 'A'shows a polymer molecule before hydration in water, whilst 'B'shows a polymer molecule after hydration in water.

It is these long, linear molecules that produce the increase in viscosity. However, it should be remembered that this hydration only occurs at a specific pH range. Outside this range, the Page 29


BJ Services’ Frac Manual 5. Fluid Systems hydration rate can be very slow and sometimes almost non-existent. Different polymers have different pH ranges, and buffers may have to be used to make the polymer hydrate. If a polymer that hydrates at a neutral pH is added to water, it may start to hydrate very rapidly. This leads to the formation of “clumps” of non-hydrated polymer, surrounded by partially hydrated polymer, surrounded in turn by hydrated polymer. These are known as fish-eyes and are a sign that the gel has been poorly mixed. Several techniques can be employed to prevent the formation of fish-eyes. i) ii) iii) iv)

Buffer the water so that the pH will prevent hydration. Once the polymer powder is thoroughly dispersed in the water, a different buffer is used to change the pH to a point where the polymer will hydrate. Add the polymer through a high shear device (such as a jet mixer) to ensure that the polymer does not form clumps. Circulate the hydrating gel through a high shear device, such as a choke, to break up any fish eyes. Slurry the polymer into a hydrocarbon-based fluid (such as diesel, kerosene or even methanol). The slurry is then added to the water, allowing the polymer to disperse before it hydrates.

A combination of these methods can also be used. Common polymers used for linear gels include:Starch Guar Hydroxypropyl Guar (HPG) Carboxymethyl Hydroxypropyl Guar (CMHPG) Carboxymethyl Guar (CMG) Cellulose Hydroxyethyl Cellulose (HEC) Carboxymethyl Hydroxyethyl Cellulose (CMHEC) Xanthan ® ® ® Xanthan derivatives (e.g. Bioxan , Xanvis , XC Polymer etc) The most commonly used polymers for fracturing are Guar, HPG and CMHPG, mostly as the basis for crosslinked systems (see below). HEC is probably the most widely used polymer for linear gel fracturing, due to its popularity for fracturing low temperature, high permeability formations. BJ’s range of water-based linear gel frac fluids includes the Aqua Frac system, which is based on guar and its derivatives. Gelling agents are GW-3, GW-4 & GW-27 (guar), GW-32 (HPG), GW-38 (CMHPG) and GW-45 (CMG). Also in BJ’s product range is the Terra Pack system, which is primarily designed for gravel packing, but can also be used for fracturing. The gelling agent for Terra Pack II is GW-21 (HEC) and for Terra Pack III is GW-22 (Xanthan).

5.2

Water-Based Crosslinked Systems

The majority of hydraulic fracturing treatments are carried out using water based crosslinked gels. These systems offer the best combination of low cost, ease of use, high viscosity and ease of fluid recovery. Generally, water based crosslinked gels will be used unless there is a specific reason not to use them – they are the default option. The starting point for a crosslinked system is a linear gel, as described above in Section 5.1. When used for crosslinked systems, linear gels are often referred to as base gels. The most commonly used linear gels are guar and its derivatives; HPG, CMG and CMHPG.

Page 30


BJ Services’ Frac Manual 5. Fluid Systems A crosslinked gel, as illustrated in Figure 5.2a, consists of a number of hydrated polymer molecules, which have been joined together by the crosslinking chemical. This series of chemical bonds between the polymer molecules greatly increases the viscosity of the system, sometimes by as much as 100 times. In order for an efficient crosslink to occur, two separate things need to happen. First, the base gel needs to be buffered to a pH which will allow the crosslinking chemical to work. Usually, this is at a different pH to that required for polymer hydration, so a different pH buffer has to be used. Secondly, the crosslinking radical needs to be present at sufficient concentration. If both these conditions occur, the gel will experience a dramatic increase in viscosity.

A

B

Figure 5.2a – A crosslinked polymer. ‘A’ shows the hydrated polymer prior to addition of the crosslinker. ‘B’ shows the crosslink chemical bonds between the polymer molecules.

Obviously, a fully crosslinked polymer is extremely viscous, and can result - under the wrong conditions - in a high level of fluid friction as it is pumped downhole. To counter this, it is quite common to use a delayed crosslinker. A delayed crosslinker can take anything up to 10 minutes before the gel is fully hydrated, depending upon the temperature, initial pH and shear that the fluid experiences. The ideal crosslink delay system would delay the onset of crosslink as long as possible, but would still have the fluid fully crosslinked by the time it reaches the perforations. The most commonly used crosslinking systems are as follows:Borates “Exotic” Borates Zirconates Aluminates Titanates

Zirconates

Aluminates

Organic Titanates

Borates

0

1

2

3

4

5

6

7

8

9

10

Figure 5.2b – pH ranges for crosslinkers (after SPE 37359)

Page 31


11

12

BJ Services’ Frac Manual 5. Fluid Systems Of these, the borates and “exotic” borates are by far the most commonly used, followed by the zirconates. Figure 5.2b illustrates the pH ranges of these crosslinkers, whilst Figure 5.2c shows their temperature ranges:Zirconates

Aluminates

Titanates High Temperature Borates Conventional Borates

100

150

200

250

Temperature,

300

350

400

oF

Figure 5.2c – Temperature range for crosslinkers (after SPE 37359)

All crosslinked gels tend to be shear thinning, which means that the apparent viscosity of the fluid decreases with shear rate. This is because the shear acts to break the crosslink bonds between the hydrated polymer molecules. Borate crosslink bonds will reconnect and produce a good quality gel after the shearing has taken place. However, zirconate bonds are much more shear sensitive and may not reconnect. Therefore, it is essential to consider the level of shear that a fluid will experience when selecting a crosslinker. Like most fracturing companies, BJ Services tends to classify its crosslinked fluids systems by the type of crosslinker used:Viking™ is a guar-based system that uses conventional borates for the crosslink. It is a cheap, easy to use fluid intended for low temperature applications. There is no crosslink delay. Crosslinkers used are XLW-4, XLW-32 or XLW-10. Viking D is the delayed crosslink version of Viking, and uses the crosslinkers XLW-30 or XLW-30A. ®

Spectra Frac G is probably the most commonly used of all BJ’s borate frac fluid systems. It is guar based, and uses an organo-borate crosslinker for a much greater temperature range than the Viking systems. The system is a premium system at lower temperatures, typically providing more viscosity. The crosslinker can be delayed, and the length of time for the delay can be varied over a significant range. The crosslinker for the system is XLW-24. ®

®

Spectra Frac G HT is the high temperature version of Spectra Frac G . It is guar based, and uses an organo-borate crosslinker for a much greater temperature range than the Viking systems. The crosslinker also employs a self-breaking mechanism, which helps to reduce the viscosity over a period of time above +/- 230°F. The crosslinker can be delayed, and the length of time for the delay can be varied over a significant range. The crosslinker for the system is XLW-56. Lightning™ is a new fracturing fluid system that uses a newly developed low-residue guar polymer, GW-3. The system uses the same borate crosslinkers as Viking™. ®

Medallion Frac is a CMHPG based system that uses a zirconate crosslinker. Unlike the ® borate systems, which operate at a pH above +/- 9.0, Medallion Frac operates at a pH below neutral, usually around 4.5 to 5.5. Because of its low pH, it is the fluid of choice for CO2 foam fracs, pads for acid fracs, and for formations which are sensitive to high pH’s. Page 32


BJ Services’ Frac Manual 5. Fluid Systems Crosslinkers for the system are XLW-41, XLW-53 or XLW-60. XLW-60 is a delayed crosslink, whilst XLW-41 and XLW-53 are designed for a rapid crosslink. The crosslinkers can be used together in varying proportions to adjust the crosslink time as desired. ®

®

Medallion Frac HT is a high pH version of Medallion Frac . It uses a different buffer to achieve the required pH (usually around 8.0 to 9.0), but otherwise is the same as Medallion ® Frac . The high pH zirconate system is more temperature stable than the low pH. Generally, the high pH system uses XLW-60 as the crosslinker. Vistar™ is a low or high pH, zirconate crosslinked system, designed so that only very low polymer loading is needed, as compared to other fluid systems. The base gel is a guarderivative (GW-45). Crosslinkers for the system are XLW-63 (lower temperatures) and XLW14 (high temperatures). Crosslinked systems are also characterized by the quantity of polymer used in the base gel. For instance, a “35 lb” system has the base gel mixed with 35 lbs of polymer in every 1000 ® gals of water. If this base gel were to be used in Spectra Frac G , the fluid system would be ® known as Spectra Frac G HT 3500.

LFC, XLFC, VSP and GLFC LFC (which stands for Liquid Frac Concentrate) and XLFC are slurried polymer systems, usually designed to carry 4 lbs of polymer in every gallon of slurry. Simply add the slurry to water and the base gel will form. Slurrying the polymer in an oil-based system helps disperse the polymer in the water (preventing fish-eyes) and is much easier to meter when hydrating gel on-the-fly. The liquid base for the slurry is usually diesel or a low toxicity diesel-derivative. However, LFC and XLFC systems have been developed that use vegetable or fish oil as the base liquid, although these hold reduced amounts of polymer per gallon. In addition to the base oil and polymer, LFC and XLFC also contain suspending agents to prevent settling during storage, dispersants to help mix the slurry and wetting agents to help the polymer hydrate quickly once the LFC or XLFC is added to water. A pH buffer can also be incorporated to help the polymer hydrate more quickly, especially at low temperatures. LFC-1, GLFC-1 and XLFC-1 contain guar (GW-27) LFC-2, GLFC-2 and XLFC-2 contain HPG (GW-32) LFC-3, GLFC-3 and XLFC-3 contain CMHPG (GW-38) XLFC-5 contains GW-3 XLFC is the updated version of LFC. VSP (or Vistar Slurried Polymer) is a version of XLFC developed for the Vistar™ system and contains CMG (GW-45). GLFC systems, which can be mixed with guar, HPG or CMHPG, use an organically-derived base oil in order to meet increasingly tight environmental regulations in many areas of the world.

5.3

Oil-Based Systems

As stated previously, the very first hydraulic fracture treatment was carried out using gasoline gelled with war surplus napalm. The operation was performed on Pan American Petroleum’s Klepper No 1 well, Grant county, Kansas, (part of the Hugoton gas field) in 1947. The treatment was aimed at 4 gas bearing limestone formations, at about 2500 ft. The gasolinebased fluid was selected, as it was perceived to be more compatible with the formation. This continues to be the primary reason for selecting an oil-based fluid. For the record, the treatment on Klepper No 1 failed to produce a significant production increase, and it was decided that the "Hydrafrac” process would never compete successfully with acidising in this type of formation.

Page 33


BJ Services’ Frac Manual 5. Fluid Systems The first widely-used oil-based fluid system, was based on the reaction of an acidic material (tallow fatty acid) and basic material (caustic) to form a polymeric salt, in a process similar to the manufacture of soap. These fluids provided viscosity, but where very unstable at elevated temperatures. As time progressed, this system was replaced by others based on the use of aluminium phosphates, which were able to provide significantly increased viscosity and more stability at elevated temperatures. In the early 70’s, the aluminium phosphate systems were replaced by the aluminium ester systems. The association of aluminium and phosphate esters is illustrated in Figure 5.3a. These systems used a combination of two products to produce the required viscosity. The relative ratio of these two products was extremely critical – so critical that it was difficult to mix these systems on the fly. Consequently, a great deal of time and effort was spent in pregelling tanks full of hydrocarbons, resulting in considerable spillage and waste of chemicals. R

R O

P

O

O

P

O O

O

H

Al

H

O

O

O R

P

O

O Al

O O

R O

O

Al

H

R

O

P

O O

H

R

Figure 5.3a – Aluminium phosphate association polymer

More recently, BJ Services has introduced a much more user-friendly system known as Super RheoGel. The ratios of the various components of the system are not nearly as critical, so that the gel can now be mixed on the fly. The following products are used in Super RheoGel:GO-64 (gelling agent) and XLO-5 (activator) are the main components of the system. They are added in equal quantities, at different stages of the blending procedure, to produce the required viscosity and stability. NE-110W is a critical surfactant blend used in the continuous mix gelled oil system. This material aids in fluid recovery by acting as a hydrotropic material in the system. It helps to reduce emulsion tendencies of oils and also acts as a long-term breaker for the system. NE110W also helps to counteract the oil-wetting surfactants contained in products such as diesel. GBO-5L, GBO-6 and GBO-9L are the breakers for the system. Most gelled oil systems can be prepared with a wide variety of hydrocarbon based fluids, including diesel, kerosene, “frac oil”, condensate and many lease crudes. Because the fluid used to fracture the well is itself hydrocarbon based, the well can be put straight on to production after the treatment is over. This makes the fluid recovery phase of the operations much easier. The Super RheoGel system does not work like a conventional water-based crosslink system. There is no base gel viscosity when the GO-64 is added, as it does not react with the base hydrocarbon. Instead, the GO-64 disperses in the hydrocarbon. When the XLO-5 is added, the crosslinker joins up the GO-64 molecules, trapping the hydrocarbon molecules within the GO-64/XLO-5 matrix and producing viscosity. Because the GO-64 does not react with the base hydrocarbon, it is possible to gel any fluid system in which this product can be dispersed, hence the ability of the system to be used in a wide variety of hydrocarbon-based fluids. Page 34


BJ Services’ Frac Manual 5. Fluid Systems Methanol can also be used as the base for fracturing fluids, although the systems designed for oil-based fluids (such as Super RheoGel) are not suitable. Instead, a polymer is used to produce a base gel and a specialised crosslinker is used to provide the viscosity necessary for proppant transport. Methanol-based fracturing fluids are used in water- and fluid-sensitive reservoirs where fluid recovery after the treatment is critical. The methanol reduces interfacial tension between the fracturing fluid and the connate water and also helps remove and prevent capillary water blocks. This allows for much easier recovery of the fracturing fluid from dry gas and watersensitive reservoirs. BJ Services'methanol-based fracturing fluid is called Methofrac XL. The system is designed for continuous or batch-mixed applications. GM-55 is the guar-derivative polymer powder, whilst XLFCM-1 is the slurried polymer concentrate. The crosslinker is XLW-40, which is a titanium-based, is usually diluted before use. The diluted versions (XLW-40B, -41A and -41B) are mixed by adding 2.5 to 10% of XLW-40 by volume to methanol or iso-propyl alcohol, as appropriate (see BJ Services'Mixing Manual instructions). The breaker for the Methofrac XL system is GBW-5. When mixing with lease crude or condensate, obtain fresh samples of the hydrocarbon and test to make sure that the system performs as required. This practice should also be followed when mixing with fluids such as kerosene or diesel, as local variations in product quality can have a significant effect on fluid performance. Additionally, be aware that BJ Services has strict safety and operations standards for the use of hydrocarbon based fluids, and for the handling of low flash point liquids. These standards can be found in BJ’s Standard Practices Manual and BJ’s Corporate Safety Standards and Procedures Manual.

5.4

Emulsions

In general, emulsions are only rarely used in fracturing operations, but in some parts of the world they have been found to have an ideal combination of fluid loss characteristics, formation compatibility and downhole viscosity. As a result, in these areas their use is common. Most of these systems are oil-in-water emulsions and operate in a similar fashion. Water is gelled with a standard gelling agent and held in a tank(s). During the job, water and oil are mixed together at the ratio of 2 parts oil to one part gel. An emulsifier is either pre-blended in the water phase (the gel) or added on the fly. The fluids very quickly form a brown emulsion, the viscosity of which is largely proportional to the initial viscosity of the water phase. Some systems require an external breaker in order to destroy the emulsion and allow the fluids to be recovered. However, in most systems, the emulsion quickly falls apart after exposure to the formation. BJ Services emulsion-based fluid system is known as Polyemulsion for which the emulsifying agent is E-2.

5.5

Visco-Elastic Surfactant Fluids

Visco-elastic surfactant (VES) systems are water-based fluids that employ a completely different method from all other water-based fluid systems for obtaining viscosity. They do not rely upon the hydration of a polymer. Instead, they use the unique properties of certain surfactants when mixed at certain concentrations in brine-based fluids. In aqueous fluids, surfactants will tend to expel their lipophilic (water-repelling) tails out from the surface of the fluid. As the concentration of the surfactant increases, close packing occurs Page 35


BJ Services’ Frac Manual 5. Fluid Systems and no more surfactant molecules can expel their tails. At this point, the surfactant molecules will start to form spherical aggregates (or micelles) with the lipophilic tail facing inwards, and the hydrophilic head facing outwards. The concentration at which these micelles start to form is called the critical micellar concentration (CMC), and is often around 0.5% by volume of surfactant. The CMC will decrease as the molecular weight of the surfactant increases. As the surfactant concentration increases further, and in the presence of a suitable counter ion (such as those produced by brines), these micelles can come together to form worm- or rod-shaped aggregates or micelles. It is these rod-shaped micelles that impart viscosity to the water. VES fluids have some rather unique properties, as follows:1. 2. 3. 4. 5.

VES fluids are extremely shear thinning, with the property to quickly re-heal after the shear is removed. This means that the fluids have an extremely low friction pressure, whilst at the same time retaining excellent proppant transport characteristics. VES fluids are very easy to mix. Simply start with the base brine and add the surfactant on the fly. VES fluids can be made to be very environmentally friendly, depending upon the combination of surfactant and brine used. VES fluids often require no breaker system, as micelles can be disrupted by changes in pH, high temperatures, dispersion in formation waters or by contact with hydrocarbons. The VES system is as formation and proppant pack friendly as the base brine used to mix it. The systems contain no polymers, and therefore produce no polymer residue. Therefore, these fluids are capable of providing zero formation damage and 100% regained proppant pack permeability.

The two main disadvantages of VES fluids are that they are relatively expensive and that they are limited by temperature. Proppant transport characteristics are rapidly lost above temperatures of +/- 230°F. Development work continues, however. Another problem with VES fluids is leakoff. Because they contain no polymers, they do not have any wall-building characteristics, and so leakoff control is entirely dependent upon the fluid’s viscosity and/or additives used in the system. BJ Services’ has two VES fluid systems, called ElastraFrac and AquaClear. ElastraFrac uses the surfactant MA-1. The system uses potassium, ammonium or magnesium chloride as the base brine, although more exotic phosphate-based brines are used at temperatures above +/- 200°F. The surfactant used is anionic (unlike competitor’s products that use cationics and hence risk oil-wetting the formation). AquaClear is also a surfactant-based fracturing fluid. The system uses either a combination of FAC-1W and FAC-2, or the single surfactant FAC-3W as the VES. It is designed for mixing on-the-fly and is suitable for use up to +/- 250 °F. The system is easily used as an energised fluid and does not need an additional foaming agent).

5.6

Energised Fracturing Fluids

Energised fluids consist of a liquid phase – usually a water-based linear or crosslinked gel – and a gaseous phase, which is typically N2, CO2 or a combination of these. Such treatments involve large amounts of equipment and personnel. Consequently, they are relatively expensive. These treatments are also referred to a foam fracs, as foam is generally what is arriving at the formation. Because of the safety implications of working with both cryogenic fluids and energised fluids, the procedures detailed in BJ’s Standard Practices Manual and BJ’s Corporate Safety Standards and Procedures Manual, should be closely followed at all times.

Page 36


BJ Services’ Frac Manual 5. Fluid Systems Foamed fluids have several unique properties that make them advantageous under certain circumstances:i) ii) iii) iv)

Viscosity and proppant transport. Stable foams have a comparatively high viscosity and make excellent fluids for carrying and suspending proppant. Foams have very good leakoff properties. This is due to the multi-phase flow effects as the foam tries to move through the formation' s porosity. Because foams are typically only 30 to 40% liquid, they are more compatible with water sensitive formations than frac systems which are 100% liquid. The extra energy stored in the fluid, coupled with the low hydrostatic head of the foam, makes fluid recovery relatively easy.

Foam Quality

Proppant Transport

The foam quality, often expressed as a percentage or just simply as a quality (i.e. “70 quality” or even “70Q”) is the percentage of foam or energized fluid that is gas, at the anticipated bottom hole conditions. In order to design a foam treatment, an Engineer must have a reasonable idea of the expected bottom hole treating pressure and temperature, as the volume occupied by the gas phase will vary depending on both of these (although the temperature is much less significant than the pressure). As illustrated by Figure 5.6a, foam viscosity (and hence it’s ability to transport proppant) is heavily influenced by the quality. If the bottom hole pressure is significantly less than anticipated, the foam quality will be too high, and the gas phase will expand to make a mist, rather than foam.

STABLE FOAM

0

20

40

60

80

100

Foam Quality Figure 5.6a – Proppant transport as a function of foam quality. This graph is a combination of the work performed by several individuals and organisations. It is intended as a qualitative illustration of the effect foam quality has on the ability of the fracturing foam to transport and suspend proppant.

Gas assisted fluids use lower gas quality (typically 20 to 40%) than foamed fluids. The main purpose of the gas phase is to reduce hydrostatic head and hence aid fluid recovery. In such treatments, the proppant transport and fluid leakoff properties for a fully foamed fluid system are not required.

Proppant Concentration Because proppant is added to the liquid phase of the foamed frac fluid, there is a limit to the overall proppant concentration that can be achieved downhole. Because it is not possible to blend and pump proppant at more than 18 or 19 ppg in the liquid phase, by the time the liquid phase has been mixed with the gaseous phase, the overall proppant concentration has been Page 37


BJ Services’ Frac Manual 5. Fluid Systems reduced to 7 or 8 ppg. For this reason, it is not possible to place the very high proppant concentrations required for fracturing high permeability formations. This means that foam fracturing is limited to medium and low permeability reservoirs, for skin bypass fracturing (although the extra cost can defeat the low cost objectives of this type of treatment) and for coal bed methane fracturing.

Constant Internal Phase vs Constant External Phase Foams can be thought of as being a multi-phase fluid, with a gas-internal phase, and a liquid external phase. The difficulty comes in deciding whether or not the proppant is part of internal phase or the external phase. The traditional method of modeling foams and designing treatment schedules uses the constant external phase method. This assumes that the proppant is part of the external phase. It is easier to operate on location, as both the slurry rate and the gas rate remain constant. However, under constant external phase, the actual fraction of the foam that is liquid can be severely reduced as higher proppant concentrations are reached. Obviously, the proppant has no properties that act to hold the foam together, so foams can become very unstable as the proppant concentration increases. The modern way of modeling foam is to use the constant internal phase method. This models the proppant as being part of the gas phase. Therefore, in order to keep foam quality constant, the gas rate has to go down as the proppant concentration rises, and then increase rapidly as the treatment goes to flush. This method is harder operationally, but provides much more stable foam.

Foam Stability The stability of foam is its ability to remain as foam, rather than separating out into two or even three phases. Ideally, the fluid should remain as foam long enough for the fluid to be recovered as foam after the treatment. Obviously, temperature and fluid contamination will act to reduce foam stability. There are three main methods for maximising foam stability:i)

ii) iii)

Mixing the liquid and gas phases at high shear, such as with a foam generator, or by passing the mixed phases through a high shear device, such as a choke. The greater the shear that the foam experiences, the more stable it becomes. High shear acts to reduce the average size of the gas bubbles, which in turn makes it harder for then to separate out. Crosslinking the fracturing fluid after the foam has been formed. By using a delayed crosslinker, the onset of crosslink can be timed to take place after the foam has been generated, so that the gas bubbles are literally crosslinked into position. Foaming agents. These surfactants act to increase the surface tension of a material, so that the gas bubbles are much more stable.

Often a combination of these methods is used.

Foam Viscosity The viscosity, proppant transport characteristics, fluid leakoff and stability of the foam are all influenced by the same foam characteristics - the liquid phase viscosity, the average gas bubble size, the foam quality and the surface tension properties of the liquid phase. All of these are affected by temperature and two of these are significantly affected by pressure. This means that calculating the viscosity – and hence the friction pressure and fluid leakoff – of the foam is very difficult. Consequently, calculated bottom hole treating pressures for foam fracs are extremely unreliable and should not be used for analysis unless there is absolutely no alternative whatsoever. The results from such an analysis should be considered as educated guesswork only. Page 38



N2 Foam Fracs N2 foamed fracs are the most straightforward of all the types of energized fluid fracs performed. Nitrogen is stored as a cryogenic liquid, in specialised, highly insulated tanks on location. Prior to the treatments, each tank uses a heat exchanger to vapourise a small amount of the liquid into gas. This has the effect of pressuring up the tank, so that liquid nitrogen is forced from the tank to the N2 pumpers. Before liquid N2 can be pumped, the pump itself has to be cooled down. This is done by flowing liquid N2 though the pump and out of a vent. Initially gas will bleed out if the vent. Eventually, as the unit cools down, liquid will be seen coming out of then vent, indicating to the operator that the unit is now ready to pump. Therefore, when designing N2 foam fracs, sufficient liquid nitrogen should be on location for cooling the N2 pumpers down at least 3 times (once for the minifrac, once for the main treatment and one spare). It is much easier to convert a liquid from low to high pressure, than it is to convert a gas from low to high pressure. Consequently, the N2 pumpers will be working on liquid N2 that is stored and pumped at around –320°F. This means that specialised equipment is required for pumping this cryogenic liquid. These pumpers also include a vapouriser, which will heat the high pressure liquid and convert it into a gas (for this reason, N2 pumpers are often referred to as “converters”). These vapourisers can be diesel fired or run from the engine coolant. As N2 is chemically inert, there are no limitations on the fluid systems it can be used with.

CO2 Foam Fracs CO2 has a number of properties that make its use significantly different from N2. To start with, liquid CO2 is stored at –20°F. The much higher temperature means that the liquid can be pumped with a standard frac pumper (provided they have been specially prepared – see BJ’s Standard Practices Manual). It also means that the liquid CO2 does not have to be converted into a gas before it is mixed with the liquid phase – this will happen automatically as the CO2 heats up. The second major property difference of CO2 is its tendency to form a solid (“dry ice”) if stored or pumped under the wrong conditions. Obviously, this must be avoided. Dry ice will only form below +/- 80 psi. So at every stage, the liquid CO2 is kept well above this pressure. Typically, CO2 is stored at between 150 to 300 psi. There are several different methods for pumping the liquid CO2 from the tanks to the pumpers. One method involves forcing it out with N2 pressure applied above the fluid level in the CO2 tank. Another method employs specialised boost pumps. Yet another method employs a combination of these two systems. Once again, BJ’s Standard Practices Manual should be consulted before designing any treatments. The third major difference is that unlike N2, CO2 is not chemically inert. Specifically, on contact with water based fluids, some of the CO2 will dissolve into the water to form an acid. This has the effect of lowering the pH of the system. This means that CO2 is not compatible with high pH fracturing fluids, such as borate crosslinked gels.

Binary Fracs Binary Fracs involve the use of a mixture of both CO2 and N2 to provide the foam. They were originally developed as a method of getting around one service company’s patent on CO2 foam fracturing. Since then, the method has been extensively developed and is now the preferred method of foam fracturing for many operating companies. Binary fracs are the most complicated stimulation operations performed, requiring the use of no less than three service supervisors (one for the CO2, one for the N2 and one for the liquid phase, who is in overall control). Consequently, these are relatively uncommon.

Page 39



Poly CO2 Poly CO2 is a highly specialised fluid developed by Nowsco in Canada. In this fluid, a specialised additive is mixed into the water-based liquid phase, which causes the waterbased gel and the liquid CO2 to form an emulsion, rather than foam. The emulsion is not particularly stable, and will break down after the fluid contacts the formation. This fluid system has only ever been used in low temperature applications, and it is unclear as to whether the stimulation benefits come from the placing of proppant, or from the thermal shock experienced by the formation. However, in certain formations it has proved to be highly successful.

5.7

Additives

There are an enormous number of additives used in the preparation of the various types of fracturing fluids, and an exhaustive list is beyond the scope of this manual. However, below is a description of the general types of additive, together with the most commonly used examples from BJ’s product range.

Gelling Agents Water-based gelling agents are designed to increase the viscosity of water. This water can be fresh (rarely), 2% KCl, 3% NH4Cl, seawater or any of a myriad of different kinds of brines. Nearly all the gelling agents are some kind of polymer. A wide range is available, depending upon hydration pH, temperature stability and polymer residue:Guar High-yield guar Hydroxypropyl Guar (HPG) Carboxymethyl Hydroxypropyl Guar (CMHPG) Carboxymethyl Guar (CMG) Hydroxyethyl Cellulose (HEC) Carboxymethyl Hydroxyethyl Cellulose (CMHEC) Xanthan Polysaccharide

GW-4, GW-27 GW-3 GW-32 GW-38 GW-45 GW-21, GW-24L, AG-21R GW-28 GW-22, GW-22L, GW-37 GW-23

Oil-based gelling agents are designed to increase the viscosity of oil-based fluids. These gelling agents work on a wide variety of hydrocarbons, but are primarily designed for diesel and kerosene. Any other hydrocarbon fluid should be tested prior to application. GO-64 GM-55

Gelling agent for Super Rheo Gel Gelling agent for Metho Frac XL

Crosslinkers and Complexers Crosslinkers and complexers are designed to dramatically increase the viscosity of an already gelled fluid, so that high viscosity can be maintained for extended periods of time at high temperatures. For many fluid systems, the crosslinker is the chemical that really defines its characteristics. XLW-4, XLW-32, XLW-10 XLW-30, XLW-30A XLW-14, XLW-63 XLW-24 XLW-56 XLW-60 XLW-40B, XLW-41A, -41B XLO-5 Page 40

Crosslinkers for Viking and Lightning Crosslinkers for Viking D Crosslinkers for Vistar Crosslinker for SpectraFrac G Crosslinker for SpectraFrac G HT Crosslinker for Medallion Frac & Medallion Frac HT Crosslinker for Metho Frac XL Complexer for Super RheoGel



Breakers Breakers are designed to reduce the viscosity of the fracturing fluid to a minimum, so that the fluid can be easily recovered after the treatment. They are also designed to minimise polymer residues, so that damage to the proppant pack is minimised. GBW-5, GBW-7, GBW-41L GBW-23, GBW-24 GBW-26C GBW-12CD GBW-14C High Perm CRB GBO-5L, GBO-6, GBO-9L

Oxidizing breakers Delayed oxidizers Enzyme breakers for cellulose + derivatives Enzyme breaker for guar + derivatives Enzyme breaker for xanthan + derivatives Encapsulated oxidizing breaker Breakers for Super RheoGel

Buffers Buffers are designed to either raise the pH or lower the pH, as required. Low pH buffers High pH buffers

BF-1, BF-10L, BF-10LE BF-7, BF-7L, BF-8L, BF-9L, caustic soda

Surfactants The word Surfactant comes from the phrase SURFace ACTive AgeNT, and includes any chemical that affects the interface properties between materials. Because this covers such a wide range of materials, it is necessary to discuss this group of products in more detail. Surfactants can also be grouped according to the type of charge they possess, so that some surfactants are anionic (negative charge), some are cationic (positive charge), some are amphoteric (cationic at low pH and anionic at high pH), some are Zwitterionic (both cationic and anionic simultaneously) and some are non-ionic. Generally speaking, it is best not to mix anionic and cationic products together, as they might form viscous deposits. Details of this can be found in BJ’s Mixing Manual. Most of BJ’s surfactant products are designed to leave the formation water wet. This means that the relative permeability of the formation to water has been lowered, and the relative permeability of the formation to oil has been raised. However, it is important to note the following:Cationic surfactants will leave sandstones oil wet and carbonates water wet Anionic surfactants will leave sandstones water wet and carbonates oil wet. Amphoteric surfactants can behave either way depending upon the pH. At acidic pH’s (less than 7), amphoteric surfactants show cationic properties, whilst at alkaline pH’s (greater than 7), they display anionic properties. At neutral pH, they behave like non-ionic surfactants. Non-emulsifying surfactants are designed to prevent the formation of emulsions between the crude oil in the formation and the treatment fluid. All water-based treatments should have a non-emulsifying surfactant added to them, unless they are being pumped into a water injection well or dry gas reservoir with no trace of condensate. Inflo 100, Inflo 102 NE-13 NE-110W NE-118 NE-940

Blend of cationic and nonionic Blend of cationic and nonionic Anionic Nonionic Nonionic

Note that some non-emulsifiers will also act to break existing emulsions.

Page 41


BJ Services’ Frac Manual 5. Fluid Systems Foaming agents work by increasing the surface tension of the fluid. This helps increase foam stability. Most foaming agents also acts as detergents and dispersants FAW-4 FAW-18W FAW-20 FAW-21 FAO-25

Anionic Anionic Anionic Amphoteric Nonionic – foaming agent for oil-based fluids

Note that FAW-4, FAW-18W and FAW-20 will leave carbonate formations oil wet. Low surface tension modifiers act to reduce the surface tension of the fluid. This helps the fluid penetrate into very small places, such as the pore spaces in low permeability reservoirs. These products also help the treatment fluid flow back out of the well after the treatment is finished. Flo-Back-20, Flo-Back-30 Inflo-100 Inflo-150

Nonionic Blend of cationic and nonionic Nonionic

Mutual solvents will dissolve hydrocarbon based deposits and allow them to disperse water based fluids. US-2, US-40 Inflo-40

Nonionic Nonionic

Emulsifiers are used to deliberately create emulsions. They only should be used as part of an emulsion-based fluid system AE-7 E-2

Cationic Cationic

Biocides Biocides, also known as Bactericides, are designed to kill bacteria. Any bacteria – especially sulphate reducing bacteria – will eat the polymer used in frac fluids. A colony of bacteria can reduce a tank of good quality gel into foul-smelling slick water in less than an hour. Biocides are used to prevent this. Initially, all tanks used for frac fluids should be as clean as possible. This will help reduce the risk of bacterial contamination. However, the water used to mix the gel can still contain these bacteria, especially if the climate is hot or seawater is being used. The biocide should be added either directly to the tank before the water is added, or it should be thoroughly mixed into the water prior to the addition of any polymer. Once the biocide has been added, it will quickly kill any bacteria that are present in the water. It is recommended that a biocide is used on any treatment with involves pre-gelling the fluid. It should be remembered that biocides are designed to prevent a colony of bacteria from developing in the first place, rather than for killing an existing colony - any gel that is suspected of being contaminated should be discarded, and its tank thoroughly cleaned. In order to break down the gel, bacteria secrete enzymes (similar enzymes to the breakers described above). These enzymes will cause a tank of gel to degrade, so that even if all the bacteria in a tank have been killed, their enzymes are still present in the tank. This is why contaminated tanks of gel need to be discarded, and not used again. It should also be noted that in their concentrated form, biocides are very dangerous materials (after all, they are designed for killing living things) and should be handled with extreme care. Magnacide 575 XCide 102, 207

Page 42



Gel Stabilisers Gel stabilisers are used to prolong the viscosity of crosslinked gels at high temperatures. They work by one of two methods:- they can scavenge the oxygen in the fluid; or they can chelate cations which can contribute to the degradation of the gel. GS-1, GS-1L, GS-9 Methanol

Clay Control Additives Clay control additives are used to prevent the swelling, migration and disintegration of clay minerals such as illite, smectite, chlorite and montmorillonite. Fresh water by itself will cause these problems. The addition of chloride ions to fresh water will prevent these problems in most formations, so that most treatments carried out with seawater do not need any additional clay stabilisers. However, exceptionally water sensitive formations may need additional protection, which is where BJ’s range of synthetic clay control additives is applied. KCl, NH4Cl, NaCl etc CaBr2, ZnBr2, etc Clay Treat 3C Clatrol Claymaster 5C, FSP

standard salts for brines specialised salts for high density completion brines (some of these may be incompatible with BJ’s crosslinked fluids). KCl substitute, recommended for Vistar.

Note that any salts containing calcium or magnesium should not be mixed with frac fluids, as these are incompatible with some crosslinkers. Also note that some of the synthetic clay control additives are cationic in nature and should not be mixed with any anionic products.

Fluid Loss Control Fluid loss control additives can be used for two main reasons; firstly, to lower a very high matrix leak off rate; and secondly, to prevent fluid loss down natural fractures. The use of fluid loss additives is becoming less and less common, as the understanding of fluid leakoff increases. Most Engineers also believe that pumping more fluid is preferable to using additives that can potentially produce permanent damage. Silica flour, 100 mesh sand 5% diesel ® Adomite Regain

Used for blocking natural fractures

References BJ Services’ Mixing Manual BJ Services’ Stimulation Engineering Support Manual BJ Services’ Products and Services Manual BJ Services’ Product Bulletins BJ Services’ Standard Practices Manual BJ Services’ Corporate Safety Standards and Procedures Manual

Page 43


BJ Services’ Frac Manual 5. Fluid Systems Rae, P., and Di Lullo, G.: “Fracturing Fluids and Breaker Systems – A Review of State-of-theArt”, paper SPE 37359, presented at the SPE Eastern Regional Meeting, Colombus OH, Oct 1996. Brannon, H.D., and Ault, M.C.: “New, Delayed Borate-Crosslinked Fluid Provides Improved Conductivity in High Temperature Applications”, paper SPE 22838, presented at the SPE Annual Technical Conference and Exhibition, Dallas TX, Oct 1991. Cramer, D.D., Dawson, J., and Ouabdesselam, M.: “An Improved Gelled Oil System for High Temperature Fracturing Applications”, paper SPE 21859, presented at the Rocky Mountain Regional Meeting and Low-Permeability Reservoirs Symposium, Denver CO, Apr 1991. Blauer, R.E., and Kohlhaas, C.A.: “Formation Fracturing with Foam”, paper SPE 5003, th presented at the 49 Annual Fall Meeting of the SPE, Houston TX, Oct 1974. Grundman, S.R., and Lord, D.L.: “Foam Stimulation”, paper SPE 9754, JPT pp 597 – 602, Mar 1983 Valkó, P., and Economides, M.J.: “Foam Proppant Transport”, paper SPE 27897, presented at the SPE Western Regional Meeting, Long Beach CA, Mar 1994. Tjon-Joe-Pin, R, DeVine, C.S., and Carr, M.: “Cost Effective Method for Improving Permeability in Damaged Wells”, paper SPE 39784, presented at the SPE Permian Basin Oil and Gas Recovery Conference, Mar 1998. Di Lullo, G., Ahmad, A., and Rae, P.: “Towards Zero Damage: New Fluid Points the Way”, paper SPE 69453, presented at the SPE 2001 Latin American and Caribbean Petroleum Engineering Conference, Buenos Aires, Argentina, March 2001.

Page 44


BJ Services’ Frac Manual 6. Proppants

6.

Proppants

The word proppant comes from the abbreviation of two words - “propping agent”. Proppants are granular materials, which are placed inside the fracture in order to “prop” the fracture open as the pressure falls below closure. The conductivity of the fracture is directly related to the quantity of proppant within the fracture, the type of proppant, the producing conditions and the size of the proppant grains. The purpose of hydraulic fracturing is to place the right amount of the right kind of proppant in the right place. When this is done correctly, the well is effectively stimulated.

6.1

Proppant Pack Permeability and Fracture Conductivity

As discussed in Section 2, one of the major factors affecting post-treatment well performance is the fracture conductivity. This is the product of the proppant pack permeability and the width of the fracture. In other words, the fracture conductivity is a function of the type of material holding the fracture open and the amount of this material within the fracture. The permeability of the proppant pack is controlled by several factors:i)

Proppant Substrate. The material that the proppant is made from obviously has a big effect on the permeability of the proppant pack. Some materials are stronger than others and are better able to withstand the enormous forces trying to crush the proppant as the fracture closes. The weaker the material, the more the proppant grain will deform. Proppant deformation reduces the porosity of the pack and reduces the overall fracture width. The more brittle the proppant is, the more likely it is that the proppant will produce fines as the grains are pushed together in a series of point to point contacts. Any fines will significantly reduce the proppant pack permeability.

ii)

Proppant Grain Size Distribution. A normal sedimentary formation has a wide variety of grain sizes, depending upon how well “sorted” the individual rock grains are. In general, any sandstone will be a mixture of small, medium and large grains. The mixture of grain sizes acts to reduce the formation' s permeability and porosity, as the smaller grains will occupy the pore spaces between the larger grains and will also tend to plug up the pore throats. However, if a set of particles are of almost identical size, then there will be no fines to block up the pore spaces and pore throats, so that the porosity (and hence the permeability) are maximised. This is why proppants are generally produced within a specific grain size distribution. This uniformity of grain size is one of the main reasons why proppant is usually several orders of magnitude more permeable than the formation, and also one of the main reasons why so much effort is spent in ensuring this uniformity of size. This is illustrated in Figure 6.1a, below;

Uniform

Natural

Figure 6.1a – The effect of uniform and natural grain size distribution on porosity

Page 45


BJ Services’ Frac Manual 6. Proppants Proppants are supplied within a specific grain size range. This grain size refers to the size of sieve used to sort the proppant. For instance, 20/40 size means that the vast majority of the proppant will fit through a size 20 sieve (20 holes per square inch), but will not fit through a size 40 sieve (40 holes per square inch). This is sometimes confusing, as larger grain sizes correspond to smaller mesh numbers. Common proppant sizes are 8/12, 12/20, 16/30, 20/40 and 40/60, although theoretically any combination of sizes can be produced. iii)

Average Proppant Grain Size. Generally, the larger the average proppant grain size is, the higher the permeability of the proppant (provided the grain size distribution is reasonably uniform). This is because larger grains produce larger pore spaces and pore throats, allowing an increased flow rate for a similar porosity. However, the larger grains are more susceptible to producing permeability reducing fines than are the smaller grain sizes. This is because larger grains distribute the closure pressure across fewer grain-to-grain points of contact and so the point contact loads tend to be greater. This is illustrated in Figure 6.1b;

Figure 6.1b – Diagram illustrating how larger grains have larger pore spaces and hence greater permeability.

iv)

Sphericity and Roundness. These quantities define how spherical the proppant grains are and how many sudden, sharp edges the grains have. Obviously, the smoother and more spherical the proppant grain is, the higher the pack permeability. There are standard API procedures for checking these quantities, but unfortunately they rely on some subjective analysis. Consequently, it is often difficult to see a clear trend between one proppant type and another. However, in general, artificial proppants will have better sphericity and roundness than naturally occurring types. This is illustrated in Figure 6.1c;

Figure 6.1c – Diagram illustrating the difference between a proppant with good sphericity and roundness (left), and a proppant with poor sphericity and roundness (right).

Coarse, angular grains also tend to produce more fines, as corners and edges tend to get broken off as compressive stress is applied. Therefore, proppants with good sphericity and roundness also tend to retain greater permeability at high stresses. In addition, because proppant with low sphericity and roundness will produce a more convoluted flow path for the produced fluids, non-Darcy pressure losses tend to be

Page 46


BJ Services’ Frac Manual 6. Proppants greater in these materials (see Section 10.9), leading to decreased effective proppant pack permeability. iv)

Frac Fluid Quality. The amount of residue left by the fracturing fluid can also have a big influence on the permeability of the proppant pack. In order to assess the effect of these fluids, a quantity called Regained Permeability is measured. Put simply, a sample of the proppant is put into a load sell and is subjected to a closure pressure, at an elevated temperature. A standard, non-damaging fluid is then flowed through the test cell. By analysing the pressure drop and flow rate, the permeability of the pack can be calculated. Next, the frac fluid is flowed through the test cell, and allowed to remain there for a specific time, during which it is designed to break. Once the fluid has broken, the permeability of the pack is measured again, by the same method as before. The two permeabilities are compared and the result (the regained permeability) is given as the percentage of the original permeability that remains after the test. Figure 6.1d, below, illustrates the difference between fluids with a high and low regained permeability;

Figure 6.1d – Three SEM micrographs showing the effects of frac fluid residue. The micrograph on the left shows undamaged proppant before the addition of the frac fluid. The center micrograph shows the residue left by a poorly designed crosslinked system. The final micrograph shows the same proppant pack after an enzyme breaker has been used.

Proppant packs can lose significant proportions of their permeability to fluid damage. Cheap, poorly designed fluids can cause regained permeabilities to be as low as only 30% or even less, whereas the state-of-the-art fluids can produce values in excess of 90%. v)

Closure Stress. As the proppant is crushed by the closure of the formation, it will start to produce fines. As discussed above, these fines will reduce the permeability of the pack. The stronger the proppant, the fewer fines are produced - nevertheless, all proppant types experience a decrease in permeability as closure stress increases, to a greater or lesser extent. In addition, most proppants also have a “maximum” stress, above which whole-scale disintegration of the proppant substrate starts to occur, rather than simple fines production. At this point, pack permeability falls dramatically. It should be noted that the reservoir pressure has an influence on the closure stress experienced by the proppant. This phenomenon is discussed in greater detail later in this manual (see Section 7.6). The relationship between reservoir pressure and closure pressure is dependent upon a number of factors - there are circumstances under which a decrease in reservoir pressure can result in an increase in closure stress. Additionally, there can be localised areas of low reservoir pressure (such as near the wellbore during drawdown) where once again the proppant experiences higher closure pressure. This potential increase in stress with the life of the well must be allowed for when selecting a proppant.

Page 47


BJ Services’ Frac Manual 6. Proppants vi)

Non-Darcy Flow. This effect will be discussed on more detail in Section 10.9. However, as the flow rate through the proppant pack increases, the pressure drop will increase at a rate faster than that predicted by Darcy’s law. This is due to the effects of inertial energy loses, as the fluid rapidly changes direction as it moves through the pore spaces. As the fluid velocity increases, the pressure drop due to inertial flow effects increases with the square of the velocity. So at low flow rates, (such as in a reservoir rock), non-Darcy effects can safely be ignored, whilst at high rates (such as in a proppant pack), the effective proppant permeability has to be reduced to reflect this effect. The phenomenon is particularly significant in high rate gas completions.

vii)

Multi-Phase Flow. Multi-phase flow has a similar effect upon proppant pack permeability as it does on formation permeability. It reduces it, by an amount that is dependent upon the absolute permeability, and the relative saturation of each phase. As it is very rare for a reservoir to produce a single phase (with the exception of some gas reservoirs), it is also very rare for proppant to conduct only a single phase. Therefore, the actual effective permeability of the proppant pack may be significantly less than the published data, which is generally produced for single-phase flow only (although this situation is improving).

6.2

Proppant Selection

As illustrated in Section 6.1, there are a substantial number of variables that must be taken into account when selecting proppant. However, in many cases the selection process has been simplified. All proppant suppliers and manufactures publish data for pack permeability against closure stress, for all their proppant types and grain size distributions. Provided the closure stress is known (taking into account any subsequent loss in reservoir pressure), the absolute permeability of the proppant pack can be easily found. This eliminates the need for the Frac Engineer to hold data on sphericity, roundness, crush resistance, grain size distribution, substrate material etc. Simply look up the proppant you are interested in, and see what the permeability is for a given closure stress. Most fracture simulators already have this data for most major proppant types. This allows the simulator to predict the fracture conductivity for most given proppant/closure stress combinations. Usually, there is also a “proppant damage factor”, which allows the user to simulate the regained permeability effects of the fracturing fluids. Some - but not all - fracture simulators will also model the effects of non-Darcy flow, showing a decrease in effective permeability as production rate rises. However, no current fracture simulators allow for the effects of multi-phase flow. Data on this has been published by a few sources, the most notable of these being the Stim-Lab Consortium’s PredictK software and Carboceramics' FracFlow proppant permeability simulator. Table 6.2a gives guidelines as the maximum closure stress each of the major proppant types can withstand, before substrate failure begins to occur. Obviously, these limits are very generalised, and are highly dependent upon factors such as grain size and the quality of the manufacturing process and/or source of sand. More detailed information is available from the manufacturers or in the references;

Page 48



Type Frac Sand Low Density Ceramics Intermediate Density Ceramics Sintered Bauxite

Maximum Closure Stress, psi 5,000 9,000 12,000 14,000

Product Example Brady, Ottawa, Colorado CarboEconoprop, CarboLite, ValueProp CarboProp, InterProp Carbo HSP, Bauxite

Table 6.2a – Generalised maximum closure stresses for the main proppant types.

Important Note The quality of the proppant, and the subsequent conductivity of the fracture, has a bigger effect on post treatment production than virtually anything else under the Frac Engineer’s control. In most cases, an economy made on proppant selection is a false economy. For instance, although low-density ceramics cost two to three times as much as frac sand, they have four to five times the pack permeability - even at low closure stresses - due to their high sphericity and roundness.

Resin-Coated Proppant Many operating companies prefer to use resin-coated proppant or sand for some or all of their fracture treatments. There are many different types of resin coat and the manufacturers are continually improving and updating their products. Therefore, the reader is advised to consult the manufacturer’s specifications for details of any specific product. However, broadly speaking, resin-coated proppant can be divided into two main categories as follows:Curable Curable resin-coated sand or proppant is coated with a resin designed to harden when exposed to temperature and/or closure stress. This allows the resin-coated grains to adhere to each other, and hence dramatically reduce the effects of proppant flowback (see Section 10.7 for more details). At low temperatures, an activator is added to the fracturing fluid in order to improve the adhesion. Tempered or Pre-Cured Tempered or Pre-cured resin coatings are harder than curable resin coats. They rely more on closure pressure than temperature in order to make the sand or proppant grains adhere to each other. These resin-coatings also have a secondary effect. Because the resin coat acts to reduce the localised contact stresses between proppant or sand grains and because any fines produced by this process are kept within the resin coat, these materials tend to have a higher closure pressure resistance than the same material without the resin coat. This means that they retain permeability under higher crush loadings and so can – for instance – extend the range over which a cheaper material, such as frac sand, can be used. Resin coated proppant or sand has a number of significant drawbacks, however:1. Cost. Coating the grains with resin can substantially increase the cost of the proppant, especially when coatings designed for high pressure and temperature are used. 2. Resin coats tend to affect the properties of the fracturing fluid. The exact variation in properties depends upon the pH of the frac fluid and the type of resin coat. However, it is common for resin-coated proppant to make frac fluids much harder to break. It is recommended that when resin-coated proppant or frac sand is being used, testing is performed on the frac fluids with the proppant in the fluid. 3. Resin coat tail-in. Many operators like to save money on a treatment by only using resin coated sand or proppant for the last 20 or 30% of the treatment. The theory being that only the part of the proppant close to the wellbore actually needs to Page 49


BJ Services’ Frac Manual 6. Proppants adhere together to prevent proppant flowback. However, due to the effects of proppant convection and settling (see Section 10.6), there is no guarantee that the proppant pumped last in the treatment will be the proppant that ends up right by the wellbore. In fact, the only way to guarantee this is to pump 100% resin coated material.

6.3

BJ Services’ FlexSand and LiteProp

BJ Services’ has two proprietary products that have significant technological advantages over conventional proppant systems

FlexSandTM FlexSand™ is designed to prevent proppant flowback by dramatically increasing the internal friction of the grains inside the proppant pack. Put simply, in order for proppant flowback to occur, individual proppant grains have to be able to move relative to each other. FlexSand acts to prevent this by introducing deformable particles into the proppant pack. The FlexSand grains are designed to be slightly deformable relative to the proppant itself. The theory, as illustrated in Figure 6.3a, is that as the formation closes on the proppant, the proppant causes the FlexSand to deform slightly, allowing the proppant grains to “key into” the FlexSand and as a result making it much harder for the grains to move relative to each other. Typically, FlexSand is mixed into the sand or proppant at between 10 and 15% by weight. The material can be either added on the fly – using a process controlled FlexSand “Bazooka” – or can be dry blended into the proppant or frac sand prior to the treatment.

Figure 6.3a – SEM micrograph of FlexSand grain clearly showing the indentations caused by the closure of the surrounding proppant grains.

Because the FlexSand has to be only slightly deformable relative to the proppant, there are three different types supplied, for different sand or proppant types; FlexSand™ LS, FlexSand™ MSE and FlexSand™ HS. These are made of different materials and – in the case of the FlexSand™ HS material – different shapes. BJ Services’ patent for FlexSand™ describes the method of preventing proppant flowback, and does not limit BJ to any specific material, nor to any grain size or shape. FlexSand also has a secondary effect. Because the FlexSand grains deform slightly, they act to “cushion” the sand or proppant grains, reducing the localised point contact stresses between grains. The reduces the quantity of fines produced, particularly by frac sand, and helps to preserve proppant or sand permeability at higher closure stresses. Thus using FlexSand can also lead to improved fracture conductivity in addition to preventing proppant flowback.

LitePropTM LiteProp™ is a proprietary low-density proppant system, designed to be neutrally buoyant in the fracturing fluid. Currently, it comes in two versions, LiteProp™ 125 and LiteProp™ 175, with SG’s of 1.25 and 1.75 respectively. Table 6.3a illustrates how this compares to other types of proppant. Because the LiteProp is designed to be neutrally buoyant, there is no need to use an expensive crosslinked fracturing fluid. Instead, any brine with the same SG as the LiteProp can be used. This in turn significantly reduces the cost and complexity of fracturing operations. However, there are a few points to be aware of when using LiteProp:Page 50



Proppant Type LiteProp 125 LiteProp 175 Frac Sand Carbolite Carbprop Carbo HSP

Specific Gravity 1.25 1.75 2.65 2.71 3.27 3.56

Table 6.3a – Specific gravity of selected proppant types

1. Although expensive fracturing fluids do not have to be used, heavy-weight brines will still have to be mixed, if neutral density is required. For instance, 1.25 SG calcium chloride brine requires 2860 lbs of CaCl2 per 1000 gals of brine – mixing this quantity of material on location could present a logistical challenge in itself. 2. Although the proppant does not require fluid viscosity in order to stay in position within the fracture, this is not the only reason for having viscosity in the fluid. If the fracture is experiencing significant tortuosity (see Section 10.1), viscosity will be required to transport the proppant through the near wellbore region. A system without viscosity may experience premature screenouts. 3. Fracturing fluids are also designed to reduce leakoff. The polymer in a typical crosslinked gel will form a filter cake against the wall of the fracture, reducing the rate at which fluids leave the fracture. Brines will not have this polymer and so will leak off into the formation much more quickly. Therefore, significantly higher fluid volumes may be required. 4. At the time of preparation of this manual, LiteProp is limited to fairly shallow formations, as the maximum closure stress that can be sustained by the material is about 5,000 psi (for LiteProp™ 125). Nevertheless, in spite of these limitations LiteProp has the potential to revolutionise the way fracturing treatments are performed.

References Technical Data Interactive CD ROM, Carbo Ceramics Inc, 2000 onwards. www.carboceramics.com, Carboceramics Inc. website, 2001 onwards PredictK software, Stim-Lab Consortium, 1999 onwards BJ Services’ Mixing Manual BJ Services’ Stimulation Engineering Support Manual Vincent, M.C., Pearson, C.M., and Kullman, J.: “Non-Darcy and Multiphase Flow in Propped Hydraulic Fractures: Case Studies Illustrate the Dramatic Effect on Well Productivity”, paper SPE 54630, presented at the SPE Annual Technical Conference and Exhibition, Houston, Oct 1999. nd

API Recommended Practice 56 Testing Sand Used in Hydraulic Fracturing Operations, 2 Edition, American Petroleum Institute, December 1995.

API Recommended Practice 60 Recommended Practices for Testing High Strength nd Proppants Used in Hydraulic Fracturing Operations, 2 Edition, American Petroleum Institute, December 1995.

Page 51


BJ Services’ Frac Manual 6. Proppants Rickards, A., Lacy, L., Brannon, H., Stephenson, C. and Bilden, D.: “Need Stress Relief? A New Approach to Reducing Stress Cycling Induced Proppant Pack Failure”, paper SPE 49247 presented at the 1998 SPE Annual Technical Conference and Exhibition, New Orleans, Louisiana, Oct 1998. Wood, W.D., Brannon, H.D., Rickards, A.R. and Stephenson, C.: “Ultra-Lightweight Proppant Development Yields Exciting New Opportunities in Hydraulic Fracturing Design”, paper SPE 84309, presented at the 2003 SPE Annual Technical Conference and Exhibition, Denver, Colorado, Oct 2003.

Page 52


BJ Services’ Frac Manual 7. Rock Mechanics

7.

Rock Mechanics

Rock mechanics is the study of the mechanical properties of a rock, especially those properties which are of significance to Engineers. It includes the determination and effects of physical properties such as bending strength, crushing strength, shear strength, moduli of elasticity, porosity and density, and their interrelationships.

7.1

Stress

Consider the situation illustrated in Figure 7.1a, in which a block of material is subjected to a force F:-

F Area = A

Figure 7.1a – A block of material subjected to a force F.

The block of material has an area A, on the plane at right angles to the line of action of the force. Therefore the stress, σ, is given by:-

σ

F = A .................................................................................. (7.1)

Note that this is very similar to the formula for calculating pressure. Stress and pressure have the same units and are essentially the same thing – stored energy. The main difference between the two is that in liquids and gases, the material will flow away from an applied force, until the force and stress (or pressure) is the same in all directions (i.e. an equilibrium has been reached). However, solids cannot deform in such a manner, so these materials will always have a plane across which the stresses are at a maximum. They will also have a plane perpendicular to this, across which the stresses are at a minimum. Properties such as mass and volume are said to be scalars – they require only a magnitude to define them. Quantities such as force and velocity are vectors – they require not only a magnitude, but also a direction in which they are acting in order to be fully defined. Stress takes this one step further, and is a tensor property – it can only be fully defined by a magnitude and an area across which it is acting.

7.2

Strain

Strain is measure of how much the material has been deformed when a stress is applied to it. Figure 7.2a illustrates how the block of material is compressed by the force F:-

Page 53



F

x1 x2

Figure 7.2a – Strain produced by the application of force F

As the force is applied, the height of the block of material changes from x1 to x2. The strain, ε, is given by:-

ε

=

x1 - x2 x1 ........................................................................... (7.2)

Note that the strain is defined in the same direction as the applied force F and perpendicular to the plane across which the stress acts. Strain is important as this is the way we measure stress – by observing the deformation of a known piece of material. Strain is dimensionless.

7.3

Young’s Modulus

Young’s modulus, E, (also known as modulus of elasticity or elastic modulus) is defined as follows:E

=

σ ε ................................................................................... (7.3)

E is the ratio of stress over strain. As strain is dimensionless, E has the same units as stress. Young’s modulus is a measure of how much a material will elastically deform when a load is applied to it. This is another term for hardness. On a more fundamental level, if stress and pressure are closely related (apply a pressure to a surface and it will induce a stress), then in fracturing, we can think of Young’s modulus as a measure of how much a material (i.e. rock) will elastically deform when a pressure is applied to it. As pressure is stored energy, E is also a measure of how much energy it takes to make the rock deform. Materials with a high Young’s modulus, such as glass, tungsten carbide, diamond and granite, tend to be very hard and brittle (susceptible to brittle fracture). Conversely, materials with a low E, such as rubber, Styrofoam and wax, tend to be soft and ductile (resistant to brittle fracture). Caution – Elastic vs Plastic. Elastic deformation is reversible – if the force (or pressure, or stress) is removed, the material returns back to its original size and shape. If so much force is applied to a material that it passes beyond its elastic limit then the material will start to plastically deform. This is permanent. A good illustration of this is the small spring from a ball Page 54


BJ Services’ Frac Manual 7. Rock Mechanics point pen. When the spring is lightly stretched, it will return to its original shape. However, if the spring is stretched too far, it will be permanently, or plastically, deformed. Young’s Modulus only applies to elastic deformation. As a group of materials, rocks tend not to plastically deform very much. Instead they will elastically deform and then fracture if the stress gets too high. Notable exceptions to this are salt beds, soft carbonates (e.g. chalk) and young coals. Static Young’s Modulus is the standard measure of E and is applicable to hydraulic fracturing. The material is being deformed slowly and in only one direction. Dynamic Young’s Modulus is the rock property measured by special sonic logging tools. The material is no longer static – it is being continually stretched and then compressed rapidly. There is often a significant variation between static and dynamic values for E due to a process known as hysteresis. Hysteresis is a retardation of the effects of forces, when the forces acting upon a body are changed (as if from viscosity or internal friction). In this situation, it represents the history dependence of the physical systems. In a perfectly elastic material, elastic stress and strain is infinitely repeatable. In a system exhibiting hysteresis, the strain produced by a force is dependent upon not only the magnitude of that force, but also the previous strain history (see Section 7.10) Plane Strain Young’s Modulus. In hydraulic fracturing, the strain in the direction perpendicular to the fracture plane (i.e. the direction in which fracture width is produced) is effectively zero. This is because in this situation the denominator in Equation 7.2 (the “x1”) is so large that the strain is effectively zero, even though there has been measurable material deformation. This is known as “plane strain”, which implies that strain only exists in a directions perpendicular to the direction in which strain is zero. To account for this anomaly, fracture simulators use the plane strain Young’s modulus, E’, to calculate the fracture width:E’

=

E 2 ............................................................................. (7.4) (1 - ν )

In fracturing, Young’s modulus will typically have values ranging from as low as 50,000 psi (for a shallow, very soft chalk or weak sandstone) to as high as 6,000,000 psi for deep, tight, shaley sandstone. It should be noted that Young’s modulus may not be constant in weak or unconsolidated formations.

7.4

Poisson’s Ratio

Poisson’s ratio, ν, is a measure of how much a material will deform in a direction perpendicular to the direction of the applied force, parallel to the plane on which the stress induced by the strain is acting. This is illustrated by Figure 7.4a:F

x1 x2

y1 y2 Figure 7.4a – Application of force F also produces a deformation in the y direction

Page 55


BJ Services’ Frac Manual 7. Rock Mechanics The strain in the x direction, εx, is given by Equation 7.2 (see Section 7.2). The strain in the y direction is given by the following:-

εy

=

y1 - y2 y1 ........................................................................... (7.5)

Note that this value is negative – this is a result of the way the forces and the direction the forces act in are defined. Compressive strain is positive and tensile strain is negative. Poisson’s ratio is defined by Equation 7.6:-

ν

=-

εy ................................................................................. (7.6) εx

Poisson’s ratio is an important factor in determining the stress gradient of the formation, but is less important in defining fracture dimensions, although it does have some effect. Typical values for ν for rocks are between 0.2 and 0.35 (ν is dimensionless).

7.5

Other Rock Mechanical Properties

Tensile Strength. The tensile strength of a material is the level of tensile stress that is required in order to make the material fail. Usually, as stress is applied the material will elastically deform (reversible), plastically deform and then fail. In most rocks this amount of plastic deformation is negligible and the material will, for all practical purposes, elastically deform and then fail. This property is important in hydraulic fracturing, as this stress level has to be overcome in order to split the rock. Usually, the frac gradient (which is the pressure – a.k.a. the stress – needed to make the rock fracture) has two components – the stresses induced by the overburden, and the tensile strength of the rock. See Section 7.6 below for a more detailed explanation of in-situ stresses. It should be noted that materials also have a Compressive Strength, which is the compression load, beyond which a material will fail. Failure mechanisms are more complex, as the material is often compressed in several directions at once. Generally, rocks are much stronger in compression than in tension, a fact which we take advantage of during fracturing. Shear Modulus. The shear modulus is similar to the Young’s modulus, except that it refers to the material being in shear, rather than in compression or tension. It defines how much energy is required to elastically deform a material in shear:-

x F

h

a

b

Figure 7.5a – Force F applied to produce a shear stress

Page 56


BJ Services’ Frac Manual 7. Rock Mechanics With reference to Figure 7.5a, the shear stress, τ, is given by:-

τ

F = A .................................................................................. (7.7)

where A is the area of the block of material parallel to the line of action of the force F, (this is the plane along which the shear stress acts) and is equal to a × b. The shear strain, γ, is defined as follows:-

γ

x = h ................................................................................... (7.8)

Therefore, the shear modulus, G, is equal to the shear stress divided by the shear strain:G

Fh t = g = x A ....................................................................... (7.9)

Bulk Modulus. This is another elastic constant, which defines how much energy is required to deform a material by the application of external pressure. This is a special form of compressive stress, in which the applied compressive stress is equal in all directions. Suppose we have a block of material, which originally has a pressure P1, applied to it, and has a volume V1. This pressure is increased to P2, which causes the volume to decrease to V2, as illustrated below in Figure 7.5b. The increase in bulk stress is the same as the increase in pressure, P2 – P1. The bulk strain is equal to the change in volume, V2 – V1 divided by the original volume, V1. Thus, the bulk modulus, K, is given by:K

V1(P2 - P1) P2 - P1 = - (V - V )/V = - V - V ....................................... (7.10) 2 1 1 2 1

Figure 7.5b – Volume changes from V1 to V2 as pressure increases from P1 to P2.

K

dP = - V dV ......................................................................... (7.11)

The minus sign is introduced into the equation due to the fact that the term V2 – V1 will always be of the opposite sign to the term P2 – P1. The bulk modulus is therefore a measure of how much energy it takes to compress a material using externally applied pressure. Relationships Between the Four Elastic Constants. The four main elastic constants – Young’s modulus, shear modulus, bulk modulus and Poisson’s ratio - are all related to each other. If two of these material properties are known, the other two can be deduced:-

Page 57


BJ Services’ Frac Manual 7. Rock Mechanics E

= 3K (1 – 2ν ) ................................................................. (7.12)

K

=

E ......................................................................... (7.13) 3 - 6ν

G

=

E ........................................................................ (7.14) 2 + 2ν

ν

=

3K - E 6K ......................................................................... (7.15)

Therefore, if the Young’s modulus and the Poisson’s ratio are known, the shear modulus and the bulk modulus can be deduced. Thus, fracture simulators only require the input of E and ν.

7.6

In-Situ Stresses

In situ stresses are the stresses within the formation which act as a load (usually compressive) on the formation. They come mainly from the overburden, and these stresses are relatively easy to predict. However, factors such as tectonics, volcanism and plastic flow in underlying formations can significantly affect the in-situ stresses – these factors are much harder to predict. In addition, the act of producing a localised anomaly – such as an oil well – can also significantly affect the stresses in a specific area. The stresses due to the overburden are simply the sum of all the pressures induced by all the different rock layers. Therefore, if there has been no external influences – such as tectonics – and the rocks are behaving elastically, the vertical stress, σv, at any given depth, H is given by:-

σv

=

H 0

ρnghn ....................................................................... (7.16)

where ρn is the density of rock layer n, g is the acceleration due to gravity and hn is the vertical height of zone n, such that h1 + h2 + ..... + hn = H. This is usually modified (after Biot et al) to allow for the effects of pore (or reservoir) pressure, such that:-

σv

= γob H - αPres ............................................................... (7.17)

where γob is the overburden pressure gradient (usually between 1.0 and 1.1 psi/ft) and α is Biot’s poroelastic constant, and is a measure of how effectively the fluid transmits the pore pressure to the rock grains. α depends upon variables such as the uniformity and sphericity of the rock grains. By definition α is always between 0 and 1, usually it is taken to be between 0.7 and 1.0 for petroleum reservoirs. Stresses under the ground do not just act on a single plane. There is a complex three dimensional stress regime. To simplify things, stresses are usually resolved into three mutually perpendicular stress components; the vertical stress, σV, and two horizontal stresses, σH, min and σH, max. Additionally, as the stresses are three dimensional, so are the strains. The elastic relationship between these stresses and strains in three mutually perpendicular directions, x, y and z, is governed by Hooke’s law:-

εx

Page 58

1 = E [σx - ν(σy + σz)] ........................................................ (7.18)


BJ Services’ Frac Manual 7. Rock Mechanics Now, for the case of elastic deformation with no outside influences (such as tectonics) in subterranean rock strata, there are two important things to note. First, σH,min = σH, max, as the stresses will be symmetrical on the horizontal plane. Secondly, as each individual unit of rock is pushing against another identical unit of rock with the same force, εH, min = εH, max = 0 (i.e. no deformation on the horizontal plane). Therefore:-

σH

=

σv ν

1-ν

............................................................................ (7.19)

As a result of the work of Terzaghi, Biot and Handin et al., this Equation is generally modified to allow for the effects of the pore pressure:-

σH

=

ν(σv - 2αP) + αP ........................................................ (7.20) 1-ν

From Equation 7.20 we can see that the Poisson’s ratio can have a considerable influence on the horizontal in-situ stresses.

7.7

Stresses Around a Wellbore

A wellbore is essentially a pressure vessel with a very thick wall. Consequently, the same theories that are applied to thick walled pressure vessels can also be applied to wellbores, providing that the in-situ stresses and reservoir pressure are accounted for. Figure 7.7a illustrates how the stresses at any given point near the wellbore can be resolved into three principle stresses. Once again, these are perpendicular to each other.

σv

σt σr

σr

σt σv Figure 7.7a – Three dimensional stresses around a wellbore

From Deily and Owens (1961) we can get expressions for the radial and tangential stresses induced by a pressure in the wellbore, Pwb, at a radius r, from the centre of the well. The vertical stress is as given in Equation 7.17;

σt Page 59

= -[Pwb - α(Pres + Pwb - Pr)]

2

2

rw rw 2 2 r + 1 + r σv .............. (7.21)



σr

2

rw = (Pwb – Pres) r2

+

ν

1-ν

2

rw 1 - r2 (Pob – Pres) ......... (7.22)

where Pob is the pressure due to the overburden (see reference for more details). At the wellbore face, the stresses due to wellbore pressure will be at a maximum. Also, this is by definition the point at which the fracture initiates. Therefore, these are the stresses which interest us most. At the wellbore r → rw and Pr → Pwb so that:2ν (γ H - αPres) – (Pwb – αPres) .......................... (7.23) 1 - ν ob

σt

=

σr

= Pwb - Pres ..................................................................... (7.24)

Furthermore, Barree et al (1996) went on to show that provided the rock does not have any significant tensile strength and no significant plastic deformation, failure of the rock (i.e. breakdown) occurred when the tangential stresses were reduced to zero; Pb

7.8

=

2ν (γ H - αPres) + αPres ...................................... (7.25) 1 - ν ob

Fracture Orientation

Fractures will always propagate along the line of least resistance. In a three dimensional stress regime, a fracture will propagate so as to avoid the greatest stress. This means that a fracture will propagate parallel to the greatest principal stress, and perpendicular to the plane of the greatest principle stress. This is a fundamental principle – therefore the key to understanding fracture orientation is to understand the stress regime itself. Propagation parallel to the greatest principle stress usually means that the fracture will propagate on a vertical plane. We can see from Equations 7.16 to 7.20 that the horizontal stresses in an undisturbed elastic formation will always be less than the vertical stress. However, there are some exceptions to this. Magnitude of In-Situ Stress

Magnitude of In-Situ Stress

σV

Depth

Depth

Formation lost due to erosion

σV

σH

Original Stress Regime

σH

Stress Regime After Loss of Height by Erosion

Figure 7.8a – Changes in stress regime due to erosion

Page 60


BJ Services’ Frac Manual 7. Rock Mechanics Equations 7.17 and 7.20 define the magnitude of horizontal and vertical stresses in undisturbed formations. The horizontal stresses are induced by the vertical stresses. There is evidence to suggest that these horizontal stresses somehow get “locked” into place (Economides, et al), and remain relatively constant, regardless of what later happens to the vertical stress. Figure 7.8a illustrates what happens when the vertical stress is reduced. If formation is lost due to erosion, then the overburden stresses are reduced. However, because the horizontal stresses are “locked-in”, they have not been reduced. Therefore, there is a region, close to the new surface, where the horizontal stresses are greater than the vertical stresses. This means that the fracture will propagate horizontally – a “pancake frac”. Thus, in shallow formations in areas with a history of surface erosion, horizontal fracs are not only possible, they are in fact likely. This does not apply to formations which are very weak or unconsolidated, as stresses cannot be “locked in” if the rock strata have no strength. Another consequence of this phenomenon is that in formations where the σV and the σH are approximately equal, it can be very hard to predict fracture orientation. The action of outside forces, such as tectonics and volcanism, can also significantly affect fracture orientation. The extra stresses imposed by the movement of the Earth’s crust, which does not usually alter the overburden stress, but can significantly alter the horizontal stresses. In addition, formations can sometimes be bent and buckled. In Barbados, there is a formation that has experienced so much tectonic stress that it now runs vertically. Its stresses have been locked into place, so now the original vertical stress is horizontal, and vice versa. So the fractures propagate horizontally. Influence of Wellbore Orientation. Drilling a well can significantly alter the stress regime in an area around the well. The distance away from the wellbore that is affected by this change is dependent upon the Young’s modulus of the formation. Hard formations (high E) tend to transmit stress more easily than soft formations (which will deform to reduce the stress). Therefore hard formations are affected more than soft formations. In the area around the wellbore – the area affected by the new stress regime – fractures may propagate parallel to the wellbore, even if the wellbore is highly deviated or even horizontal. As the fracture propagates away from the wellbore, it will eventually reach a point at which the normal stress regime of the formation becomes more significant than the near wellbore stress regime. At this point, the fracture will change orientation. Sometimes this re-orientation can be quite sudden, resulting in sharp corners in the fracture, which can cause premature screen outs.

7.9

Breakdown Pressure and Frac Gradient

The breakdown pressure is the pressure it takes to initiate a fracture from the wellbore. Due to the effects of the stresses induced by the presence of the wellbore, the breakdown pressure is usually significantly greater than the fracture - or frac – gradient, which is a measure of how much pressure it takes propagate the fracture through the formation, away from the influence of wellbore effects. Both are usually expressed as pressure gradients (i.e. in psi/ft or kPa/m) so that similar formations in different wells at different depths can be more easily compared. The frac gradient is a very important quantity in fracturing, as it is the most significant contributor to the bottom hole treating pressure, which in turn helps to define the surface treating pressure, the loading on the completion and the proppant selection. In order to produce a fracture in the formation, two forces have to be overcome. The first force is the in-situ stress, which is defined in Equations 7.19 and 7.20 when there are no external influences such as tectonics etc. The second force that has to be overcome is the tensile strength of the rock, which is usually in the region of 100 to 500 psi. Roegiers, in his chapter on Rock Mechanics in Economides and Nolte’s excellent Reservoir Stimulation, defined the breakdown pressure in the following Equations:Pb, upper Page 61

= 3 σH,min - σH,max – P + T ............................................... (7.26) Version 1.0 June 2005 Uncontrolled Copy – DO NOT DISTRIBUTE


Pb, lower

=

3σHmin - σHmax - 2ηP + T ............................................. (7.27) 2(1 - η)

where η is a parameter defined by the Poisson’s ratio and Biot’s constant, as follows:-

η

=

α(1 - 2ν) ..................................................................... (7.28) 2(1 - ν)

Pb, upper is the breakdown pressure assuming no fluid invasion into the formation (and hence the maximum possible theoretical breakdown pressure), Pb, lower is the lower boundary for breakdown pressure, assuming significant alteration of the near wellbore pore pressure due to fluid invasion, σH,min is the minimum horizontal stress, σH,max is the maximum horizontal stress, P is the reservoir pressure and T is the tensile strength of the rock. From this we can see that the higher the reservoir pressure, the easier it is to fracture the rock, so that depleted reservoirs tend to have higher breakdown pressures than undepleted reservoirs. In addition, we can see that when we have fluid invasion, the breakdown pressure can be significantly reduced, which implies that lower viscosity fluids provide lower breakdown pressures. In the case where there are no significant external influences on the stress regime, the two horizontal stresses are equal and the Equations can be simplified to:and

Pb, upper

= 2 σH – P + T ................................................................ (7.29)

Pb, lower

=

2σH - 2ηP + T ............................................................ (7.30) 2(1 - η)

The breakdown gradient is simply the breakdown pressure, Pb, divided by the TVD. The frac gradient is the pressure required to make the fracture propagate, outside of the influences of the wellbore (the region referred to as “far-field”). As stated above, this is often significantly lower than the breakdown pressure, depending upon the viscosity of the frac fluid, the reservoir pressure and the contrast between maximum and minimum horizontal stresses. In general, the far field fracturing pressure is equal to the minimum horizontal stress, modified to allow for the effects of pore pressure. In general, any external effects such as tectonics or faulting, will only act to increase the stresses. Therefore Equation 7.18 defines the frac gradient, gf, as follows:gf

=

1 TVD

ν(σv - 2αP) + αP ....................................... (7.31) 1-ν

Important Note. The best way to get the frac gradient for a formation is to pump some fluids into it and measure the response. There are many influences on the formation that Equations 7.26 and 7.27 do not account for, such as tectonics (there are very few areas of the world that are completely free of tectonics), and the only way to account for these is to actually measure them. The second best way to get the frac gradient is to look at data from offset wells. Make sure that you are looking at data from the same formation. Compare values for Poisson’s ratio and reservoir pressure. If these values are similar (provided they come from the same formation), then the frac gradient will probably be similar as well. Once these two methods have been rejected, the remaining way to get the frac gradient is to use the Equations above. This method should only be used if attempts at carrying out the other two methods have failed.

Page 62



7.10

Rock Mechanical Properties from Wireline Logs

Certain types of open hole wireline logs can be used to provide useful information about the mechanic properties of the formations involved in the fracturing process. The dipole sonic or sonic array is the main tool used to do this. This is a special tool, and is different from the sonic logging tool used to generate the sonic transit time seen on most logs. In order to be able to quantify the rock mechanical properties, the logging tool must be able to generate and measure two completely different types of sonic waveform, the shear or s-wave and the compression or p-wave, as illustrated in Figure 7.10a, below:-

Figure 7.10a – The left hand side shows the shear or s-wave, whilst the right hand side shows the compression or p-wave. In both diagrams, the blue arrows illustrate the overall movement of the sonic waveform, whilst the red arrows indicate the movement of individual particles.

For the shear wave, the material is continually sheared in one direction and then the opposite direction, back and forth. The plane across which the material is being sheared is perpendicular to the direction the shear wave is travelling. For the compression wave, the material is subject to alternating compression and tension, on a plane that is again perpendicular to the direction of wave travel. The dipole sonic tools measures the transit time of both the shear wave, ts and compression wave, tp. These values are usually expressed in units of µsec/ft, so that the transit time is the reciprocal of the wave velocity. The transit time of the sonic waves through the formation can be used to derive dynamic rock mechanical properties as follows:2

νd

0.5(ts/tp) - 1 = (t /t )2 - 1 ............................................................... (7.32) s p

Ed

= 2 t 2 (1 + νd) ............................................................ (7.33) s

ρb

ρb

= 26,950 t 2 (1 + νd) (in field units)............................. (7.34) s Where νd is the dynamic Poisson’s ratio, Ed is the dynamic Young’s modulus (see below for an explanation of dynamic and static properties) and ρb is the bulk density, usually taken from the corrected bulk density log. For Equation 7.34 in field units, ρb is in g/cc and ts is in µsec/ft 6 – the units most commonly used on logs - whilst Ed is in psi x 10 . Other rock mechanical properties can also be found (in “log” field units):-

Page 63


BJ Services’ Frac Manual 7. Rock Mechanics Ed ..................................................................... (7.35) 2(1 + νd)

Gd

=

Kd

1 1 = 26,950 ρb t 2 - 3t 2 ..................................................... (7.36) p s

cb

1 = K ............................................................................... (7.37)

cr

=

α

cr = 1 - c .......................................................................... (7.39)

d

1 ............................................... (7.38) 1 4 26950 ρb t 2 - 3t 2 ma sma

b

Where Gd is the dynamic shear modulus, Kd is the dynamic bulk modulus, cb is the bulk compressibility of the formation, cr is the rock, or zero porosity, compressibility tma is the rock matrix compression wave transit time (see below), tsma is the rock matrix shear wave transit time (see below) and α is Biot’s poroelastic constant. Table 7.10a lists commonly used values for tma and tsma:-

Rock Matrix

Commonly used values for sonic wave rock matrix transit times, µsec/ft Compression Wave tma

Shear Wave tsma

Quartz

55.5 or 51.0

83.3

Calcite

49.7

90.0

Dolomite

43.5

78.7

Anhydrite

50.0

87.7

Granite

50.8

89.3

Salt

66.7

125.0

Table 7.10a – Commonly used values for compression and shear wave rock matrix sonic transit times (after Schlumberger, 1989)

Figure 7.10b shows an example dipole sonic log with interpreted values for Poisson’s ratio, Young’s modulus and horizontal stress. Stress can be derived by using the dynamic Poisson’s ratio and the density log (to give the vertical stress) using Equation 7.19. However, it should be remembered that these “stress logs” are based on dynamic properties (see below) and do not take poroelastic effects into account. Nevertheless, whilst the absolute values for these logs cannot be trusted, they can be useful for determining stress contrasts.

Page 64



Dynamic and Static Rock Mechanical Properties In most applications, including hydraulic fracturing, we are using values for rock mechanical properties that are based on static material properties. However, because of the conditions that sonic-based rock mechanical properties are measured under, they are said to be dynamic, and there is often a significant difference between the static properties that the Frac Engineer needs and the dynamic properties that are measured by open hole logs. 0 0

GR

200

4

HCA L

14

PR

1

50

DTCO

250 1

YO UNGS

3.5

50

DTSM

250 500 0

HSTRES S

750 0

9, 050

9, 100

9, 150

9, 200

Figure 7.10b – Example interpreted dipole sonic log. The left track shows gamma ray (GR) and caliper (HCAL) logs. The center track shows compression (DTCO) and shear (DTSM) wave transit times. The right track shows interpreted values for Poisson’s ratio (PR), Young’s modulus (YOUNGS) and horizontal stress (HSTRESS).

To put things simply, when a stress related event happens to a material, the changes that occur to the stresses in the material to not occur instantly. Instead, any change to the stress will spread through the material at the speed of sound in that material. Usually, the time taken for this to happen is so small compared for the time taken for the applied stresses to change (as in fracturing) that it does not affect the process. However, when the stresses applied to a material alter at a speed that is a significant fraction of the speed of sound of that material, then the time taken for the change in stress to propagate can significantly affect the stresses themselves. For instance, when a compression wave is passing through a material, any given portion of that material is constantly being subjected to alternating tensile and compression loads. The speed at which the load changes is directly proportional to the frequency of the sound wave, whilst the speed that the compression wave moves through the material is the speed of sound for that material. At low frequencies, the length of time taken for a piece of the material to undergo one full stress cycle is much less than the length of time it takes one sound wave to travel past that piece of material. However, as the frequency increases, the length of time between stress cycles decreases, whilst the wave transit time stays constant, and it becomes increasingly difficult for the material to return to it’s original state before the next wave passes through. This causes a deviation away from linear elastic behaviour, as illustrated by Figure 7.10c. Page 65


BJ Services’ Frac Manual 7. Rock Mechanics Stress

Stress

TENSION

TENSION

Strain COMPRESSION

Strain COMPRESSION

Figure 7.10c – Static (left) and dynamic (right) cyclic stress loading.

The left side of Figure 7.10c shows cyclic loading under static conditions. As the load is alternated between tension and compression, the relationship between stress and strain is linear (proportional to the Young’s modulus) and follows the same path on the stress strain plot every time, provided the elastic limit is not exceeded and the material is not plastically deformed. This relationship between stress and strain is referred to as linear elastic. The right side of Figure 7.10c shows the dynamic case. The behaviour of the material under loading is now dependent upon the stress history of the material. The relationship between stress and strain is different depending upon whether the loading is being applied or removed and whether it is tensile or compression. This deviation away from linear behaviour becomes more pronounced as the frequency of the sound waves (i.e. the frequency of the stress cycling) increases. When an alternating stress is applied to a material the induced alternating strain moves through this material at the speed of sound, for that material. However, as the frequency of the changes gets closer to the speed of sound in that material, the material has insufficient time to return to its original state, before the next deformation occurs. Thus the subsequent deformation is influenced by the previous deformation. This deviation from linear elastic behaviour under high frequency stress cycling is often referred to as hysteresis. Hysteresis is a general term used throughout science and engineering to denote when the behaviour of a material under certain conditions is dependent upon the historical application of these conditions. The behaviour of a material that does not exhibit hysteresis (such as that shown on the left hand side of Figure 7.10c), is the same every time, regardless of what has happened previously. In order to convert from dynamic to static properties, several correlations are available. Usually these are based on empirical data derived from tests on core samples and then extrapolated back to BH conditions. As such, there is a degree of inaccuracy associated with them. Lacy’s method (1997) is recommended for Young’s modulus:E

2

= 0.018 Ed + 0.422 Ed ................................................... (7.40)

However, there is no such correlation for Poisson’s ratio.

Page 66



Frac Gradient The horizontal stresses (assumed to be equal) can be calculated using the log data and Equation 7.20. To use this Equation, three important pieces of information must be acquired:• • •

Vertical stress. Usually found by taking the bulk density back to the surface and using Equation 7.16. Pore pressure. Biot’s poroelastic constant, usually found using Equation 7.37 – 7.39 and table 7.10a. Otherwise use 0.8 for a poorly consolidated formation and 1.0 for a consolidated formation

References Economides, M.J., and Nolte, K.G.: Reservoir Stimulation, Schlumberger Educational Services, 1987. Economides, M.J.: A Practical Companion to Reservoir Stimulation, Elsevier, 1992 Biot, M.A.: “General Theory of Three-Dimensional Consolidation,” Journal of Applied Physics , 1941, 12, p155-164. Biot, M.A.: “General Solutions of the Equations of Elasticity and Consolidation for a Porous Material,” Journal of Applied Mechanics, 1956, 23, p91-96. Deily, F.H., and Owens, T.C.: “Stress Around a Wellbore”, paper SPE 2557, presented at the Annual Fall Meeting of the SPE, October 1969. Barree, R.D., Rogers, B.A., and Chu, W.C.: “Use of Frac-Pack Pressure Data to Determine Breakdown Conditions and Reservoir Properties”, paper SPE 36423, presented at the SPE Annual Technical Conference and Exhibition, Denver, October 1996. Handin J., Hager, R. V. Jr, Friedman, M., and Feather, J. N.: “Experimental Deformation of Sedimentary Rocks Under Confining Pressure: Pore Pressure Tests,” Bulletin AAPG, 1963, 47, p717-755. Terzaghi, K. van: “Die Berechnung der Durchlassigkeitsziffer des Tones aus dem Verlauf der Hydrodynamischen Spannungserscheinungrn,” Sber. Akad. Wiss, Vienna, 1923, 123, p105 (in German) Bradley, H.B. (Editor), Petroleum Engineering Handbook, Society of Petroleum Engineers, Richardson, Texas, 1987, 51. Log Interpretation Principles/Practices, Schlumberger Educational Services, Houston, Texas, 1989, 5. Lacy, L.L.: “Dynamic Rock Mechanics Testing for Optimized Fracture Design”, paper SPE 38716, presented at the 1997 SPE Annual Technical Conference and Exhibition, San Antonio, Texas, Oct 1997

Page 67


BJ Services’ Frac Manual 8. 2-D Fracture Models

8.

2-D Fracture Models

2-D fracture models were the industry’s first attempt at mathematically modelling the process of fracture propagation. By today’s standards, they are crude approximations. However, there are two important points to note. First, in order to understand how the modern 3-D models work, it is first necessary to understand the 2-D models. Second, there are some circumstances in which certain 2-D models can be valid. These include coal bed methane fracturing (KZD) and fracturing in massive, uniform formations (radial).

8.1

Radial or Penny-Shaped

R

H

Wmax

Figure 8.1a – Propagation of a radial or penny-shaped fracture

Figure 8.1a shows the propagation of a radial or penny-shaped fracture. In this model, the height, H, is a function of the radius or half-length of the fracture, R, such that H = 2R. This produces a fracture, which is circular in shape. The width of the fracture is given by:2

Wmax

=

8 ( 1 - ν ) ∆P R ............................................................ (8.1) πE

Where ∆P is the net pressure, ν is the Poisson’s ratio and E is the Young’s modulus. In this model, the width at any part of the fracture is a function of the distance between the center and the edge of the frac such that:w(r) w ¯

= Wmax

1-

r R .......................................................... (8.2)

8 = 15 Wmax ........................................................................ (8.3)

Note the following points, which are applicable to all the 2-D fracture models:i)

Wmax is inversely proportional to the Young’s modulus. This means that as the formation gets harder (i.e. the Young’s modulus increases), the net pressure required to produce a given width increases. So it takes more energy to produce width in a hard formation than it does in a soft formation.

Page 68


BJ Services’ Frac Manual 8. 2-D Fracture Models ii)

Wmax is directly proportional to the half-length of the fracture – if the half-length is doubled, the width is doubled. Note that this is the created width, not the final propped width, which is what the post treatment production increase will be partially dependent upon. The propped width will always be equal to or less than the created width, and is a function of the volume of proppant placed per unit area of the fracture.

iii)

Wmax is relatively insensitive to changes in Poisson’s ratio. An increase in ν from 0.2 2 to 0.25 (an increase of 25%) will change the term (1 - ν ) from 0.96 to 0.9375, a decrease of only 2.34%. Therefore, it is pointless to spend too much time trying to get accurate values for ν. However, as seen in Chapter 6, ν can have a significant effect on the magnitude of the horizontal stresses – if the frac gradient is unknown, then finding accurate values for ν can be important.

The radial model has no limits to height growth. As long as the fracture is growing outwards (i.e. R is increasing), then it will also be growing up and down the wellbore (i.e. an increase in H). This type of propagation can be found in a massive uniform formation with no vertical variations in rock properties and hence no “barriers” to height growth. It can also be found for small fractures that have not contacted any “barriers”, such as in skin bypass fracturing. The volume of the fracture is obtained from the volume of fluid pumped into the fracture, less the volume of fluid leaked off. The volume of fluid leaked off is a function of the leakoff area of 2 the fracture (which is equal to 2πR ), so that if the fluid efficiency (η), injected volume of fluid, E, ν and ∆P are known, R can be easily obtained:R

=

3

3ηQtE ....................................................... (8.4) 2 16 ( 1 - ν ) ∆P

where Q is the average pump rate and t is the pump time.

8.2

Kristianovich and Zheltov - Daneshy (KZD)

This model was originally developed by two Russians, Kristianovich and Zheltov, and was later modified by Daneshy, Geertsma and de Klerk , and also by Le Tirant and Dupuy. Often, this model is referred to as GDK, after Geertsma and de Klerk. In this model, the height is fixed, and remains constant throughout. It is usually set as the gross height of the formation:Wmax

L

H

Figure 8.2a – Schematic showing the general shape of the KZD fracture

Page 69


BJ Services’ Frac Manual 8. 2-D Fracture Models As we can see from Figure 8.2a, the KZD model produces a fracture with a constant height. This means that there must be slippage between the formation being fractured and the formations above and below. This is unlikely (but not unknown) in most situations, but can happen when fracturing coal beds. The maximum width is related to the half length L by the following Equation:2

Wmax

=

4 ( 1 - ν ) ∆P L ............................................................ (8.5) E

Note that for a given net pressure and half length, the maximum width of a KZD fracture is greater than the maximum width of a radial fracture by a factor of π/ 2. The average width is given by:w ¯

π

= 4 Wmax .......................................................................... (8.6)

Therefore, for two “wings”, the length of the fracture is given by L

=

2

ηQtE .................................................. (8.7) 2 2 π ( 1 - ν ) ∆P H

where η is the fluid efficiency, Q is the average pump rate and t is the pump time.

8.3

Perkins and Kern - Nordgren (PKN)

This fracture model was originally conceived by Sneddon and later developed by Perkins and Kern, with further work by Nordgren, Nolte and Advanti et al. In this model, the maximum width is related to the height of the fracture, such that:2

Wmax =

2 ( 1 - ν ) ∆P H ............................................................... (8.8) E

whilst the average width, w ¯ , is given by:w ¯

=

π 5 Wmax ............................................................................. (8.9)

Wmax

H

L

Figure 8.3a – The Perkins and Kern - Nordgren fracture

Page 70


BJ Services’ Frac Manual 8. 2-D Fracture Models Thus, both fracture height and width are constant down the length of the fracture. Figure 8.3a illustrates the shape of this fracture. The length of the fracture can be determined by a method similar to those used for the radial and KZD fractures:L

=

5ηQtE 2 2 ........................................................ (8.10) 4 π ( 1 - ν ) ∆P H

The PKN fracture geometry was used for many years by the industry as the standard, until the advent of pseudo-3D fracture simulators and an improved understanding of fracture propagation and fracture mechanics (see Sections 9 and 11).

References Abé, H., Mura, T., and Keer, L.M.: “Growth Rate of a Penny-Shaped Crack in Hydraulic Fracturing of Rocks”, J. Geophys. Res. (1976) 81, 5335. Zheltov, Y.P., and Kristianovitch, S.A.: “On the Mechanism of Hydraulic Fracturing of an OilBearing Stratum”, Izvest. Akad. Nauk SSR, OTN (1955) 5, 3-41 (in Russian) Daneshy, A.A.: ”On the Design of Vertical Hydraulic Fractures”, JPT, Jan 1973, 83-93. Geerstma, J., and de Klerk, F.A.: “Rapid Method of Predicting Width and Extent of Hydaulically Induced Fractures”, JPT, Dec 1969, 1571-81 Le Tirant, P., and Dupuy, M.: “Fracture Dimensions Obtained During Hydraulic Fracturing Treatments of Oil Reservoirs”, Rev. Inst. Français du Pétrole (1967) 44-98 (in French). Sneddon, I.N.: “The Distribution of Stress in the Neighbourhood of a Crack in an Elastic Solid”, Proc. Royal Society of London, (1946) 187, 229. Perkins, T.K., and Kern, L.R.: “Widths of Hydraulic Fractures”, JPT, Sept 1961, 937-949. Nordgren, R.P.: “Propagation of a Vertical Hydraulic Fracture”, SPEJ, Aug 1972, 306-314. Nolte, K.G.: “Determination of Proppant and Fluid Schedules From Fracturing Pressure Decline”, SPEPE, July 1986, 255-265. Advanti, S.H., Khattib, H., and Lee, J.K.: “Hydraulic Fracture Geometry Modeling, Prediction and Comparisons”, paper SPE 13863, presented at the SPE/DOE Low-Permeability Gas Reservoirs Symposium, Denver, May 1985. Howard, G.C., and Fast, C.R.: Hydraulic Fracturing, Monograph Series Vol 2, SPE, Dallas, Texas (1970). Gidley , J.L., et al.: Recent Advances in Hydraulic Fracturing, Monograph Series Vol 12, SPE, Richardson, Texas (1989). Economides, M.J., and Nolte, K.G.: Reservoir Stimulation, Schlumberger Educational Services, 1987.

Page 71


BJ Services’ Frac Manual 9. Fracture Mechanics

9.

Fracture Mechanics

Fracture mechanics is the study of how fractures propagate through a material. The aim of fracture mechanics is to predict how fast a crack will grow, and at what point the fracture becomes “critical” – i.e. the fracture will suddenly spread across the entire material causing catastrophic failure. In hydraulic fracturing we use fracture mechanics to predict how far our fracture will grow – both horizontally and vertically. When reading this section, it should be remembered that stress and pressure are essentially the same thing. This means that a pressure in a fracture puts a stress of equal magnitude in the formation at the fracture face, in a direction perpendicular to the fracture face. Therefore, when the fracture is propagating, the critical stress (the stress needed to make the fracture grow) has to be equal to the net pressure.

9.1

Linear Elastic Fracture Mechanics and Fracture Toughness

Linear Elastic Fracture Mechanics (LEFM) is all about the prediction of how much stress (i.e. energy) it takes to make a fracture propagate. LEFM assumes linear elastic deformation (constant Young’s modulus) followed by brittle fracture – it is assumed that no significant energy is absorbed by non-linear or non-elastic effects. That is to say, energy stored as stress in the material is transferred directly to fracturing the material, and no energy is lost to plastic deformation. LEFM was used almost exclusively in the earlier fracture models (see Section 8), and is still used – to a greater or lesser extent - in a number of fracture models currently available in the industry (e.g. MFrac, StimPlan – see Section 11).

The Griffith Crack The first person to adopt a meaningful analytical approach to studying the mechanics of fracture propagation was Griffith, in the 1920’s. Figure 9.1a illustrates the concept of the Griffith crack, which can be expressed with the following Equation; δU δa

2

=

2πσ a E ..............................................................................9.1

σ

2a

σ Figure 9.1a – The Griffith crack

Page 72


BJ Services’ Frac Manual 9. Fracture Mechanics where U is the elastic energy (i.e. the energy used to produce elastic stress on the material), a is the characteristic fracture length, σ is the far field stress (i.e. the “bulk” stress away from the influence of the fracture) and E is the Young’s modulus. Therefore, Equation 1 describes the amount of additional energy (δU) required to make the fracture grow from length a to length a + δa. Usually, δU/δa is replaced by 2G. G is referred to as the “elastic energy release rate” and also the “crack driving force”, such that; G

πσ2a

= E

............................................................................. (9.2)

In order to reach this relationship, Griffith makes a significant assumption – that there is no energy lost at the fracture tip and not used to propagate the fracture. Energy is used either to elastically deform the material or to rupture the material. Therefore, there can be no plastic deformation at the tip, and the Griffith model is only applicable to materials liable to elastic deformation followed by brittle fracture.

Griffith Failure Criterion Given that for a uniform material with constant geometry δU/δa is a constant, there is a critical value of stress, σc, at which the material will experience catastrophic failure, i.e. the fracture propagates at high velocity across the material causing failure. This critical stress is defined as follows;

σc

=

2

EGIc ....................................................................... (9.3) πa

The critical energy release rate, Glc, is determined experimentally and is a material property, although it will vary with both temperature and the overall geometry of the test specimen. Equation 9.3 also defines - for a given stress - a critical fracture length. If the fracture is less than this critical length, the material will not fail. However, if the fracture grows above this critical length, the material will fail. The subscript "I" refers to the failure mode, as illustrated in Figure 9.1b. Failure mode I is the "opening mode", mode II is the "sliding mode" and mode III is the "tearing mode". In hydraulic fracturing, we are usually only concerned with failure mode I. Mode I Opening

Mode II Sliding

Mode III Tearing

Figure 9.1b – Failure modes in Linear Elastic Fracture Mechanics

Page 73



Stress Intensity Factor With reference to Figure 9.1c, the stresses in the principal directions, at some point away from the fracture tip, can be expressed as follows;

σxx

=

K 3θ θ θ cos 2 1 - sin 2 sin 2 ...................................... (9.4) 2πr

σyy

=

K θ θ 3θ cos 2 1 + sin 2 sin 2 ...................................... (9.5) 2πr

and

where K is the stress intensity factor.

σ

y

r Fracture

θ x

a

σ Figure 9.1c – Coordinate system for stress intensity factor

Considering the plane strain situation (i.e. εzz = 0, an object with a thickness large enough to make strain on the z-axis negligible), and the case that a >> r, then the stress in the ydirection – “across” the line of the fracture (i.e. θ = 0) – can be expressed as follows;

σyy

=

K ........................................................................... (9.6) 2πr

Obviously in Equation 9.6, as r tends to 0, σyy tends to . This represents a fundamental flaw in this approach to modelling fractures – it fails close to the fracture tip. Using this approach, K is the only factor that affects the magnitude of the stress at a given distance from the fracture tip. Whilst K is a material property, it is also a variable, depending upon the gross geometry of the fracture and its surroundings, as well as temperature. Assuming a constant temperature in any given instance, relationships linking K, a and σ for most situations have been solved, either analytically or numerically. At material failure, σc can be described in terms of a critical stress intensity factor, KIc, which is more commonly referred to as the fracture toughness;

σc

=

KIc ........................................................................... (9.7) β πa

This is the fundamental Equation of Linear Elastic Fracture Mechanics, where β is a geometrical factor and is equal to 0.4 for a radial fracture. KIc is related to GIc as follows; Page 74


BJ Services’ Frac Manual 9. Fracture Mechanics 2

GIc

2 KIc = (1 - ν ) E ................................................................... (9.8)

For a given geometry, the fracture toughness is a material property. Equation 9.8 shows that it represents the amount of mechanical energy a material can absorb before it fails by brittle fracture. Put simply, a material with a low KIc is brittle and a material with a high KIc is tough. 2 The term E/(1 - ν ) is often referred to as the plane strain Young' s modulus or E' , so that Glc is equal to Klc divided by E' .

9.2

Non-Linear and Non-Elastic Effects

From the extensive work done in this field, it is clear that LEFM alone does not adequately account for the pressure needed to make the fracture grow. There is a tip over pressure effect, which means that more pressure (energy) is required than is predicted by LEFM. Two possible – and not necessarily mutually exclusive – theories for this are described below.

Crack Tip Dilatency The theory of crack tip dilatency was first put forward by Cleary et al, and has been used extensively by them in the FracPro model. This approach has almost entirely done away with the concept of fracture toughness, which means that users of simulators based on this model find that changes to input fracture toughness values have little or no effect on fracture geometry. Instead the theory states that deep underground, the effect of the confining stress is much more significant than the effect of the fracture toughness. Thus KIc can be ignored if the following condition is satisfied;

σ π R >> KIc ..................................................................................... (9.9) where R is the radius of the fracture and is analogous to the LEFM characteristic fracture length. Equation 9.9 shows us that fracture toughness is more significant for small fractures in shallow formations, such as during skin bypass fracturing. The fracturing fluid does not penetrate to the very end of the fracture. This means that there is a very rapid change in net pressure at a distance ω from the tip of the fracture, as illustrated in Figure 9.2a.

Pnet

Dilation Contribution

ω

r Figure 9.2a – The Cleary et al approach.

If the condition described in Equation 9.9 is satisfied, then ω can be found as follows;

Page 75



ω

R

≈ 2

2 Pnet ...................................................... (9.10) Pnet + Pc

Because the fluid does not penetrate into the tip of the fracture, energy is lost as the tip of the fracture deforms. It is postulated that this deformation occurs in a non-linear or dilatent fashion. This crack tip dilatency reduces the energy left for the fracturing fluid to propagate the fracture, and hence reduces the size of the fracture, for a given Pnet.

Crack Tip Plasticity The crack tip plasticity theory allows for a significant region of plastic deformation at the fracture tip. All materials exhibit some level of plastic deformation prior to failure – it is assumed in LEFM, and most of the other approaches to modeling hydraulic fracturing, that this is not significant. This may be true in some formations. However, there may be a wide range of circumstances under which significant plastic deformation is not only possible, but probable. Even so-called brittle materials can experience plastic deformation when exposed to extreme tri-axial stresses. As the load on a material containing a fracture increases, the stresses around the fracture tip also increase. Because of the geometry of the area of the fracture tip, these stresses are usually far in excess of the overall stress on the material - as illustrated in Equation 9.6. As the overall stress increases, the stress around the fracture increases to a point where it exceeds the yield point of the material (σy). The material then starts to plastically deform, and to move in a direction that will relieve the stress – away from the crack tip. This produces a crack tip of finite radius, as opposed to the infinitely small fracture tip modeled in LEFM. The diameter of the fracture tip, d, is given by the following Equation: 2

d

=

2

KI (1 - ν ) .................................................................... (9.11) Eσy

For a long, narrow fracture, having a tip of finite radius can significantly reduce the overall length of the fracture. This is illustrated in Figure 9.2b:-

σyy σy

rp

Fracture

r

d The plastic zone

Figure 9.2b – Crack tip diameter and the plastic zone. Note that rp is the radius of the plastic zone.

Figure 9.2b shows the plastic zone as a circle around the fracture tip. However, this is not necessarily the case. By using the principle stresses given by Equations 9.4 and 9.5, and assuming plane strain (εzz = 0), the von Mises yield criterion gives the following result:

Page 76



σy2πrp

1 =4

2

KI

3 2 2 2 sin θ + (1 - 2ν) (1 + cosθ ) ............................ (9.12) 2

By plotting, in polar coordinates, σy2πrp/KI (dimensionless plastic radius) against θ, we can see the shape of the plastic zone at the fracture tip, as illustrated in Figure 9.2c. This produces two plastic “ellipses” either side of the fracture plane. 0.5

σy2 π r p

ν = 0.25

2

KI

θ -0.5

0.5

-0.5

Figure 9.2c – The shape of the plastic zone, for a Poisson’s ratio of 0.25 (after Martin, 2000)

In hard rocks, the actual size of the plastic zone is quite small, compared to the volume of the fracture. However, as Young’s modulus and yield stress decrease, the relative size of the plastic zone increases until it reaches a relatively large volume. At this point, the energy absorbed by the plastic deformation of this volume becomes a significant fraction of the energy contained in the fracturing fluid. This means that in a formation liable to significant plastic deformation, it requires significantly more energy to propagate the fracture than is predicted by LEFM. As discussed below, if the fracture tip takes more energy, the fracture will be smaller and will have less width.

9.3

The Energy Balance

The process of propagating a fracture through a formation is all about the transfer of energy from the frac pumps to the formation. Energy transfer occurs as shown in Figure 9.3a. Reducing all the processes occurring in the creation of a fracture to energy, allows them to be related to each other in the most fundamental fashion. To start with, we must remember that pressure and stress are essentially energy per unit volume. Therefore, the total energy per unit time (a.k.a. power) in the fluid available for creating a fracture is:U ˙

= BHTP.Q ...................................................................... (9.13)

Not forgetting that:BHTP

= STP + HH – Pfrict ....................................................... (9.14)

Therefore, the total energy available to the fracturing fluid is given by:-

Page 77


Page 78

The Frac Fluid At the Wellhead STP

Energy from Hydrostatic Head

By the Perforations BHTP

Friction Pressure in the Surface Lines and W ellbore

In the Near W ellbore Region, P net

Perforation Friction & Tortuosity

Moving Down the Fracture, P net

Fluid Leakoff

Ufluid =

Overcoming the Closure Stress of the Formation

Compression of Formation to Produce Frac W idth Fluid Friction in Fracture

At the Fracture Tip P net- Fluid Friction in Fracture

Propagation of Fracture

tp U ˙ dt.................................................................... (9.15)

0

Figure 9.3a – Sources of Energy Gains and Losses for the fracturing fluid. Energy Gains + Energy Losses = 0.


Chemical Energy from the Diesel Fuel is Changed into Mechanical Energy by the Engine, which is Changed into Pressure and Kinetic Energy by the Frac Pump


Energy Losses

Energy Gains

BJ Services’ Frac Manual 9. Fracture Mechanics Where tp is the total pumping time. Equation 9.15 looks intimidating, but it is simply the area under the graph of (bottom hole) horsepower against time. A substantial portion of the energy is used up, simply by overcoming the in-situ stresses of the formation. Another portion of the energy is used up in overcoming friction in the near wellbore area. Therefore, the final amount of energy available for fracturing the formation is given by:= tp P Q dt................................................................ (9.16) U fluid

0

net

Given that in most cases the rate is relatively constant, a plot showing Pnet versus time can show a great deal about how much energy is being used to create the fracture. This is the basis of Nolte analysis (see Section 10.2). Most fracture simulators spend a great deal of time quantifying these energy loses and gains, so that the amount of energy left in the fracturing fluid for propagation and the production of width can be a found. If the Young’s modulus is known, the fracture width – for a given Pnet – can be easily determined. This then leaves the amount of energy available for the propagation of the fracture, which in turn defines how big the fracture gets. This is the ultimate goal of the fracture simulator.

References Griffith, A.A.: “The phenomena of rupture and flow in solids”, Phil. Trans. Roy. Soc. of London, A 221 (1921), pp. 163 – 167 th

Broek, D.: Elementary Engineering Fracture Mechanics, Kluwer Academic Publishers, 4 Ed. (rev), 1986. Economides, M.J., and Nolte, K.G.: Reservoir Stimulation, Schlumberger Educational Services, 1987. Cleary, M.P., Wright, C.A., and Wright, T.B.: “Experimental and Modeling Evidence for Major Changes in Hydraulic Fracturing Design and Field Procedures”, paper SPE 21494, presented at the SPE Gas Technology Symposium, Houston TX, Jan 1991. de Pater, C.J., Weijers, L., Savi , M., Wolf, K.H.A.A., van den Hoek, P.J., and Barr, D.T.: “Experimental Study of Nonlinear Effects in Hydraulic Fracture Propagation”, paper SPE 25893, SPEPF, Nov. 1994, pp. 239 – 246. Dugdale, D.S.: “Yielding of steel sheets containing slits”, J. Mech. Phys. Sol., 8, 1960, pp. 100 – 108. Martin, A.N.: “Crack Tip Plasticity: A Different Approach to Modeling Fracture Propagation in Soft Formations”, paper SPE 63171 (revised), presented at the SPE Annual Technical Conference and Exhibition, Dallas TX, Oct 2000.

Page 79


BJ Services’ Frac Manual 10. Advanced Concepts

10.

Advanced Concepts

In this section we shall deal with some of the more advanced concepts used in the process of designing hydraulic fracture treatments, as well as in diagnosing what may (or may not) have happened during a frac or minifrac treatment.

10.1

Tortuosity

Hydraulic fractures are created by pressure, not rate. Often we use rate to help generate the required pressure, but we shouldn’t loose sight of the fact that it’s pressure that splits the rock. Over a long perforated interval, fractures can form anywhere that the fluid pressure exceeds the local frac gradient. Generally, the rock will have one point that is weaker than the rest and the initial fracture will form here. However, if the pressure continues to rise, additional fractures may be formed. Potentially, every single perforation is a source of fracture initiation. Many of these fractures will be very small – but some may be large enough to take a significant proportion of the treatment fluid. Away from the artificial stress environment around the wellbore, treatments tend to produce a relatively small number of larger fractures. Normally, fractures do not tend to join together – the stress regime around the fracture tip tends to keep fractures apart. However, under the influence of the complex stresses around the wellbore and perforations, fractures can join together, sometimes giving several narrow paths towards a single, large fracture. So the treating fluid has to travel from a region containing a large number of small fractures to a region containing a small number of large fractures. In doing so, the fluid has to move through a series of convoluted, narrow fractures – or put another way, through a tortuous path. This tortuosity can produce a significant loss in pressure, resulting in a smaller than expected fracture and possible early screenouts. Screenouts can also be caused by tortuosity for another reason – the width of these channels through the rock is often not large enough to carry the proppant concentration passing through it. This causes the proppant to bridge off, preventing any further flow of proppant. Tortuosity manifests itself as a pressure drop through the near wellbore region. There are also other phenomenon that can result in a near wellbore pressure loss (such as poor quality perforations). However, the important point is that there is a loss of pressure, which can be a substantial proportion of the observed net pressure (i.e. the total energy available to propagate the fracture). Because the pressures inside the fracture drive the pressures at the surface, the pressure loss due to the tortuosity actually produces a higher BHTP and hence higher STP. This gives the surface observer the impression that the net pressure is higher than it really is. For instance, for a well with 200 psi net pressure and 300 psi pressure loss due to tortuosity, it appears, to an Engineer who is unaware of the tortuosity, that the net pressure is 500 psi. This means that Engineer thinks that the frac fluid has much more energy for creating fracture volume than it has in reality, potentially resulting in a treatment design that contains more proppant than can physically fit into the fracture. It is therefore important to understand the magnitude of the near wellbore pressure loss, so that this can be allowed for when designing the treatment. Hard Rocks (that is, rocks with a high Young’s modulus and low fracture toughness) tend to be more susceptible to tortuosity than soft rocks. In this type of brittle formation, there is already a fracture formed at each perforation by the explosive action of the perforating charges – all we are doing when we pump fluid is making these fractures extend, through a medium that allows easy fracture extension. Because of the high Young’s modulus, the stress concentrations at the fracture tip are more intense and so these smaller fractures are less likely to link up. This means that hard rocks are more likely to produce a large number of small fractures than soft rocks. Deviated Wellbores tend to be more susceptible to tortuosity than vertical wells. As fractures propagate, they compress the rock either side of them. This makes it harder for other Page 80


BJ Services’ Frac Manual 10. Advanced Concepts fractures to propagate in this region. As discussed in section 7, fractures tend to propagate on a vertical plane. This means that the more deviated the wellbore, the less each fracture interferes with its neighbour, and so they propagate more easily. Additionally, as each fracture is further apart, there is less joining together of fractures. Finally, there is some evidence to suggest that on some deviated wells, the fracture can initiate along the wellbore. At some point not too far from the wellbore, the fracture grows out far enough so that the influence of the wellbore is less significant than the influence of the in-situ stresses. At this point the fracture has to change its orientation, rapidly if the rock is very hard. This produces a “corner” around which the fluid and proppant has to flow, which causes further loss of pressure. Thus, highly deviated wells in hard rocks are more likely to experience tortuosity problems than vertical wells in soft formations. This does not mean that significant tortuosity will not be encountered in soft formations or in vertical wells – it simply means that it is less likely. Perforation Strategy. Often Service Companies are asked to treat wells which are already perforated. In such wells, it is very difficult to control fracture initiation. However, sometimes the well to be treated is new and we can perhaps influence the perforation strategy. This can have a significant effect on the tortuosity, and is explained in detail in Section 14. Horizontal Stress Contrast. As illustrated in Figure 10.1a, the contrast between the maximum and minimum horizontal stresses can also influence the tortuosity. For the left hand side of Figure 10.1a, there is a large contrast between σh,max and σh,min. This produces a narrow fracture close to the wellbore and a very tight radius turn for the fracture. For the right hand side of Figure 10.1a, there is little difference between the two horizontal stresses, so the fracture starts with a wider width and gradually changes direction. Therefore, depending upon the initial fracture orientation (which is turn is affected by the perforation strategy and wellbore deviation), the contrast between horizontal stresses can have a significant effect on tortuosity.

σh,max ~= σh,min

σh,max >> σh,min

σh,max σh,min Figure 10.1a. Diagram illustrating the effects of horizontal stress contrast on tortuosity (after GRI-AST 1996).

Curing Tortuosity. If tortuosity is detected before the main treatment (see Sections 15 and 16), it can sometimes be cured. This is done by pumping proppant slugs. The first company to successfully accomplish this on a regular basis was Mærsk Olie og Gas, a Danish company operating in the North Sea. Several SPE papers have been produced by Mærsk and their contractors to document this. Mærsk had the advantage that they were operating off a large frac boat, mixing gel with seawater on the fly. This meant that they had an effectively limitless supply of both gel and proppant at their disposal – most of the time this is not the case. To start with, Mærsk would pump a proppant slug in the minifrac, ideally at the maximum anticipated proppant concentration for the main treatment. If this slug passed into the formation without a significant rise in pressure, they could be reasonably sure that the tortuosity would not significantly affect the treatment. Sometimes they would pump a series of slugs, mixed at increasing proppant concentrations. If these slugs encountered a significant

Page 81


BJ Services’ Frac Manual 10. Advanced Concepts rise in pressure, or worse still screened the well out, they knew they had a problem. The cure was to deliberately screen the tortuosity out. This is done by pumping proppant slugs and then shutting down with the slug in the perforations and near wellbore region. The effect of this is to block up the narrow channels and force open the wide channels. After a few years, Mærsk became so proficient at this – and so familiar with their formations – that they developed a standard method used on every treatment. This involved pumping a relatively long stage of 100 mesh sand at 1 or 2 ppa during the minifrac, followed by a relatively short stage with 20/40 proppant at 4 or 5 ppa. The minifrac was shut down with the 20/40 proppant in the perforations. The 100 mesh sand blocked the narrow channels, whilst the 20/40 proppant held open the wide channels, so that they would accept fluid when the main treatment started. Using this method, Mærsk achieved a near perfect record for placing treatments, in an area notorious for tortuosity problems.

10.2

Nolte Analysis

Nolte analysis is a branch of frac theory originally developed by Ken Nolte of Amoco in the early 1980’s. It uses a plot of log Pnet against log time to determine the shape of the fracture, as illustrated in Figure 10.2a:-

log Pnet

b II I

a

III

IV

log (job time) Figure 10.2a – The Nolte plot

Basically, pressure is stored energy – or in the case of the fracturing fluid, stored energy per unit volume. As work (a.k.a. horsepower) is the rate of using energy, on a graph of pressure against time the gradient is the amount of work being performed. In this case, it is the amount of work being performed by the fracturing fluid on the formation. Nolte used a mathematical analysis to show that at certain gradients on the log Pnet against log job time plot, certain fracture geometries will apply (with reference to Figure 10.2a):Mode I Mode II Mode IIIa Mode IIIb Mode IV -

Page 82

Good height containment, fracture propagates preferentially in the horizontal direction. Even fracture growth, fracture is propagating elliptically with vertical as well as horizontal growth. Screenout, fracture is filling with proppant and is having to balloon in order to cope with the volume of fluid entering the fracture. Screenout, near wellbore event. It is no longer possible to pump proppant into the fracture. Uncontrolled height growth. Also radial fracture geometry.


BJ Services’ Frac Manual 10. Advanced Concepts Nolte’s work was carried out with respect to the three main 2-D models that were widely used at the time. However, it is still a useful tool for the Frac Engineer to assess fracture geometry without using a fracture simulator, or as a back up to a simulator. Nolte analysis became popular at the same time that Service Companies began to use computer monitoring and data storage systems on location. It became possible to have a Nolte plot running real time – providing the industry with its first real-time fracture simulation and diagnosis tool.

10.3

Dimensionless Fracture Conductivity

Dimensionless Fracture Conductivity (FCD or – as recently redefined by the API - CfD) or Relative Fracture Conductivity is a measure of how conductive the fracture is compared to the formation. In order to produce a production increase, the propped fracture has to be more conductive than the formation (setting aside the effects of bypassing the skin damage). In Section 2 we defined the fracture conductivity (FC) as being the product of the fracture width and the permeability of the proppant. Dimensionless fracture conductivity is defined as follows:CfD

Fc =x k f

=

¯ kp w xf k ......................................................... (10.1)

¯ is the average where xf is the fracture half length, kp is the permeability of the proppant, w fracture width and k is the permeability of the formation. In order for the fracture to be more conductive than the formation, the dimensionless fracture conductivity has to be greater than one. Equation 10.1 compares the ability of the formation to deliver fluids to the fracture, with the ability of the fracture to delivery fluids to the wellbore. If the CfD is less than one, then post treatment production increase is limited by the relatively low conductivity of the fracture, and the fluids will flow more easily through the formation. If the CfD is significantly greater than 1, then the limiting factor is the formation’s ability to deliver hydrocarbons to the fracture. Of the four components on the right hand side of Equation 10.1, the permeability of the formation is fixed, whilst the permeability of the proppant is defined by the proppant type, the closure stress and the producing conditions. In order to maximise CfD, it is necessary to control the fracture half-length, whilst at the same time getting the width and the proppant permeability as large as possible. Under most circumstances – for any given fracture situation – there is a fixed relationship between width and length. For so much length created, there will be so much width created. However, created width is not the same as propped width, unless the well has screened out from tip to wellbore. The more proppant that is placed per unit area of the fracture, the wider the propped fracture will be. Therefore, two ways to increase CfD are; one – pump more proppant; or two – pump better quality proppant. In higher permeability formations, this is not enough. Even with the fracture completely full of good quality proppant, the CfD can still be less than one. Therefore, a technique called the Tip Screen Out must be used (see section 10.4, below).

10.4

The Tip Screen Out (TSO)

The Tip Screen Out is a technique used to artificially increase the width of the fracture, without increasing the length. As previously discussed, for any given fracture, there is a fixed relationship between width and length. If we can artificially overcome this, then we can dramatically increase the CfD. Figure 10.4a illustrates this:-

Page 83



Proppant

Fracture Tip

∆P Figure 10.4a – The Tip Screen Out

The TSO is a technique that is generally used in high permeability formations. The high formation permeability means that it is very difficult to get a CfD greater than one. In order to generate the TSO, proppant is pumped into the fracture earlier than would normally be the case. As the formation has high permeability, the fracturing fluid is leaking off relatively quickly. This acts to dehydrate the proppant-laden slurry. If the treatment is correctly designed, this dehydration will cause the proppant to collect at the fracture tip. In order for the fracture to continue to propagate, a positive Pnet must be maintained at the fracture tip. As the proppant builds up in the fracture tip, fluid has to flow through it to reach the tip and maintain the Pnet. Whilst flowing through the proppant build-up, the fluid loses pressure due to friction as it passes between the proppant grains. When the proppant build-up gets large enough, the ∆P of the fluid equals and then exceeds Pnet and the fracture ceases to propagate. At this point, fluid is still being pumped into the fracture and has to go somewhere. Some of this fluid is leaking off, but not all of it – so the fracture volume still has to grow. This means that the fracture starts to get wider. It also means a rise in net pressure as the formation gets increasingly compacted – this is how the onset of a TSO is spotted during a treatment. The TSO technique relies on two things; high permeability (and hence high fluid leakoff), and low Young’s modulus. High leakoff is necessary so that the slurry will dehydrate sufficiently to allow proppant build-up at the tip. Low Young’s modulus is necessary to allow the width to increase. If the formation is too hard (i.e. Young’s modulus too high), the pressure will rise very rapidly and quickly exceed the maximum treating pressure at surface.

10.5

Multiple Fractures and Limited Entry

As previously discussed, any perforation is potentially a source of fracture initiation. All it takes is for the fluid pressure to exceed the fracture extension pressure at any given point and a fracture is formed. How large that fracture is depends upon the volume of fluid the fracture receives. Usually, most of the small fractures get “squeezed out” as larger fractures close by develop. However, if the fractures are far enough apart (which is easy enough on deviated wellbores), more than one fracture will develop into a significant size. This is often detrimental, as multiple fractures that cover the same vertical plane are largely wasted, unless they are widely spaced out. In addition, as the rate (and hence frac fluid volume) is split between two or more fractures, the treatment ends up producing a range of smaller, narrower (i.e. less conductive) fractures, rather than a single large fracture. Finally, although each fracture receives only a fraction of the total rate, the proppant concentration remains unchanged. As the fracture width is less, and the slurry velocity down each individual fracture is decreased, there is a much greater chance of proppant bridging and a premature screenout. In short, multiple fractures can lead to less effective stimulation and an increased chance of job failure. The majority of wells worldwide are completed with more than one set of perforations. Unless something is done to isolate these perforations and control the point of fracture initiation, multiple fractures are likely. However, there is one situation where this is deliberately used to produce stimulation of an entire interval at one go. This is called Limited Entry fracturing. Page 84


BJ Services’ Frac Manual 10. Advanced Concepts Limited Entry Fracturing. Even whilst fracturing, fluids follow the path (or paths) of least resistance. The resistance to the flow of the fluids comes from three sources:- perforation friction; tortuosity; and the formation’s fracture extension pressure. All of these can vary with fluid rate. However, the fracture extension pressure and tortuosity are not controllable, whereas the perforation friction is. Therefore, if the fracture extension pressure of each formation is known, as well as the tortuosity (usually assumed to be zero), the number and size of the perforations can be varied to balance the fluid flow, so that each set of perforations receives the same proportion of fluids. This technique is called Limited Entry, as we are trying to limit and control the amount of fluids entering each zone. This technique can be taken one step further. By further varying the number of perforations, the proportion of each fluid going into each zone can be adjusted to produce the optimum treatment for that zone – more fluid enters zones needing most stimulation, for example. Obviously, the calculations for working out the size and number of perforations can get pretty complex – once there are more than two zones you need a computer model to keep things straight. In addition, the results are only as good as the data input – if you are guessing at the frac gradient, then you are also guessing at the number of perforations needed. Finally, this analysis also assumes perfect perforations – something that cannot be guaranteed. Therefore, limited entry fracturing is unreliable unless exact data is available. In addition to being unreliable, limited entry fracture treatments tend to be very big. The treatment is trying to place effective fractures in several zones simultaneously. This means lots of rate and large fluid volumes, as well as lots of proppant, as this treatment is trying to do the work of several smaller treatments in one go.

10.6

Proppant Convection and Settling

Proppant Convection. Proppant Convection is caused by variation in slurry density, and can lead to the majority of the proppant being placed in the bottom of the fracture. Put basically, a 10 ppa slurry is much denser than a – for instance – 5 ppa slurry. This means that if a 10 ppa slurry follows a 5 ppa slurry into the formation, it will tend to slide beneath the lighter slurry, leading to the placement of the higher proppant concentration at the bottom of the fracture, where it may not necessarily connect with the perforations. This is illustrated in Figure 10.6a. Obviously, proppant convection is not really an issue on TSO designs, as the plan is to completely fill the fracture from tip to wellbore. However, when fracturing lower permeability formations, proppant convection can cause significant problems. The way to prevent this is to use long proppant stages mixed at the same concentration. Once in the formation, slurries will dehydrate with time due to leakoff - increasing the ppa of the slurry - so it may be necessary to gradually increase the proppant concentration at the blender as the treatment progresses.

Figure 10.6a – Proppant convection. As the heavier slurry enters the fracture it sinks and displaces the lighter slurry upwards

Page 85


BJ Services’ Frac Manual 10. Advanced Concepts Proppant Settling. Proppant settling occurs when the frac fluid has insufficient viscosity to suspend the proppant inside the fracture. Proppant moves downward, leading in the worst cases to a fracture that only has proppant right at the bottom. This may be completely unconnected to the wellbore. Once again, this phenomenon is not an issue when a TSO treatment is being performed. However, on lower permeability formations, especially those with very long closure times, settling can be a significant issue. The key to preventing proppant settling is to design the frac fluid correctly. In order to prevent settling, the frac fluid must exhibit good proppant transport qualities at BHST for at least the anticipated job time, plus the anticipated closure time, plus a safety factor. This can be tested by the use of the model 50 high temperature rheometer. A widely accepted criterion for -1 proppant transport is to have at least 200 cp apparent viscosity at a shear rate of 40 sec . Note that this criterion is not an API standard and is somewhat subjective – different standards are used in different places. Equation 10.2 gives an Equation for calculating the terminal velocity (i.e. the maximum possible velocity) for a spherical particle falling through a power law fluid (note that this assumes the fluid is almost at rest):vt

n' +1 (SGp - SGf) 1 n' 0.04212 dp = 36 ............................ (10.2) K'

where vt is the terminal velocity (ft/sec), dp is the proppant grain diameter (inches), SGp is the proppant absolute specific gravity, SGf is the fluid specific gravity, n’ is the flow behaviour n’ –2 index (dimensionless) and K’ is the consistency index (lbs.sec ft ).

10.7

Proppant Flowback

Proppant flowback is when the proppant that been placed in the fracture flows back into the well during production. It has been the subject of intense industry debate and investigation over the last 10 years. Some of the causes of proppant flowback are listed below:i)

Stress Cycling. Every time the well is drawn down, the closure stress on the proppant increases, as the reservoir pressure in the fracture is effectively reduced. When the well is shut in, the pressure builds up again and the closure pressure is reduced. This is stress cycling, which was first identified in 1994 by Shell and StimLab as a major cause of proppant flowback. As the well is opened and closed, the proppant pack expands and contracts slightly, weakening its integrity. If the stress is cycled enough times – or too suddenly – the pack will literally break apart, allowing proppant to flow back into the wellbore. Wells that have been fractured should be handled with care – don’t shut them in unless there is no alternative, and if it has to be done, then it should be done slowly.

ii)

Weak Formations. Obviously, if the formation holding the proppant in place falls apart, then the proppant will flow back. Formations that are susceptible to this need to be frac and packed, rather than just fractured.

iii)

Insufficient Fracture Conductivity. If the propped fracture has insufficient conductivity, especially in the near wellbore area, then the higher velocity of the produced fluids, coupled with the increased pressure gradient along the plane of the fracture, will result in an increased net force acting to push individual proppant grains out if the fracture.

iv)

Poor Quality Frac Fluid. If the frac fluid does not have sufficient viscosity to keep the proppant suspended until the fracture closes, then the proppant will settle into the bottom part of the fracture. In extreme cases, this can result in the bottom half of the fracture having all the proppant, whilst the top half closes up on nothing. This creates a void space at the top of the proppant pack, as illustrated in Figure 10.7a.

Page 86



No Proppant Void Space

Proppant

Figure 10.7a – Illustration of the “Pipelining” effect.

As the well is produced, fluid flows rapidly across the top of the proppant pack, through the void space, as this is the path of least resistance. As it does so, it picks up proppant grains and can carry these out of the fracture and even up to the surface. This effect is known as “pipelining” and can result in almost all of the proppant being produced back out of the fracture.

Preventing Proppant Flowback Once proppant flowback has started, it is usually very difficult to stop. Therefore, the best option is to prevent proppant flowback from happening in the first place. Obviously, a well designed treatment using a good quality frac fluid, together with good well management, can go a long way to mitigating proppant flowback. However, it is also true that for some formations, this is not sufficient. To combat this, there are several different methods which can be employed:i)

Resin Coated Proppant. By far the most common method for controlling proppant flowback, resin coated proppant (RCP) is simply proppant which has been coated with a layer designed to make the proppant grains stick together. Usually, it requires temperature and a closure stress for this to happen. RCP tends to come in two main varieties, curable and pre-cured (or tempered). Curable RCP has a softer coating, which is designed to chemically cure when exposed to temperature. Pre-cured RCP has a harder resin coat, which relies more on the closure pressure to make the proppant grains stick together. RCP has an additional effect, in that it makes the proppant more tolerant to closure pressure, as the resin coat will capture permeability-reducing fines produced as the fracture closes. RCP is generally used as an alternative to ordinary proppant, either for the whole treatment, or for the last few proppant stages. This latter method, whilst being cheaper, is less reliable as there is no guarantee that the stage which is pumped last will be the stage that is positioned by the wellbore (see earlier section on Proppant Convection). RCP can be highly effective, but has three main disadvantages. First, it is expensive, often being more than twice as expensive as the non-coated proppant. Second, it can have a significant effect on the fracturing fluid, especially at high pH’s, as some of the resin is stripped off and dissolves in the fluid. Finally, the standard bulk pneumatic systems generally used for handling large volumes of proppant cannot be used for RCP, as it the resin coat can be chipped off.

Page 87


BJ Services’ Frac Manual 10. Advanced Concepts ii)

Micro-Fibers. Another method for preventing proppant flow back is to pump very small fibers with the proppant. These fibers, if used in sufficient quantity, will form a three dimensional mesh within the proppant pack, acting to prevent individual grains slipping past one another. The use of these fibers can result in a slight decrease in proppant pack permeability, but this can be allowed for in the frac design. The fibers are usually made from a polymer. The main problems with this system (other than its cost) are operational. Because the fibers are very small, they have a very high surface area to mass ratio. This in turn means that it can very difficult to actually mix the fibers into a fluid, especially on the fly during a treatment. Because of the large difference in specific gravity between the proppant and the fibers, it is also very difficult to mix the proppant and fibers together before adding them to the gel. The fibers also have a limited maximum temperature above which, they will disintegrate. This significantly reduces the number of wells that are suitable for this type of treatment. Finally, if used in the wrong proportions with the proppant (due either to poor design or ineffective mixing), the fiber itself can be produced out of the formation, sometimes resulting in a “hair ball” somewhere in the production facilities.

iii)

Micro-Sheets. In order to get around the patent held by one service company for the micro-fibers, a competitor introduced a product that uses small sheets or platelets of polymer, which act to wrap around the proppant grains. This has several effects. First, and unfortunately foremost, is a significant reduction in permeability of the proppant pack. Secondly, the sheets will form a three dimensional mesh, acting in a similar fashion to the micro-fibers. The sheets also act a little like a resin coat, in that they can cushion the proppant grains and tie up fines. Unfortunately, the micro-sheets also suffer from many of the same operational and temperature limitation problems experienced by the micro-fibers.

iv)

Deformable Particles, such as BJ’s FlexSand, is another approach. These particles, mixed at 10 to 15% by weight with the proppant, will deform – to a limited extent – around the proppant as the fracture closes. This acts to lock the proppant grains together and reduce the tendency for them to slide past each other. The deformable particles also have the effect of cushioning the proppant grains and increasing the grain to grain area of contact. This acts to increase the proppant pack permeability, by reducing the production of fines. The main disadvantage of the deformable particles is the extra equipment needed to handle it and mix it at the correct proportions. However, this is no worse than for the micro-fibers and the micro-sheets.

10.8

Forced Closure

Forced closure is a technique used to produce a very tight proppant pack in the near wellbore area. As soon as the treatment is finished, the well is opened up and flowed back at 0.5 to 1.0 bpm. This is before the fracture has closed and before the fluid has broken. Although the exact mechanism by which this prevents proppant flowback is not clear, there is sufficient empirical evidence to make this a valid technique, in a suitable formation. However, there are no methods for deciding which formations are suitable, apart from actually trying the technique. Modern treatment monitoring software and fracture simulators are set up to allow for forced closure. Many of them even allow input from a flowmeter placed on the flowback line whilst monitoring the post-treatment pressure decline.

Page 88



10.9

Non-Darcy Flow

Darcy defined fluid flow through a porous media, in terms of the flow dimensions, the fluid viscosity, the pressure differential and the permeability of the media, in an Equation that is fundamental to the oil industry:q

=

kh∆P

µ (re/rw)

....................................................................... (10.3)

However, this doesn’t tell the whole story. Whilst this Equation can be very reliable for fluid flow through relatively low permeability media (such as rocks), it does not take into account inertial flow effects. On the microscopic scale, the fluid is constantly changing direction as it moves through the pore throats and pore spaces. This represents a loss of kinetic energy, and so also an increased loss in pressure per unit distance. This effect is quantified in the Forcheimer Equation:-dP L

µv = k p

(1)

(2)

2

+ β ρ v ............................................................. (10.4) (3)

The term –dP/L is the pressure drop per unit length along the propped fracture, µ is the viscosity, v is the overall “bulk” velocity of the fluid, kp is the permeability of the proppant pack, β is a constant (the “beta” factor, non-Darcy flow factor or turbulence factor), and ρ is the density of the fluid. In Equation 10.4, parts (1) and (2) are essentially the Darcy Equation. Part (3) is the nonDarcy term, and is basically kinetic energy per unit volume. Obviously, the effect of the nonDarcy term varies with the square of the velocity, so at lower flow rates (such as for oil flowing through a permeable rock) this effect is negligible. However, at high flow rates (such as for gas flowing through a highly permeable fracture) this term becomes highly significant and can produce a pressure gradient many times greater than that caused by Darcy flow. Obviously, the magnitude of the non-Darcy effect is also highly dependent upon the beta factor. The magnitude of beta is determined by a number of factors, but experimental determination of beta factors, has revealed two relationships:-

β

∝ D................................................................................. (10.5)

where D is the average grain diameter, and:-

β

∝

1 ........................................................................... (10.6) kp

It is also true that artificial proppants tend to have lower beta factors than naturally occurring sands, due to their greater sphericity and roundness. In practice, beta factors have been determined for a wide range of proppants and closure stresses, and can be easily obtained from the proppant manufacturers, such as in Table 10.1:Closure Stress, psi 2000 4000 6000 8000 10000

12/18 0.0001 0.0002 0.0007 0.0018

β, atm sec2/gram 16/20 0.0001 0.0002 0.0003 0.0007 0.0023

20/40 0.0002 0.0003 0.0004 0.0007 0.0015

Table 10.1 – Beta factor data for CarboLite artificial ceramic proppant (Carbo Ceramics Inc)

Page 89


BJ Services’ Frac Manual 10. Advanced Concepts As the expected production rate from the treatment increases, so does the pressure loss due to non-Darcy effects, and this should always be taken into account when selecting proppant and predicting production increase.

References Wright, C.A., Weijers, L., and Minner, W.A.: Advanced Stimulation Technology Deployment Program, report GRI-09/0075, Gas Research Institute, March 1996 Cleary, M.P, et al.: ”Field Implementation of Proppant Slugs to Avoid Premature Screen-Out of Hydraulic Fractures with Adequate Proppant Concentration”, paper SPE 25892, presented at the SPE Rocky Mountain Regional/Low Permeability Reservoirs Symposium, Denver CO, April 1993. Kogsball, H.H., Pits, M.J., and Owens, K.A.: “Effects of Tortuosity in Fracture Stimulation of Horizontal Wells – A Case Study of the Dan Field”, paper SPE 26796, presented at the Offshore Europe Conference, Aberdeen, UK, Sept 1993. Nolte, K.G.: “The Application of Fracture Design Based on Fracturing Pressure Analysis”, paper SPE 13393, SPEPE (Feb 1988) p31-42. Nolte, K.G., and Smith, M.B.: “Interpretation of Fracturing Pressures”, paper SPE 8297, JPT (Sept 1981) p1767-75. Gidley , J.L., et al.: Recent Advances in Hydraulic Fracturing, Monograph Series Vol 12, SPE, Richardson, Texas (1989). Vreeburg, R-J., Davies, D.R., and Penny, G.S.: “Proppant Backproduction During Hydraulic Fracturing – A New Failure Mechanism for Resin Coated Proppants”, paper SPE 27382, JPT, 1994. Ely, J.W.: “Experience proves forced closure works”, World Oil, Jan 1996, p 37 – 41. Rickards, A., et al.: “Need Stress Relief? A New Approach to Reducing Stress Cycling Induced Proppant Pack Failure”, paper SPE 49247, presented at the SPE Annual Technical Conference and Exhibition, New Orleans, Sept 1998. Forcheimer, P.: Wasserdewegung durch Boden. ZVDI (1901), Vol. 45, p. 1781. (in German) Martins, J.P., Milton-Taylor, D, and Leung, H.K.: “Effect of non-Darcy Flow in Propped Hydraulic Fractures”, paper SPE 20709 Vincent, M.C., Pearson, C.M., and Kullman, J.: “Non-Darcy and Multiphase Flow in Propped Hydraulic Fractures: Case Studies Illustrate the Dramatic Effect on Well Productivity”, paper SPE 54630, presented at the SPE Annual Technical Conference and Exhibition, Houston, Oct 1999.

Page 90


BJ Services’ Frac Manual 11. 3-D Fracture Simulators

11.

3-D Fracture Simulators

The three main fracture simulation models used in the industry today are FracPro, FracproPT and MFrac. Between them, they are used on well over 90% of all treatments currently performed. Other simulator’s, such as StimPlan, GOHFER and the proprietary simulators produced by Schlumberger, Halliburton, Shell and others, are available, but their use is limited mainly to Engineers who work for the actual company that produced the simulator. Most of the 3-D and lumped-parameter 3-D simulators described below are produced by companies whose main tasks are producing software or providing a fracturing service. As such, there is a considerable amount of detail concerning these simulators that is proprietary and not available in the public domain. Therefore, detailed descriptions of the actual algorithms behind the model are not possible and in any case beyond the scope of this manual. The reader is referred to the references for more information. The term pseudo or lumped-parameter 3-D is applied to most of the simulators, as they relate everything back to a single characteristic dimension (usually fracture half-length), which is found by a variety of methods. Fully 3-D models have every dimension as independent variables. As stated, most simulations are performed by one of three simulators (and it should be noted that FracPro and FracproPT are essentially the same model). In the industry, there is a perception that the FracPro-FracproPT model is more applicable to low permeability “hard” formations, whilst the MFrac model is more applicable to high permeability “soft” formations. The reliability of this perception is a matter of some debate, but it may be due to the respective origins of the two models. In any case, it should be remembered that the producers of these simulators are all competitors. Most of the discussions about the relative merits of each model are subjective and partisan. For a discussion on the limitations of the 3-D fracture simulators, refer to the discussion on pressure matching in Section 19.1.

11.1

RES’ FracPro and Pinnacle Technologies’ FracproPT

FracPro and FracproPT originally started out as one simulator, FracPro. The model was originally developed using funding from the Gas Research Institute, a joint US Government and gas industry funded organisation. Eventually, a company called Resources Engineering Systems (RES) produced a commercial simulator based on the work carried out by the GRI. However, a couple of years ago, a group of people dissatisfied with the company’s approach, split away from RES and moved over to Pinnacle Technologies. For a variety of reasons, they were able to take the FracPro technology with them and FracproPT was the result. At this point in time, there is very little difference between how the two models work. The major differences between the two models concentrate on the way they interface with the user, and with the on-screen graphics. The FracPro approach has almost entirely done away with the traditional concept of fracture toughness, which means that users of simulators based on this model find that changes to input fracture toughness values have little or no effect on fracture geometry. Instead the theory states that deep underground, the effect of the confining stress is much more significant than the effect of the fracture toughness. Thus Klc can be ignored if the following condition is satisfied;

Page 91


BJ Services’ Frac Manual 11. 3-D Fracture Simulators

σ pR

>> KIc ................................................................................ (11.1)

where R is the radius of the fracture and is analogous to the characteristic fracture length used in classical linear elastic fracture mechanics. The above Equation shows us that fracture toughness is more significant for small fractures in shallow formations, such as during skin bypass fracturing. The fracturing fluid does not penetrate to the very end of the fracture. This means that there is a very rapid change in net pressure at a distance ω from the tip of the fracture. If the condition described in the above Equation is satisfied, then ω can be found as follows;

ω

R

≈ 2

2 Pnet Pnet + Pc ....................................................... (11.2)

Because the fluid does not penetrate into the tip of the fracture, energy is lost as the tip of the fracture deforms. It is postulated that this deformation occurs in a non-linear or dilatent fashion (see Section 9.2). This crack tip dilatency reduces the energy left for the fracturing fluid to propagate the fracture, and hence reduces the size of the fracture, for a given Pnet. Once the energy absorbed by the fracture tip has been found, the model then goes on to solve the fracture geometry using a series of Equations which relate mass conservation, energy conservation, fluid dynamics and heat transfer. The model is 3-dimensional, allowing separate rock mechanical and reservoir properties to be input for each different rock strata. This model was the first to incorporate various aspects that are now taken as standard, such as near wellbore friction, proppant convection and multiple fractures. This model also incorporates a data conversion and editing facility, an acid fracture simulator and a simple production simulator. Although all of the three main simulators can model the fractures real time, only FracPro and FracproPT have the ability to predict forward to the end of a job, whilst in the middle of a treatment. This is a very powerful tool, which allows the fracture characteristics can be predicted in the middle of a treatment. The model takes the treatment data received up to that point, and then uses the remaining input treatment schedule to predict the fracture at the end of the treatment. Thus the Frac Engineer can “see how things are going” based on actual treatment data, and alter the treatment schedule as the job is being pumped. So far no other commercially available simulator has mastered this.

11.2

Meyers & Associates’ MFrac

MFrac is produced and developed by Meyer and Associates. This methodology sticks much more closely to the conventional Linear Elastic Fracture Mechanics (LEFM) approach to fracture propagation, than the FracPro/FracproPT approach. The model uses the basic LEFM criterion, which states that in order for the fracture to propagate, the stress intensity factor (K) must be greater than KIc (the critical stress intensity factor, or fracture toughness). It uses a characteristic length (referred to as Hξ) and a geometry factor, γ, in the classic LEFM Equation:-

σc

=

KIc ........................................................................... (11.3) γ Hξ

The actual value for g depends upon the fracture model being used (PKN, KZD, Ellipsoidal or 3-D), as does the dimension actually being used for the characteristic length. For the 3-D model, the characteristic length is found from a set of partial differential Equations, which relate mass conservation, mass continuity, momentum conservation, and vertical and lateral propagation rates. Page 92


BJ Services’ Frac Manual 11. 3-D Fracture Simulators The authors of this fracture propagation model acknowledge that there is a “tip over-pressure” effect that cannot be accounted for. This is handled by using an “over-pressure factor” - that has to be obtained empirically - or by using huge values for fracture toughness.

11.3

Other Simulators

StimPlan StimPlan is a pseudo 3-D numerical simulator produced by NSI Fracturing Technologies. The simulator works by performing implicit finite difference solutions to basic Equations of mass balance, elasticity, height growth, and fluid flow. The simulator is based on LEFM. It is probably the most widely used of the non-FracPro/FracproPT/MFrac simulators. Recently, NSI have started introducing E-StimPlan, a fully 3-D fracture simulator. This simulator divides the formation into a series of grids of variable size, allowing fully 3-D fracture growth and irregularly-shaped fractures (as opposed to the elliptical fractures almost always predicted by the lumped–parameter 3-D models). This model also allows 2 dimensional proppant transport. At the time of writing this manual, this simulator is still too slow for practical use, but shows great promise

GOHFER GOHFER (Grid Orientated Hydraulic Fracture Extension Replicator) has taken a completely different approach to modelling fracture growth. Of the four main models described, only GOHFER and E-StimPlan said to be fully 3-D and only GOHFER has a significant history of use. The model takes a finite element approach to fracture propagation, modelling the reservoir and the formations above and below it as a series of elements, rather than as a continuum. The fracture propagates along a plane between elements, so in order to produce fracture width, elements either side of the fracture have to be compressed. At the fracture tip, there is a single element just ahead of the fracture, so that the tip is positioned at some point on the side of the element. Fracture propagation occurs when the tensile stress in the element exceeds the failure criterion for the material, and the element splits into two pieces, along the plane of the fracture. The fracture has then propagated by a distance equal to the width of the element. The advantages of this approach are that it is very simple to give each element its own set of rock mechanical and reservoir properties, making simulation of multiple formations very easy. The main disadvantage is the use of a tensile failure criterion, which tends to make hard rocks harder to fracture than soft rocks, which tends to be the opposite way around to conventional theories. Additionally, because each element in the model can be assigned individual rock mechanical and leakoff properties, it is very easy to "dial-a-frac", that is, produce a fracture geometry that has more in common with uses wishes than with reality.

References Crockett, A.R., Okusu, N.M., and Cleary, M.P.: “A Complete Integrated Model for Design and th Real-Time Analysis of Hydraulic Fracturing Options”, paper SPE 15069, presented at the 56 California Regional Meeting of the SPE, Oakland CA, April 1986. Cleary, M.P., Wright, C.A., and Wright, T.B.: “Experimental and Modeling Evidence for Major Changes in Hydraulic Fracturing Design and Field Procedures”, paper SPE 21494, presented at the SPE Gas Technology Symposium, Houston TX, Jan 1991. Johnson, E., and Cleary, M.P.: “Implications of Recent Laboratory Experimental Results for Hydraulic Fracturing”, paper SPE 21846, presented at the SPE Rocky Mountain Regional Meeting and Low Permeability Reservoirs Symposium, Denver CO, April 1991. Page 93


BJ Services’ Frac Manual 11. 3-D Fracture Simulators Wright, T.B., Aud, W.W., Cipolla, C., Perry, K.F., and Cleary, M.P.: “Identification and Comparison of True Net Fracturing Pressures Generated by Pumping Fluids with Different Rheology into the Same Formations”, paper SPE 26153, presented at the SPE Gas Technology Symposium, Calgary, Alberta, Canada, June 1993. FracPro Version 8.0 onwards On-Line Help, RES/Gas Research Institute, March 1998 onwards. FracproPT Version 9.0 onwards On-Line Help, Pinnacle Technologies/Gas Research Institute, July 1999 onwards. Meyer, B.R.: “Design Formulae for 2-D and 3-D Vertical Hydraulic Fractures: Model Comparison and Parametric Studies”, paper SPE 15240, presented at the SPE Unconventional Gas Technology Symposium, Louisville KY, May 1986. Meyer, B.R.: “Three Dimensional Hydraulic Fracturing Simulation on Personal Computers: Theory and Comparison Studies”, SPE 19329, presented at the SPE Eastern Regional Meeting, Morgantown WV, Oct 1989. Meyer, B.R., Cooper, G.D., and Nelson, S.G.: “Real-Time 3-D Hydraulic Fracturing th Simulation: Theory and Field Case Histories”, paper SPE 20658, presented at the 65 SPE Annual Technical Conference and Exhibition, New Orleans LA, Sept 1990. Hagel, M.W., and Meyer, B.R.: “Utilizing Mini-Frac Data to Improve Design and Production”, Journal of Canadian Petroleum Technology, March 1994, pp. 26 – 35. MFrac III Version 3.5 (onwards) On-Line Help, Meyer and Associates Inc, December 1999 onwards. NSI Fracturing Technologies Web Site, www.nsitech.com Barree, R.D.: “A Practical Numerical Simulator for Three-Dimensional Fracture Propagation in Heterogeneous Media”, paper SPE 12273, presented at the SPE Reservoir Simulation Symposium, San Francisco CA, Nov 1983. StimLab division of CoreLab, Web Site, www.corelab.com/StimLab/Depts/GOHFER_ prodinfo.asp, 2002 onwards.

Page 94


BJ Services’ Frac Manual 12. Predicting Production Increase

12.

Predicting Production Increase

Being able to accurately predict a production increase from a formation is an important part of the process of designing a frac treatment. All treatments have to be economically justifiable, before approval by the operating company. In order to be able to produce an economic justification, the Engineer must have a reasonable idea of what the post fracture production increase will be. Moreover, this prediction must be reliable, as the Engineer will have a hard time justifying subsequent treatments, if previous justifications have proved to be inaccurate. In order to be able to produce an accurate prediction of the increase in production, the Engineer needs accurate pre-treatment production data. Items like permeability, skin factor, BHP and downhole producing rate are all critical. If accurate values for items such as these cannot be obtained, then the subsequent predicted production increase will also be inaccurate. Nevertheless, because of the uncertainties associated with most of the data used in the analyses below, any estimate of post fracture production remains just that – an estimate. The Frac Engineer must make this clear to any customer. As a result, it is often more reliable to base post-treatment production estimate on the results of offset wells, if any are available.

12.1

Steady State Production Increase

Steady state production is when all reservoir parameters remain unchanged during the production process. Items such as radial extent and reservoir pressure are fixed. Most of the time this does not exist, and the reservoir is at least in a pseudo-steady state (see below). Consequently, production increases based on steady state are an approximation only. However, they are often useful as a “first look”, “back-of-the-envelope” calculation, to quickly see if a fracture is viable or not. Darcy’s Equation (which is for steady state flow only) can be expressed as follows for a skin damaged reservoir:q

=

0.00708 k h ∆P ............................................................ (12.1) µ ln[re/(rw e-S)]

where q is the downhole producing rate in bpd, k is the effective reservoir permeability in md, h is the net height of the formation in ft, ∆P is the pressure differential between the edge of the reservoir and the wellbore (the drawdown) in psi, µ is the downhole viscosity of the reservoir fluid in cp, re is the radial extent of the reservoir, rw is the wellbore radius and S is the skin factor (dimensionless). Note that re and rw should always have the same units, usually either feet or inches. To provide a fair comparison between production at different times, which may be at varying drawdown, the productivity index, J, is usually used instead of the production rate. The units of productivity index (or PI) are usually bbls/day/psi, or bpd/psi. J

=

0.00708 k h

µ ln[re/(rw e-S)]

................................................................ (12.2)

To avoid confusion, the symbol J will be used to signify the PI from a real, damaged reservoir. Jo is used to represent the PI from an undamaged reservoir and Jf for the fractured reservoir. In Darcy’s Equation, the term kh is often referred to as the permeability-thickness, or conductivity. This equates to the fracture conductivity, Fc, of the propped fracture. By

Page 95


BJ Services’ Frac Manual 12. Predicting Production Increase replacing the term kh with Fc we can obtain an expression for the PI of the fractured reservoir:Jf

=

0.00708 Fc .................................................................... (12.3) µ ln(re/rw)

Equation 12.3 should be used with some caution. As explained earlier, this is a steady state approximation to a situation that in reality is far from steady state. The Equation no longer uses the skin factor term, as it is assumed that the fracture has completely bypassed the skin, rendering it irrelevant. This Equation also assumes that all production into the wellbore comes via the fracture. This is a valid assumption for fractures with a very high CfD, but becomes less and less accurate as the contrast between the fracture and reservoir conductivity becomes lower. Indeed, if the fracture conductivity is too low, this method may actually predict a production decrease – something that is theoretically impossible, unless the fluid or proppant somehow damages the formation. This Equation also assumes that the formation has no difficulty delivering reservoir fluids to the fracture – the Equation is independent of fracture length. Nevertheless, Equation 12.3, still provides a “first guess” to see how viable a fracture treatment is. However, it is less accurate for low permeability reservoirs and for fractures which relatively low fracture conductivity. The “folds of Increase” (Jf/J) can be calculated, by dividing Equation 12.3 by Equation 12.2, which gives the following:-s

Jf J

=

Fc ln[re/(rwe )] ......................................................... (12.4) kh ln(re/rw)

Another way of getting a “quick look” at potential post-treatment production is simply to use a skin factor of -5 in Equation 12.1.

12.2

Pseudo-Steady State Production Increase

Pseudo-steady state flow is when the reservoir has been producing for a sufficient period of time, so that the effects of reservoir boundary can be felt. In practical terms, this means that the reservoir has an outer boundary. Pr Radius of Disturbed Formation

Pressure

Increasing Time

Pwb 0

rw

Distance from Well

re

Figure 12.2a – Transient production. The red lines illustrate the variation of pressure with distance from the wellbore, as time increases. The radius of the disturbed formation is continually increasing

Page 96


BJ Services’ Frac Manual 12. Predicting Production Increase As the well is produced, the radius from the wellbore at which the reservoir has been disturbed by production increases at a rate proportional to the square root of the producing time. During this period, flow into the wellbore can be described as transient, as the effective radial extent of the reservoir is continually increasing. However, at some point the area of formation disturbed by the production from the well will hit an outer boundary. At this point, the radial extent of the reservoir ceases to expand, and the reservoir pressure starts to fall. At this point, the reservoir switches from transient to pseudo-steady state. The difference between transient and pseudo-steady state is illustrated in Figures 12.2a and b.

Pr

Pressure

Increasing Time

Pwb 0

rw

Distance from Well

re

Figure 12.2b – Pseudo-steady state production. The radius of the disturbed formation has reached the reservoir boundary, re, and now the reservoir pressure is decreasing

Most reservoirs will spend the majority of their producing lives in pseudo-steady state production.

McGuire and Sikora The best known method for predicting production increase during pseudo-steady state production was developed in 1960 by McGuire and Sikora. This work was based on earlier work carried out on electrical circuits by Dyes, Kemp and Caudle. Basically, they used a series of resistors and capacitors to represent the reservoir – resistors to represent permeability (the lower the resistance the higher the permeability), capacitors to represent the porosity or storage capacity of the reservoir, voltage to represent pressure and current to represent flow rate. These experimentally-derived curves, shown in Figure 12.2c, define for a given dimensionless fracture length (L/re) and a given fracture relative conductivity (see below – note that this definition is different from that used throughout the rest of this manual), the dimensionless production increase that can be expected. McGuire and Sikora used L for fracture half length, instead of the usual xf. Note that the following use a different system of nomenclature than the rest of this manual:Relative conductivity =

Wkf k

40 A ................................................... (12.5)

where W is the average propped fracture width in inches, kf is the permeability of the proppant in md, k is the formation permeability in md, and A is the well spacing in acres.

Page 97


BJ Services’ Frac Manual 12. Predicting Production Increase 14 L/re = 1.0 .9

12

7.13

r ln 0.472 r e w

10

J/Jo

.8

k = AVERAGE FORMATION PERMEABILITY, md. (BASED ON GROSS THICKNESS) L = FRACTURE LENGTH FROM WELL BORE, Ft. re= DRAINAGE RADIUS, FEET A = WELL SPACING, ACRES Wkf = CRACK CONDUCTIVITY, md-in. W = PROPPED WIDTH OF FRACTURE, in. kf = PERMEABILITY OF PROPPING MATERIAL, md. rw = WELL BORE RADIUS, FEET J = PRODUCTIVITY INDEX AFTER FRACTURING Jo= PRODUCTIVITY INDEX BEFORE FRACTURING

8

6

.7 .6 .5

.4

.3

.2

4 .1

2

0 1.E+02

1.E+03

1.E+04

RELATIVE CONDUCTIVITY,

1.E+05

Wkf k

1.E+06

40 A

Figure 12.2c – The McGuire-Sikora Curves

L Dimensionless fracture half length = r ........................................... (12.6) e

Where L is the fracture half-length (xf normally) in feet and re is the reservoir drainage radius or radial extent, also in feet. Dimensionless production increase =

J Jo

7.13 ln 0.472 (re/rw) .......... (12.7)

Where J is the pre-frac productivity index, Jo is the post-frac productivity index (Jf normally) and rw is the wellbore radius. McGuire and Sikora is an approximation based on the limits of the experimentation they conducted. The main assumption is that the fracture is significantly more conductive than the formation, so that the main rate limiting variable is the fracture half length. Vertical fluid flow is assumed to be negligible, fluids are assumed to be incompressible and in single-phase flow and skin factor is assumed to be zero. However, it is often relatively easy to find the production increase if the skin was reduced to zero. The McGuire-Sikora production increase can simply be added to this.

Skin Bypass Fracs Rae et al presented a simple method for predicting the production increase from a skin bypass frac. It combines elements of the McGuire-Sikora and Prats methods and allows for the existence of a skin factor:Jf J Page 98

-S

=

ln[re/(rw . e )] ln[4/(Fcd . xfD)] ................................................................ (12.8) Version 1.0 June 2005 Uncontrolled Copy – DO NOT DISTRIBUTE

BJ Services’ Frac Manual 12. Predicting Production Increase This method is valid for fracture with a CfD greater than 1 – i.e. more conductive than the formation.

12.3

Nodal Analysis

The most modern method for predicting production increase is the Nodal Analysis programme. These simulators work by analysing the flow from the reservoir at a node, which can be down hole at the “sand face”, at the wellhead or at some distance from the wellhead in a separator. By defining the flowing conditions at this node, the software can then calculate back to the flow rate from the reservoir. Nodal analysis can be used to produce inflow performance relationship (IPR) curves, which relate the ability of the reservoir to deliver fluids, with the ability of the completion to carry fluids out of the reservoir. These curves are particularly useful for oil wells with a GOR (i.e. real wells and not “black oil” approximations), gas wells and wells producing at significant water cuts, where the ability of the completion to carry the fluids is not always easy or straightforward to calculate. Figure 12.3a shows an example for a gas well with a fracture of varying average propped fracture width. 3000

2500

FBHP, psi

2000

1500

Tubing

1000

500

Average Propped Fracture Width, inches 0.1

0.2

5000

6000

0.3 0.4 0.5

0 0

1000

2000

3000

4000

7000

Gas Rate, mscfpd

Figure 12.3a – Nodal analysis IPR curves for a gas well with a fracture of varying propped fracture width.

With reference to the example in Figure 12.3a, note the following points:•

•

•

The blue curves represent five different production scenarios. In this case, each curve represents varying propped fracture width. However, they could just as easily be varying skin factor, permeability or water cut. This ability to test the sensitivity of the system to varying producing scenarios makes nodal analysis very powerful. The blue curves are the inflow curves. For these, the node is fixed at bottom hole (or the “sand face”). Each of these curves represents the inflow into the well from the formation for hydrocarbons at various FBHP’s (flowing bottom hole pressures). The drawdown is the difference between the reservoir pressure and the FBHP, so the smaller the FHBP, the greater the drawdown. The red curve is the outflow or tubing curve. This represents the ability of the completion to carry the hydrocarbons out from the well. In this case, the node is fixed at the wellhead. A set of wellhead conditions are specified, and then the software calculates (for a fixed

Page 99


BJ Services’ Frac Manual 12. Predicting Production Increase

•

FWHP – flowing wellhead pressure – and surface temperature) what the bottom hole pressure must be for a variety of different flow rates. The point at which the red curve and the blue curve cross represents the point at which the two sets of conditions coincide. Therefore, this is the rate and FBHP at which the well will produce. For instance, in Figure 12.3a, for a frac width of 0.2 inches, the well will flow at 4250 mscfpd at a FBHP of +/- 1450 psi.

Most nodal analysis programmes allow the user to produce the well through a propped hydraulic fracture of varying geometry. This is very useful to the Frac Engineer, who may well end up spending more time with the nodal analysis than with the fracture simulator. When using nodal analysis to predict production increase, the following steps should be followed:1. Get production data from the well. If the well is new, get production data from an offset. If no offsets are available, use the well test data. 2. History match the production data with the nodal analysis (and without a fracture being present). Vary items such as skin factor, permeability and reservoir pressure to make the nodal analysis production match the historical production data. The nodal analysis production simulator is now tuned to the real data. 3. Introduce a fracture. Vary characteristics such as fracture length and fracture conductivity (or average propped width) to produce the biggest possible increase in production. 4. Be aware of what is achievable and what is efficient. For instance, the nodal analysis may indicate that doubling the fracture length gives an extra 50% production. What it does not tell you is that doubling the fracture length means at least 4 times as much proppant, 8 times as much fluid and a corresponding increase in equipment. Such an increase in job size may not be practical and could well be uneconomic. 5. Once the optimum fracture geometry has been obtained, go to the fracture simulator and design a treatment to make a fracture of these dimensions. Often, it is at this point that the Engineer finds out what is realistically achievable and so the final design may be the product of several alternating runs on both the nodal analysis and the fracture simulator.

References Prats, M.: “Effect of Vertical Fractures on Reservoir Behaviour – Incompressible Fluid Case” Trans AIME (1961), 222 105-118 Dyes, A.B., Kemp, C.E. and Caudle, B.H.: “Effect of Fractures on Sweep-Out Pattern”, Trans AIME (1958), 213, 245 McGuire, W.J. and Sikora, V.J.: “The Effect of Vertical Fractures on Well Productivity”, Trans AIME (1960), 219, 401-403 Gidley , J.L., et al.: Recent Advances in Hydraulic Fracturing, Monograph Series Vol 12, SPE, Richardson, Texas (1989). Economides, M.J., and Nolte, K.G.: Reservoir Stimulation, Schlumberger Educational Services, 1987. Rae, P., Martin, A.N., and Sinanan, B.: “Skin Bypass Fracs: Proof that Size is Not Important”, paper SPE 54673, presented at the 1999 SPE Annual Technical Conference and Exhibition, Houston, Texas, Oct 3–6 1999. Archer, J.S. and Wall, C.G.: Petroleum Engineering Principals and Practices, Graham & Trotman, London, 1986. TM

Perform (Well PERFORMance Analysis ) Nodal Analysis Software, version 3.00 and higher, PSG/IHS Energy Group, Richardson, Texas, USA, 1999 onwards. Page 100


BJ Services’ Frac Manual 13. Candidate Selection

13.

Candidate Selection

Virtually any zone in any well is a potential candidate for hydraulic fracturing. Given a free hand, it is possible to produce an increase in productivity index in almost any formation using hydraulic fracturing. However, often the Frac Engineer is limited by considerations such as water-oil contacts, gas-oil contacts, poor cement bonding, completion restrictions and placement of perforations. Moreover, the formation must also have the reserves and production potential to economically justify the large expense often associated with fracturing. This section of the manual is designed as a guide to the science and art of frac candidate selection. Guidelines will be given, as to when an interval is a good candidate for fracturing and when it is not. However, there are often considerable “grey areas” between good candidates and poor candidates. In these cases, there is no substitute for experience. It should never be forgotten that the best wells are also the best candidates for fracturing. Fracturing cannot add reserves (although economically recoverable reserves and drainage efficiency can be improved) nor can it increase reservoir pressure – if there is nothing there to start with, there will be nothing there afterwards. A 50% increase in production from a good well is often more valuable than a 500% increase from a poor well.

13.1

Economic Justification for Fracturing

Fracturing – as with any other operation performed on an oil or gas well – has to be economically justified. That is to say, the increased revenue generated by the treatment must satisfy economic criteria set by the operating company. This is vitally important – it is not enough for the Frac Engineer to simply produce a production increase. Instead, the Frac Engineer must usually either produce at least a minimum production increase or increase economically recoverable reserves, in order to meet the economic criteria. Part of the skill in designing a fracture treatment is deciding whether or not these economic justifications can be met. An inability to meet these criteria is adequate grounds for rejecting a well as a candidate for fracturing. However, given that a treatment such as a skin bypass fracture can cost less than $20,000 to carry out, usually any reasonable criteria can be satisfied, unless the well has very low productivity indeed. Economic criteria can often be simple. For instance, many companies insist that the cost of the treatment be paid back within a period of three months. In such a case, the Frac Engineer has to estimate the increase in production and from that the total extra production over the first three months. Once the extra production has been calculated, the total extra revenue can easily be calculated by multiplying by the oil or gas price, as appropriate. If the total extra revenue is greater than the cost of the treatment, then the treatment is economically justified. All parties involved in the fracturing operation must be willing to accept a certain element of risk. Fracturing is not an exact science. Although many of the theories associated with fracturing are very rigorous and thoroughly proven, the fact remains that they are only as good as the available data. Often this data is of poor quality or is absent entirely. Even when considerable time, effort and expenditure have been taken to obtain data, it is usually only valid for a few inches around the wellbore. In order to complete a frac design, the Engineer has to assume this data is valid for sometimes hundreds of feet from the wellbore, encompassing a huge volume of rock. In addition to this lack of adequate data, the Frac Engineer also has to cope with the fact that no one really understands how the fracture propagates through the formation. This is illustrated by the fact that there are several different fracture simulators on the market, all using different methods to model the fracture. This uncertainty regarding how the fracture will propagate is in addition to the standard risks associated with any operation on an oil or gas well.

Page 101



Internal Rate of Return Many operating companies use a criterion known as the Internal Rate of Return. This is a percentage value, and any potential project requiring an AFE (Authorisation For Expenditure), must make a return on investment greater than this value. The theory is that the company would be better off spending the money elsewhere if a project cannot meet this criterion. For instance, if a Company Man wishes to spend $1,000,000 on a project, and his company has an internal rate of return criterion of 18% over one year, then the expenditure of $1,000,000 must generate additional production worth at least $1,180,000 in the first year after the treatment. The internal rate of return is also referred to as the discount factor, or DCF.

Net Present Value Net Present Value (or NPV) is a useful tool that can be used in two ways. First, the operating company can set an NPV criterion that has to be met. Secondly, it can be used to compare different fracture designs, and decide which one is the most cost effective. For instance, a Frac Engineer may be confronted with the following question – is it worth pumping twice the quantity of proppant for only a 10% gain in production? This question can be answered by using NPV analysis. NPV is calculated using the following method:Net Revenue

= Production Increase x Price ...................................... (13.1)

where the Production Increase is the total extra production due to the fracture treatment. n

Discounted Revenue

= X=1

Net Revenue for year X X .................................. (13.2) (1 + i)

where n is the payback period, usually measured in years, and i is the internal rate of return, expressed as a fraction. NPV =

Discounted Revenue – Total Treatment Costs ............. (13.3)

Remember that the Total Treatment Costs are the total cost that the customer has to pay, which includes the cost of the frac job (i.e. BJ’s ticket), plus items such as rig time, workover costs, wireline work, well testing, coil tubing etc. Example – NPV Calculation Calculate the NPV, given the following data:Oil price Payback period Internal rate of return Total treatment costs Production gain, yr 1 Production gain, yr 2 Production gain, yr 3

$20 per bbl 3 years 15% $1,250,000 400,000 bbls 200,000 bbls 100,000 bbls

Therefore Net Revenue, yr 1

= =

400,000 bbls x $20 per bbl $8,000,000

Net Revenue, yr 2

= =

200,000 bbls x $20 per bbl $4,000,000

Page 102


BJ Services’ Frac Manual 13. Candidate Selection Net Revenue, yr 3

= =

100,000 bbls x $20 per bbl $2,000,000

And so Discounted Revenue, yr 1

= =

Discounted Revenue, yr 2

= =

Discounted Revenue, yr 3

$8000000 1 (1 + 0.15) $6,956,522 $4000000 2 (1 + 0.15) $3,024,575

=

$2000000 3 (1 + 0.15) $1,315,032

= =

$6,956,522 + $3,024,575 + $1,315,032 $11,296,129

=

Therefore Total Discounted Revenue And finally NPV

= =

$11,296,129 - $1,250,000 $10,046,129

Now, let’s return to the Engineer’s original question - is it worth pumping twice the quantity of proppant for only a 10% gain in production? Let’s say that the cost of the actual fracturing was $500,000, and of that, the cost of the proppant was $200,000 and the cost of the fluid was $50,000. If we double the amount of proppant, we will probably need to at least double the amount of fluid. So the cost of the frac job goes up by $250,000. The final cost of the frac is now $750,000 and the overall cost of the treatment is now $1,500,000. A 10% increase in production gives us the following:-

So that

and which gives

Production gain, year 1 = Production gain, year 2 = Production gain, year 3 =

440,000 bbls 220,000 bbls 110,000 bbls

Discounted Revenue, yr 1 = Discounted Revenue, yr 2 = Discounted Revenue, yr 3 =

$7,652,174 $3,327,033 $1,446,535

Total Discounted Revenue =

$12,425,742

NPV

= =

$12,425,742 - $1,500,000 $10,925,742

So the answer to the Engineer’s question is yes – in this case, using a payback period of 3 years and an internal rate of return of 15%, it is worth doubling the volume of proppant. So the answer to the Engineer’s question is yes – in this case, using a payback period of 3 years and an internal rate of return of 15%, it is worth doubling the volume of proppant. Of course, two other things that the Frac Engineer must consider are;

Page 103


BJ Services’ Frac Manual 13. Candidate Selection 1. Can the customer afford the increase in the initial treatment cost? Small operators cannot always generate enough cash flow to do this. 2. Is it physically possible to place twice as much proppant in the fracture? Is this to be accomplished by increasing the fracture length, width or by some combination of the two? In more remote locations the Frac Engineer must also make sure that the equipment needed to store and blend the extra fluids and proppant is available.

13.2

Completion Limitations

Tubing Cooldown As relatively cold fracturing fluid is pumped down a completion, the tubing will start to cool down. As it cools down, it will shrink and decrease in length. On some wells, this can result in shrinkage of several feet Usually, wells are completed using packers with polished seal bores, and tubing with seal assemblies. When the completion is run, the packer is set at the required depth. Then the tubing is run, complete with a seal assembly on the bottom. The seal assembly is a length of pipe with a number of rubber seals on the outside. The idea is that these seals slide into the polished bore of the production packer, providing the required isolation. The seal assembly is usually several feet in length, so that it can slide up and down inside the polished bore, allowing the tubing to expand or contract whilst still retaining completion integrity. However, if the tubing is cooled down too much, the seal assembly can sting right out of the polished bore, and the completion will loose its integrity. This is highly undesirable. Additionally, as the tubing re-heats after the treatment, it probably will not sting back into the polished bore, and thus will produce additional stress on the packer, wellhead and other completion components. In order to prevent this from happening, special software programmes are used to simulate the effects of tubing shrinkage, to predict if the tubing will shrink too much. BJ’s programme for predicting this is DTools. There are two obvious answers to a tubing cool down problem:1.

Reduce the size of the treatment, so that the tubing does not get cooled down as much, or pump the treatment at a lower rate, so that the fluid heats up more as it travels down the well.

2.

Heat up the treating fluid before it goes down the well. This can be done in two ways. The first way is to pump the fluid through a heat exchanger, which contains a hot fluid, such as steam or burning oil. Such heat exchangers are often called “hot oilers”. The advantage of this system is that it can be used on the fly. The second way is to circulate the fluid through a choke, using the high-pressure frac pumps. A frac tank of fluid circulated through a choke can be quickly heated up – if the choke is set small enough so that the pumps can develop significant horsepower. 4000 HHP produces the approximately the same amount of energy as a 3MW power plant. The disadvantages of this method is that heating multiple frac tanks can be very time consuming, and individual tanks will cool down as other are heated up. Therefore, hot oilers tend to be used for large treatments, whilst pumping through a choke is used for smaller treatments.

Page 104



Tubing Expansion Another effect of fracturing on tubing is to cause its expansion, due to an elevated internal –or burst – pressure. This increase in diameter is usually not too much of a problem. However, as the tubing expands radially and circumferentially, it also contracts axially, reducing the length of the tubing string. Obviously, the effect of tubing expansion and the effect of tubing cool down will combine to produce an even worse effect. Once again, software must be used to model these effects. There are two ways to help mitigate the effects of tubing expansion:1.

Put pressure on the outside of the tubing, as it is an increase in the differential pressure across the tubing wall that causes the expansion. However, this is not always possible – especially on a completion with multiple packers. If it is possible, the maximum allowable pressure may be insufficient.

2.

Reduce the pumping rate. Obviously, the BHTP is pretty much fixed. However, by reducing the rate, and hence the friction pressure, the internal pressure that most of the tubing experiences can be reduced.

Maximum Wellhead Pressure Often, a treatment will be constrained by a low maximum wellhead pressure. It is very rare that a treatment is completely prevented by this, but a low wellhead rating can sometimes severely limit what can be achieved by the treatment. One solution to this problem is to use wellhead isolation tool or WIT (commonly referred to as a “Tree Saver”). This tool, which is described in detail in Section 20, actually bypasses the wellhead, by allowing the frac fluid to be pumped directly into the tubing, rather than through the wellhead and into the tubing. Another potential solution to this problem is to reduce the friction pressure. This can be done by either reducing the pumping rate or by altering the friction properties of the fluid (which can be done by either reducing the polymer loading or by delaying the crosslink). Both of these parameters are usually flexible to a certain extent. However, some wells have a frac gradient so high that even with zero friction pressure, the maximum wellhead pressure is exceeded. A third method for reducing the wellhead pressure is to pump a high density frac fluid. This has the effect of increasing the hydrostatic head, which in turn lowers the wellhead pressure. These fluids are usually mixed using high density brine. Potassium chloride brines can be used up to about 9.6 ppg, sodium chloride to about 10 ppg and calcium chloride to 11.0 ppg. Above that, things start to get expensive and considerably less environmentally friendly. Examples of materials used to weight brine above 11.0 ppg include caesium formate and zinc bromide. It should also be remembered that heavy-weight brines are harder to recover from the well after the treatment

Completion Jewelry Completion jewelry is a general term, used describe all the various special tools that were added to the completion as it was run. Examples include:• • • •

Sub surface safety valves (SSSV) Sliding side doors (SSD) Gas lift mandrels Blanks, used to close off gas lift mandrels

Page 105


BJ Services’ Frac Manual 13. Candidate Selection • •

Gauges and gauge carriers Non-return valves

All of these items will have a pressure rating. Ideally, this should be in excess of the overall pressure rating for the completion. However, this is not always the case, and such items should be checked. SSSV’s will form a restriction in the completion and may be abraded by the proppant. These always need to be locked open during a treatment, as the potential damage caused by an accidental closing is not worth the risk. Sometimes this can be performed from the surface, using the hydraulic control lines. In other instances, this has to be performed by installing an isolation sleeve by wireline. SSD’s can be both beneficial and detrimental to a fracture treatment. They can be beneficial, as they often allow the fracture treatment to be precisely injected into a specific zone. They can be detrimental, as they can get stuck both open and closed, and even when fully functional, usually require wireline intervention in order to manipulate them. Non-return valves should be avoided. Obviously, treatments cannot be pumped through a non-return valve. Treatments can be placed above a non-return valve, provided the nonreturn valve is isolated from the treatment by a bridge plug or similar tool.

Justifying a Workover Often, the only feasible way to fracture a formation is to carry out a workover. This allows the treatment to be pumped through a dedicated workstring, usually with some kind of packer. Consequently, the Frac Engineer has maximum control over the process – the treatment is placed in the right interval and the treatment can be pumped at relatively high rates and pressures. There are two ways to justify a workover; economically and technically. Generally, the first kind carries all the influence – there are very few companies that will approve a workover purely on technical grounds alone, unless there is some kind of research project going on. Workover operations can vary from the very cheap (such as a shallow land well) to the very expensive (offshore, deep water). Consequently, the grounds for justifying such a workover can also vary. In general, the best way to justify the cost of the workover is to first obtain an estimate for cost of the workover. Then, work out two different production increases. The first production increase assumes that a workover is performed and the Frac Engineer can place the optimum treatment. The second production increase uses the best stimulation method available, assuming no workover (this may not even be a frac – it could be an acid treatment). Then calculate rate of return and net present value for both of these methods. If the frac + workover generates a better return on investment, then it is economically justified. Another way to get a workover performed is to frac a well that is already in need of a workover. Then the costs of the workover can be split between the frac and the existing completion programme. Alternatively, a well can be selected for treatment that is in need of a workover which cannot be economically justified. The combined effects of the frac and the existing need may be enough to justify the additional expense.

13.3

Things to Look For

Listed below are a number of items that may make an interval a good, or bad, candidate for hydraulic fracturing. 1.

Skin Factor. All wells have skin damage, to a greater or lesser extent, unless they have been stimulated in some fashion. This means that all unstimulated wells are producing significantly below their full potential. As a general principle, the higher the

Page 106


BJ Services’ Frac Manual 13. Candidate Selection permeability, the higher the skin factor – so that the most productive wells are also the ones which produce least efficiently. All this means that in practice, all wells are potential candidates for fracturing. Figure 13.3a shows the effect skin factor, S, has on the production of a well. The horizontal axis shows the well’s original Skin Factor. The vertical axis shows the effect this has on productivity, relative to the undamaged production (S = 0), so that production from the undamaged well equals 100%. Note that this graph does not include the stimulation effects of the frac – it merely illustrates how much production is lost due to skin factors. A hydraulic fracture will punch a highly conductive path through the skin damage, producing a production increase by two methods; through bypassing the skin damage and through stimulation of the undamaged reservoir. Therefore, an interval with a high skin factor is a good candidate for fracturing.

100

Production Relative to S = 0, %

80

60 re = 2000 ft rw = 4.25 inches 40

20

0

0.0

5.0

10.0

15.0

20.0

25.0

30.0

Skin Factor, S Figure 13.3a – The effect of skin factor upon production rate. Note that this Figure is based purely on skin factor effects. No fracture stimulation is included.

2.

Low Permeability Wells. So-called “tight” formations are where fracturing first became widely accepted by the industry. These formations cannot produce enough hydrocarbons purely because the rock matrix itself is not conductive enough. Any production loss due to the (usually) low skin factor is generally not significant. Therefore, in order to unlock the potential of the reservoir, a fairly large hydraulic fracture treatment is required.

3.

Weak or Unconsolidated Formations. Hydraulic fracturing is a very effective method for completing a weak or unconsolidated formation. Fracturing can help reduce or eliminate sand production by a number of methods:• • •

By reducing the drawdown on the formation By re-stressing the formation By acting as a filter, provided the proppant is sized correctly.

A hydraulic fracture can also be used as part of a gravel pack completion, providing a so-called frac and pack treatment. This is probably the most effective way of developing an unconsolidated formation. Page 107


BJ Services’ Frac Manual 13. Candidate Selection 4.

Water and/or Gas Contacts. In general, these are to be avoided. The presence of a water or gas contact close to the perforations can often prevent fracturing. If a propped fracture were to propagate into a water or gas zone, then the well will quickly stop producing oil, and start producing water or gas. Once a propped fracture has connected with a water or gas zone, it is very difficult to halt the water or gas production.

5.

Poor Cement Bond. If the bond between the casing and cement, or cement and formation, is poor or non-existent, then fracturing should be avoided. In these situations, it is possible to make the poor bond even worse and to connect with separate formations above and below the zone of interest. However, in the case of a “micro-annulus”, the pressures induced by the fracturing, coupled with the filter-cake building properties of the fluid, will usually permit successful fracturing operations.

6.

Corroded Casing or Tubulars. Badly corroded casing or tubulars will probably not stand up to the differential pressures produced by fracturing. Therefore these wells should be avoided.

7.

Perforation Strategy. The position of the perforations can often prove to be the difference between a successful and an unsuccessful frac. Section 14 discusses this in more detail.

8.

Logistics. This is a measure of how easy it is to get materials and equipment to location. For instance, there is a big difference between a land location a few miles down the road from the base, and an offshore location on a satellite platform with a 5 tonne crane limitation. These two locations may have wells and formation that require similar treatments. However, it is very unlikely that the offshore would be treated in the same manner to the land well, unless a stimulation vessel was available. More often than not, it is the logistics of the operations – rather than any formation parameters – that has the biggest influence on the treatment.

References Howard, G.C., and Fast, C.R.: Hydraulic Fracturing, Monograph Series Vol 2, SPE, Dallas, Texas (1970). Gidley , J.L., et al.: Recent Advances in Hydraulic Fracturing, Monograph Series Vol 12, SPE, Richardson, Texas (1989). Archer, J.S. and Wall, C.G.: Petroleum Engineering – Principles and Practices, Graham and Trotman, London (1986). Economides, M.J., and Nolte, K.G.: Reservoir Stimulation, Schlumberger Educational Services, 1987.

Page 108


BJ Services’ Frac Manual 14. Perforating for Fracturing

14.

Perforating for Fracturing

Of all the things under our control, the position, number, size and phasing of the perforations has the single biggest influence on the effectiveness of the hydraulic fracture treatment. Many times this is outside of the control of the Frac Engineer, as a high proportion of treatments are carried out on existing wells that have already been perforated. However, if a well or an interval is new, the Frac Engineer can often greatly increase the effectiveness of a treatment by perforating for fracturing, rather than in a more conventional manner. When perforating for fracturing, it is often desirable to only perforate a very limited section of wellbore, usually located towards the centre of the gross interval. This controls the point of fracture initiation and helps to reduce tortuosity. However, there are quite legitimate reasons for wanting to perforate all of the net pay (which can often result in several sets of perforations). One of these reasons is well testing, which is used by reservoir engineers to help determine the recoverable reserves in the formation - obviously a very important task. Results from well test analysis can be misleading if the entire interval is not perforated, especially if the formation contains several discrete intervals. Therefore, the need to reduce the number of perforations and to reduce the length of the perforated interval, must be balanced with the operating company' s other interests. A compromise must be reached.

14.1

Controlling Fracture Initiation

Perforations can be used to control the point of fracture initiation, as illustrated in Figure 14.1a, below. On the left-hand side, there is an interval that has been perforated across its entire section. When the treatment commences, fracture initiation takes place. At this point, it should be remembered that fractures are initiated by pressure, not by rate. As Frac Engineers, we often use rate to create pressure (as a consequence of Darcy’s law), but it’s the pressure that makes the fracture. As the pressure increases, a fracture will initiate when the pressure rises above the breakdown pressure of the weakest point along the perforated interval. This can be in at the top of the zone (frac A, below), in the middle of the zone (B), at the bottom of the zone (C) or somewhere else. There can also be more than one fracture – wherever the fluid pressure exceeds the breakdown pressure, a fracture will be initiated. Multiple fractures (see Section 10.5) can result in poor fracture conductivity and early screenouts.

A

B

D

C

Figure 14.1a – The Effect of perforations on fracture initiation

Page 109


BJ Services’ Frac Manual 14. Perforating for Fracturing If the interval is perforated as shown in the left-hand side of Figure 14.1a, the point at which the fracture or fractures initiate is beyond the control of the Frac Engineer. Fracs A and C have substantial sections propagated outside the interval. This results in poor coverage of the interval and a considerable amount of wasted proppant. There is also a risk that Frac A could penetrate into a gas cap or that Frac C could penetrate into a water zone. Alternatively, the interval could be perforated as shown in the right-hand side of Figure 14.1a. In this example, the zone has been perforated over a very short interval (5 to 10 ft). This controls the point at which the fracture initiates, and dramatically reduces the chances of multiple fractures forming. If this short perforated interval is in the center of the zone, then there is a good chance that the fracture will propagate both up and down, covering the entire section and using the proppant efficiently. Alternatively, if there is a water zone close by, the interval can be perforated towards the top. This causes the fracture to initiate near the top, reducing the chances of the fracture penetrating down into the water. Of course, once the interval has been fractured, there is nothing - other than cost – to stop a second perforation run being made to cover the rest of the interval. However, if the treatment has been effective, the fracture will be many times more conductive than the formation. Consequently, any perforation that is not directly connected to the fracture will be unproductiv:We have recognized point-source perforating improves your ability to successfully stimulate an interval........to improve our completions and ultimate recoveries. We have learned from perforating for stimulation that it does not take 100 ft of perforations to produce a 100 ft zone. We have proven that 5ft placed in the proper place will outperform all 100 ft. Robert Lestz, Production Engineer, Chevron Hart’s E&P, February 2000 Another example of perforating to control fracture initiation is the case when multiple zones are treated simultaneously in a single treatment. The conventional method is to try a limited entry treatment (see Section 10.10), but these are unreliable and difficult to control.

Figure 14.1b – Perforating for zonal coverage

Page 110


BJ Services’ Frac Manual 14. Perforating for Fracturing Figure 14.1b illustrates this concept. Conventionally, each productive section of the formation is perforated individually. When this well is fractured, a portion of the fluid (dependent upon a number of variables) will enter each of the intervals, as in the left-hand side of Figure 14.1b. Limited entry fracturing is all about controlling how much fluid goes into each interval and can be very unreliable. However, if the well has not already been perforated, another method is to perforate a small section in the center of the formation, and allow the fracture to connect up all of the individual intervals (right-hand side of Figure 14.1b). Under any circumstances, a treatment that produces a single fracture is much easier to predict and control than a treatment that produces multiple fractures. Once again, a small section (5 to 10 ft) of perforations is shot. These need to be placed roughly in the center of the interval to be covered, or slightly towards the bottom, depending upon the stress regime. Consequently, this may even mean deliberately perforating a nonproductive formation, such as a shale. It can often be quite hard to convince an oil or gas company to deliberately do this. Important Note:- Sometimes, if there is a significant contrast in Young’s modulus between the various formations, sections were the fracture is significantly narrower than the average can form. These narrow bands can act to prevent proppant transport, leaving formations above and below un-propped. The reader should not use the above method unless reliable information on Young’s modulus contrasts – such as from a sonic array log – is available. Perforating to control fracture initiation also makes fracture simulation and post-treatment pressure matching more reliable. By controlling the point of fracture initiation, the Frac Engineer defines a significant simulation variable and reduces the complexity of any possible solution by an order of magnitude.

14.2

Controlling Tortuosity

In order to minimise tortuosity, it must be as easy as possible for the fracture to propagate from the perforations. Every single perforation is a potential source of fracture initiation, so one of the steps taken is to reduce the number of perforations to an absolute minimum, consistent with the anticipated production rate. This in turn means big holes.

Figure 14.2a – Perforation strategy for vertical wells

Another important factor is the phasing of the perforations. Ideally, this should be 180°, with the guns oriented so that they shoot perpendicular to the maximum horizontal stress. This way the holes are lined up with the direction of fracture propagation, minimising any changes of direction between the hole in the casing and the main fracture. Most of the time it is not possible to orientate the guns in this fashion – the best strategy will then depend upon factors such as the contrast between the maximum and minimum horizontal stresses and the formation’s Young’s modulus. The situation is complex, but in general it is best to minimise the number of holes shot, to use big holes to minimise perforation friction, and to perforate to Page 111


BJ Services’ Frac Manual 14. Perforating for Fracturing that the holes line up along the wellbore (see Figure 14.2a), rather than produce a spiral of holes around the circumference. The best strategy for perforating for fracturing was presented by Behrman in 1998. However, it is the author’s experience that 90º phasing usually produces the least near wellbore friction in vertical or near-vertical wellbores, without getting involved in very complex strategies.

Deviated Wellbores Hydraulic fractures tend to propagate on a vertical or near vertical plane (see Section 7). On a vertical well, this means that the fracture will propagate along or close to the wellbore. This minimises the formation of multiple fractures, as the compression of the rock on either side of the fracture will make it harder for parallel fractures to grow. However, on a deviated or horizontal wellbore, the horizontal distance between potential points of fracture initiation is much greater, making it much easier to produce tortuosity and/or multiple fractures. Consequently, it is common practice for highly deviated or horizontal wells, to perforate a very short section of the formation (+/- 2 ft or less), with as many big holes as possible. This is shown in Figure 14.2b (for a horizontal well):-

Figure 14.2b – Perforation strategy for horizontal wells

14.3

Perforating for Skin Bypass Fracturing

3

2

1

Figure 14.3a – The Effect of fracture initiation point on skin bypass fracs

Skin Bypass Fracturing (SBF - see Section 3.4) is a special type of small-scale fracturing operation designed to penetrate through skin damage, and to provide effective stimulation Page 112


BJ Services’ Frac Manual 14. Perforating for Fracturing without the cost and logistics of larger-scale treatments. Whilst it is true that SBF’s may not necessarily offer such a large production increase as conventional fracturing, the stimulation is still effective, and is usually more than adequate to justify the cost of the treatment. As with any type of fracturing, the position of the perforations can have a significant effect on the results of the treatment.

A

B

C

Figure 14.3b – Multiple skin bypass fracs over a long interval

With reference to Figure 14.3a, it is easy to see how the point of fracture initiation can effect a fracture not designed to cover the entire height of the formation, such as in skin bypass fracturing. Obviously, fracture B will produce more stimulation than fractures A or C. If the entire section of the formation is perforated, it is usually not possible to control the point of fracture initiation (although a sand fill can be used to ensure that the fracture doesn’t initiate towards the bottom). Therefore, when planning a perforation strategy, it would be better to shoot holes over a small, central section, than over the entire interval. Figure 14.3b shows a different approach to perforating for SBF’s. Over a large section, one of the most cost effective methods of stimulation is to carry out several small consecutive treatments, as listed below (with reference to Figure 14.3b): Zone Lower

Middle

Upper All

Step 1 2 3 4 1 2 3 4 1 2 3 1 2

Action Perforate bottom zone Frac lower zone Recover fluids Isolate lower zone by placing bridge plug Perforate middle zone Frac Middle zone Recover fluids Isolate middle zone by placing bridge plug Perforate upper zone Frac upper zone Recover fluids Remove bridge plugs Place on production

This method ensures maximum coverage of the interval for minimum of effort, although it does involve three separate perforating runs and the use of coiled tubing to remove the bridge plugs or sand fill.

Page 113


BJ Services’ Frac Manual 14. Perforating for Fracturing

References Behrmann, L.A.: “Perforating Requirements for Fracture Stimulations”, paper SPE 39453, presented at the SPE International Symposium on Formation Damage Control, Lafayette LA, Feb 1998. Rae, P., Martin, A.N., and Sinanan, B.: “Skin Bypass Fracs: Proof that Size is Not Important”, paper SPE 56473, presented at the SPE Annual Technical Conference and Exhibition, San Antonio TX, Oct 1999. Behrmann, L.A., and Nolte, K.G.: “Perforating Requirements for Fracture Stimulations”, paper SPE 59480, SPE Drilling and Completions, December 1999, pp 228 – 234. Venkitaraman, A., Behrmann, L.A., and Chow, C.V.: “Perforating Requirements for Sand Control”, paper SPE 65187, presented at the SPE European Petroleum Conference, Paris, Oct 2000.

Page 114


BJ Services’ Frac Manual 15. The Step Rate Test

15.

The Step Rate Test

Step rate tests are usually performed before a hydraulic fracture treatment, as part of the fracture design process. Together with the minifrac (see next section), they are often referred to as calibration tests, as they are used to adjust the fracture model to the actual pressure response of the formation. There are two types of step rate test, the step up test and the step down test. One is used for determining fracture extension pressure, whilst the other is used for determining near wellbore friction. Both tests can be extremely useful when designing the treatment. Whenever possible, bottom hole pressure data should be used, as this is more accurate and reliable than calculated BHTP.

15.1

The Step Up Test

The step up test is used to determine the fracture extension pressure, Pext. This is usually 100 to 300 psi higher than the fracture closure pressure, Pclosure, which is a very important factor in fracture design. Usually the results of the step up test will be used to determine an upper boundary for Pclosure and to give the expected BHTP. To carry out the step rate test, it is common practice to use either KCl water or linear gel. However, if this test is to be combined with the minifrac (see Section 16), then the actual frac fluid should be used. The test itself consists of pumping fluid into the formation at various rates. These rates start off slowly and gradually increase. For example, these could be the pump rates for a typical test; 0.25 bpm, 0.5, 0.75, 1.0, 1.5, 2.0, 3.0, 5.0 and 10 bpm. The first step is usually the lowest rate that the pumps can manage. It is important to get as many stages at low rate as possible. At each stage, first achieve the rate, then wait for the pressure to stabilize and finally record the exact pressure and rate. Then move on to the next stage.

Pressure

Pext

Rate Figure 15.1a – The step up test

What is important with this test is to get stabilized pressure. It is not that important to get the exact rates. Often, pump operators will fiddle with the rate for 30 seconds or so in order to get Page 115


BJ Services’ Frac Manual 15. The Step Rate Test exactly 0.75 bpm. This is not necessary. Get approximately the correct rate and then leave it alone, so that the pressure can stabilise and be recorded. Once the test has been carried out, a plot of pressure against rate can be made, as illustrated in Figure 15.1a. The change in gradient of the slope shown in Figure 15.1a marks the change from Darcy radial flow (lower rates) to Darcy linear flow at higher rates. This is the point at which our fracture is created and hence this is our extension pressure. When carrying out a step up test it is important that no artificially induced fracture exists prior to the test. Thus, if any pumping above the frac gradient has already been carried out, sufficient time should be taken for the fracture to heal up before commencing the step rate test. On very tight rocks, this could be several days. The step rate test can also provide an indication of fracture toughness, at least in the formation close to the wellbore. In theory, the difference between the extension pressure and the closure pressure (usually obtained from the minifrac) is directly related to the fracture toughness. However, it is also heavily influenced by wellbore orientation, perforation strategy and the orientation and magnitude of the horizontal stresses.

15.2

The Step Down Test

This test is used to determine the nature of any near wellbore friction that may exist, i.e. to see if it is perforation or tortuosity dominated. As the name suggests, the step down test is the opposite of the step up test. Instead of starting at low rates and increasing, the rates are started high and decreased.

Pressure

Tortuosity Dominated

Perforation Dominated

Rate Figure 15.2a – The step down test

When performing the step down test, it is important that the fracture is open the whole time, otherwise the test is invalid. Therefore, this test is often carried out after a step up test. It is not uncommon to step up then step down right after. Another factor to remember when conducting a step down test is keeping the stages short as the rate is stepped down. Unlike the step up test, which starts with no fracture and ends with an open fracture, the step down test must be performed with the fracture open all the way through. Consequently, if the steps down take too long, the fracture will close before the end of the test, making the low rate data points invalid. 4 or 5 steps down, taking 10 to 15 seconds each, is all that is required. Also, make sure that before starting the step down, that the fracture has been open for at least 5 minutes - the longer the better, as smaller fractures will close more quickly than larger fractures. Page 116


BJ Services’ Frac Manual 15. The Step Rate Test Figure 15.2a shows the relationships between pressure and rate for the step down test. The different shapes of the curves indicate how the near wellbore friction is dominated by the perforations, by the tortuosity or by a combination of the two. For perforation friction:Pnwb ∝

2

Q .................................................................................. (15.1)

In theory, perforation friction follows the same theory as flow through orifices, involving Bernoulli’s Equation and stagnation pressure. Allowances have to be made for the diameter of the perforation (assumed to be constant) and for the discharge coefficient (a measure of how “smooth” the flow is as it goes through the orifice). The discharge coefficient is also assumed to be constant. As a result, the pressure loss is proportional to the rate per perforation, as illustrated in Equation 2.3. Generally, at this stage no significant volumes of proppant have been pumped and so the assumption that perforation diameter is constant is valid. For tortuosity:Pnwb ∝

Q .................................................................................. (15.2)

In theory, for tortuosity dominated near wellbore friction, as the pumping rate increases, so does the width of the near wellbore flow channels, as the width of these is dependent upon pressure – the higher the rate, the higher the pressure and hence the greater the width. This is why, for tortuosity, Pnwb does not increase as fast as rate. In reality, the relationship between rate and near wellbore friction may be a lot more complex than that suggested by Equation 15.2. Recent work by the GRI suggests that Pnwb may be 0.25 proportional to Q rather than the square root of rate. On top of this, it is likely that the relationship between Pnwb and Q is also controlled by the nature of the tortuosity, so that different relationships exist for tortuosity generated by perforations, for tortuosity generated by horizontal stress contrasts, or for tortuosity generated by wellbore deviation (to name but three potential causes). To further complicate the situation, it is entirely possible that a well could experience tortuosity that is a combination of two or more causes. However, in spite of this complex relationship between pressure loss and rate, the geometry of the tortuosity will always be pressure-dependent and hence under most circumstances the pressure-rate crossplot will have the characteristic convex shape for tortuosity-dominated near wellbore friction. Of course, usually the near wellbore friction is a combination of perforation friction and tortuosity. Although Meyer’s MinFrac minifrac analysis programme is not recommended by the author, as it is based on a rather simplistic 2-D analysis, the step rate test analysis section within MinFrac is excellent, especially for the step down test. It incorporates a feature that allows the theoretical perforation friction to be backed out, allowing the user to view the total tortuosity-based friction, regardless of the exact relationship between pressure loss and rate. In addition, both MFrac and Fracpro (both versions) allow data from a step down test to be input directly into the simulator, so that the model can allow for the effects of tortuosity related pressure losses when calculating net pressure. However, given that most step rate tests are performed using a different fluid to the actual treatment (slick water rather than crosslinked gel), it must be remembered that the actual pressure loss will probably be greater than data generated by the step rate test indicates.

15.3

Step Rate Test Example – Step Up/Step Down Test

The following data was taken from a step rate test in which the rate was stepped up and then immediately stepped down again, using slick water. The data generated by the step rate test is given in table 15.3a. Page 117


BJ Services’ Frac Manual 15. The Step Rate Test Figure 15.3a shows the step up pressure-rate crossplot. Figure 15.3b shows the step down crossplot, whilst Figure 15.3c shows the same step down crossplot using surface pressure. This illustrates why bottom hole pressure must always be used for step rate test analysis, even if it has to be calculated from surface data. Rate bpm

STP psi

BHTP psi

0.5 0.9 1.0 1.2 1.6 2.0 2.3 3.2 4.2 5.2 6.3 8.4 10.2 11.8 8.4 6.3 4.2 2.0

2030 2310 2445 2600 2730 2850 2910 3120 3450 3780 4224 5290 6280 7281 5180 4160 3271 2580

5958 6211 6337 6474 6559 6623 6636 6671 6753 6783 6838 6996 7041 7076 6886 6774 6574 6353

Table 15.3a – Example step rate test data.

7200 7000

BHTP, psi

6800 6600 Fracture Extension = +/- 6570 psi 6400 6200 6000 5800 0

2

4

6

8

10

Slurry Rate, bpm Figure 15.3a – Step up pressure-rate crossplot using the example data. This plot shows the fracture extension pressure to be at +/- 6570 psi.

Page 118


12


7200

BHTP, psi

7000

6800

Tortuosity Dominated

6600

6400

6200 0

2

4

6

8

10

12

Slurry Rate, bpm Figure 15.3b – Step down pressure-rate crossplot for the example data. The convex shape of the curve indicates near wellbore friction dominated by tortuosity. 8000

7000

STP, psi

6000

5000 Perforation Dominated? 4000

3000

2000 0

2

4

6

8

10

12

Slurry Rate, bpm Figure 15.3c – Step down pressure-rate crossplot for the example data, using surface treating pressure (STP). This graph illustrates the danger of using STP for step rate test analysis, as in this case, the near wellbore friction would have been incorrectly diagnosed as being perforation dominated.

Page 119



References Lacy, L.L. and Hudson, H.G.: ”Step Rate Test Analysis for Fracture Evaluation”, SPE 29591, presented at the SPE Rocky Mountain Region/Low Permeability Reservoirs Symposium, Denver, Colorado, March 1995. Gidley , J.L., et al.: Recent Advances in Hydraulic Fracturing, Monograph Series Vol 12, SPE, Richardson, Texas (1989). Economides, M.J., and Nolte, K.G.: Reservoir Stimulation, Schlumberger Educational Services, 1987. Cleary, M.P.:, Johnson, D.E., Kogsbøll, H-H., Owens, K.A.: Perry, K.F., de Pater, C.J., Stachel, A., Schmidt, H., and Tambini, M.:” Field Implementation of Proppant Slugs to Avoid Premature Screen-Out of Hydraulic Fractures with Adequate Proppant Concentration”, paper SPE 25892, presented at the SPE Ricky Mountain Regional/Low Permeability Reservoirs Symposium, Denver CO., April 1993. Cleary, M.P., Doyle, R.S., Meehan, D.N., Massaras, L.V. and Wright, T.B.: “Major New Developments in Hydraulic Fracturing with Documented Reductions in Job Costs and th Increases in Normalized Production”, SPE 28565, presented at the SPE 69 Annual Technical Conference and Exhibition, New Orleans, Louisiana, September 1994. Wright, C.A., Weijers, L., and Minner, W.A.: Advanced Stimulation Technology Deployment Program, report GRI-09/0075, Gas Research Institute, March 1996.

Page 120


BJ Services’ Frac Manual 16. The Minifrac

16.

The Minifrac

The purpose of the minifrac is to provide the best possible information on the formation, prior to pumping the actual treatment. Any time that the quality of information available for a frac candidate is poor, a minifrac should be planned. This includes most wells, as it is not usual to have detailed rock mechanical and leakoff data for a formation (and for the non-productive formations surrounding the zone of interest). The only time a minifrac should not be pumped is when there is reliable data available from offset wells that have been fractured (as is often the case in the US). The minifrac is designed to be as close as possible to the actual treatment, without pumping any significant volumes of proppant. The minifrac should be pumped using the anticipated treatment fluid, at the anticipated rate. It should also be of sufficient volume to contact all the formations that the estimated main treatment design is anticipated to contact. A well planned and executed minifrac can provide data on fracture geometry, rock mechanical properties and fluid leakoff – information that is vital to the success of the main treatment.

16.1

Planning and Execution

Bottom Hole Data Do whatever it takes to get bottom hole pressure data for the minifrac (and also for the step rate test), as this is far more accurate than data calculated from the surface pressure. Bottom hole data can be obtained using three acquisition methods:1.

2.

3.

Real Time Gauges. These can be run on wireline or can be part of the well’s completion. These gauges allow both pressure and temperature to be recorded real time at the surface. Usually, it is possible to run a data cable so that the pressure data can be incorporated real time with the standard frac data already being recorded. This is the best possible situation for the Frac Engineer. Memory Gauges. These are gauges that are run in on wireline or slick line, and hung in either a specially designed gauge carrier, or some other suitable position (such as an empty gas lift mandrel). Alternatively, they can just be held on slick line at a specific depth. After the mini-frac and the step rate test are completed, the gauges are retrieved and the data is downloaded at the surface. This data is then merged with the surface data that has already been collected. This is the most common method of using gauges, even though there is a delay caused by the retrieval of the gauges. Dead String/Live Annulus. Both of these methods work on the same principle. With the live annulus, the well is completed with tubing but no packer (or the packer has not been set, or the packer is fitted with a circulating valve that is left open during the treatment). Basically, the annulus is exposed to the BHTP during the treatment, and shows a corresponding surface pressure. As the fluid is not moving in the annulus, BHTP can be easily calculated, provided the density of the fluid in the annulus is known. Most fracture monitoring programmes have the capability to perform this real time. A dead string relies on the same principle, but instead employs a small diameter tubing string, inside the actual treatment tubing. A common example of this is coiled tubing placed inside a large diameter frac string.

Remember that it is more important to get downhole pressure data during the minifrac and the step rate test, than it is during the actual treatment. Companies that supply gauges are often reluctant to have proppant pumped past them (this also applies to wireline cables). Consequently, it is common to have the gauges in the well for the minifrac and the step rate test, and then retrieve them prior to pumping any proppant.

Page 121


BJ Services’ Frac Manual 16. The Minifrac Most bottom hole pressure gauges also record temperature. This data, whilst not as important as pressure data, can also be useful:1. 2. 3.

The data can provide a good check of the bottom hole static temperature, to ensure that the correct temperature has been used for designing the fluid system. The data can provide a good value for the bottom hole treating temperature. This is especially important when performing treatments with nitrogen and/or carbon dioxide and also for treatments where tubing shrinkage due to cooldown is critical. If the gauges have been run on wireline or slick line, then it is possible to run the gauges past the perforations after the minifrac and the step rate test, to perform a temperature log. This is a plot of temperature against depth. By looking at how far each the perforations have cooled down – and how this cooling down varies across the perforated interval – it is possible to get a qualitative indication of where the fluids are going and hence were the fracture(s) is(are) initiating.

Because the rheology of the fracturing fluid is constantly changing as the minifrac is being pumped, and because the well is continually cooling down, calculated friction pressures can often be unreliable. This in turn means that a BHTP calculated from a STP can also be unreliable. This is why it is important to obtain reliable downhole data, from which to base the frac design.

Fluid Volumes and Rates Deciding what volume to pump for the minifrac can be difficult. Ideally, we wish to pump the minimum volume necessary to gather accurate formation and fracture data. However, remember that we are not just interested in getting data on the producing formation – we are also after data on any formation above and below that may be contacted by the fracture. This means that as a minimum, we must pump at least the two-thirds of the anticipated pad volume. On low permeability wells we may have to pump significantly more than the pad volume. The best method to decide the minifrac volume, is to run a few simulations for the minifrac, based on the data used to design the main treatment. Adjust the minifrac volume such that it will contact all the formations that main treatment will contact. As we are trying to create a treatment that is a close as possible to the actual treatment (minus the proppant), the minifrac should be pumped at the same rate as the anticipated treatment. The minifrac should be displaced with slick water. The displacement volume should be enough to displace the minifrac to just short of the perforations, to ensure that the near wellbore fracture(s) close on frac fluid, rather than slick water. To do this, it is common to under-displace by +/- 5 bbls.

Fluid Type As stated above, we are trying to create a test that is as close as possible to the actual treatment, minus the proppant. This means that the minifrac should use the same fluid as the anticipated treatment. In fact, every step should be taken to ensure that the fluids used in the minifrac and the main treatment are as identical as possible, so that fluid related data gathered in the minifrac is as valid as possible for the main treatment. Often, an operating company will suggest using slick water for the minifrac, as a way of saving time and money. This is a false economy, as the subsequent minifrac will have only a passing resemblance to the fracture that will be created by the main treatment. In particular, the fluid leakoff will be (usually) significantly greater with slick water. This results in faster than normal fracture closure, and smaller than normal fracture geometry. Page 122


BJ Services’ Frac Manual 16. The Minifrac Recently, some Engineers have argued that because of the wall building effects of the fluid used in the minifrac, the leakoff for the main treatment can be lower than that for the minifrac. To compensate for this, increased breaker loadings are used in the minifrac.

Wellbore Fluid Usually, there is some kind of fluid in the wellbore prior to the minifrac. Often, this fluid will be slick water from the step rate test, or produced fluids. Unless this fluid can be circulated out of the well ahead of the minifrac fluid, it will be injected into the formation as part of the minifrac. Obviously, having two different fluid types in the fracture makes the job of analysing the minifrac data that much more difficult, so every effort should be taken to minimise the volume of fluid ahead of the minifrac fluid. On some wells, this can be achieved by circulating the minifrac fluid into position. However, on most wells this cannot be done, and the Frac Engineer has to live with the situation.

Pressure Decline The data collected during the pressure decline (i.e. after the minifrac has been displaced and the pumps are shut down) is just as vital as the data collected whilst pumping. It is therefore important to monitor the pressure decline, sometimes for up to 2 hours after the minifrac is completed. During this period, it is important that nothing is done to compromise the quality of this data. Any opening or closing of valves, hammering on equipment or circulating of fluids should be avoided at all costs. In particular (and this may sound obvious, but it does happen) the wellhead should not be closed during this period. There should also be no pumping into the annulus, as this will affect the tubing pressure. Once the Frac Engineer is sure that the fracture has closed, the well can be shut in and normal activities resumed.

Proppant Slugs Many Engineers prefer to pump a proppant slug in the minifrac. This is a proppant stage in the middle of the minifrac, often containing as little as 500 lbs of proppant. This slug will test the near wellbore region for tortuosity. Ideally, the proppant slug should pass into the formation with no detectable pressure rise. If the pressure rises when the proppant flows into the formation (and worse still, if it rises and does not come down again), then there is restricted flow in the near wellbore region – tortuosity. A series of proppant slugs of increasing concentration can be use to effectively diagnose the severity of a tortuosity problem. See Section 10.1 for more details on tortuosity.

Multiple Minifracs Some companies, especially those operating in high permeability formations, prefer to use more than one minifrac. The first minifrac is designed to be small, to penetrate only into the zone of interest and provide good leakoff and closure data on this formation. The second minifrac is larger, designed to penetrate further and give a better idea of the overall fracture geometry. Obviously, the use of two minifracs provides more data than just a single treatment. However, in most cases this is probably not necessary. A well-designed and executed treatment should be able to provide the Frac Engineer with all the necessary data. However, there are cases when it is very difficult to interpret the minifrac data, through no fault of the treatment. Some formations are just too complex to easily analyse. In such cases, when the data from the first minifrac defies scrutiny, often the only way to proceed is to pump a second minifrac, in the hope that this data will be better.

Page 123



16.2

Anatomy of a Minifrac

Figure 16.2a shows a typical job plot from a minifrac:-

Pressure, Rate

BHTP

STP Rate

Time Figure 16.2a – Typical minifrac job plot, showing BHTP, STP and rate

Three important parameters are used – to a greater or lesser extent – in obtaining the required data from the minifrac. The BHTP (ideally actual pressure data, rather than calculated) is the main variable, as this tells us the way the fracture is behaving and the amount of work being performed on the formation by the fluid (or visa versa). The rate is important for determining the fracture geometry, as the volume of fluid pumped into the formation, less the volume of fluid which has leaked off, is equal to the volume of the fracture. In addition to these two parameters, the proppant concentration can also be important, if proppant slugs have been pumped.

BHP

Shut Down

Pressure Decline

Start Pumping

Time Figure 16.2b – Expanded plot showing BHTP

Figure 16.2b shows an expanded portion of Figure 16.2a, giving the BHTP more detail. Generally, a large number of minifracs will have this same basic shape, although by no means all. The area between the start of pumping and the shut down is often shaped like this, with the pressure declining initially and then increasing towards the end. In terms of Nolte analysis (see Section 10.2), this means that the fracture is initially growing in a shape which is radial or preferentially vertical (rather than horizontal), but after a period of time the height growth becomes more controlled, and the preferential growth direction is horizontal.

Page 124



16.3

Decline Curve Analysis

As soon as the pumps shut down, the pressure will start to decline. Initially, the net pressure will still be positive, and the fracture may still propagate. However, as soon as the fluid input into the fracture stops, the fracture will start to decrease in volume, as fluid is still leaking into the formation. As the fluid volume in the fracture (and hence the volume of the fracture itself) decreases, the fracture width also decreases until the fluid volume in the fracture is zero – the fracture has closed. The time taken for the fracture to close defines the rate at which the leakoff is occurring, whilst the pressure at which the fracture closes (and the difference between the treating pressure and the closure pressure) tells us how hard it will be to produce the required fracture. Both of these parameters have been more rigorously defined in previous sections of this manual, but suffice to say that they are both extremely important parameters for defining the size and shape of the fracture. A typical pressure decline curve is shown in Figure 16.3a. BHTP

BHP

ISIP

Closure Pressure

Linear Flow

Radial Flow

Time Figure 16.3a – Typical minifrac pressure decline curve

It is possible to see several distinct features on this curve, although it must be emphasized that Figure 16.3a is idealised and that actual minifrac pressure decline curves are rarely this clear. Features which the Frac Engineer needs to identify include:1. 2.

BHTP – the actual bottom hole treating pressure. This is the pressure inside the well, at the middle of the perforated section that is being treated. Ideally, this should be measured via a gauge or a dead string. ISIP – the instantaneous shut-in pressure, also referred to as the instantaneous shut down pressure, or ISDP. This is the bottom hole treating pressure just after the pumps shut down, and before the pressure the pressure starts declining. Often, this point is hidden by noise generated by “pipe ring” as the pressure suddenly drops. In that case, the decline curve has to be extrapolated backwards in order to find the ISIP.

The difference between the ISIP and the BHTP is due purely to friction pressure loses in the near wellbore area. Therefore, this difference can often be used as a quantitative assessment of tortuosity. 3.

Closure Pressure, Pclosure, is the pressure at which the fracture closes, and is often denoted by a change in gradient on the pressure decline curve. The difference between the ISIP and the closure pressure is referred to at the net pressure, or Pnet (see Sections 2.2 and 10.2). As discussed previously, the net pressure is a measure of how much energy is being used to create the fracture and so is a very important

Page 125



4.

parameter. However, it should be remembered that the net pressure will usually vary throughout the treatment, and that this method only captures the net pressure right at the end of the treatment. The closure pressure is also a measure of the in-situ stresses in the formation (see Section 6, Rock Mechanics). Closure Time. The closure time is the time taken for the fracture to close, after the pumps have shut down. If the geometry of the fracture is known (or, more likely, can be estimated from a model), then the volume of fluid in the fracture is also known. Therefore, if the length of time taken for the fracture to close is also known, the rate at which the fluid is leaking off can be easily calculated.

The are various different methods for helping the Frac Engineer pick closure pressure, as often it is very difficult to spot the change in gradient on the pressure decline curve. Additionally, there may be more than one closure pressure, if multiple fractures are closing. Finally, the effects of tortuosity may mask the closure pressure, as there is evidence to suggest that the tortuosity can, in some cases, close before the main part of the fracture.

Various Methods for Displaying Time In order to help find the closure pressure(s) on the pressure decline curve, various methods have been developed for plotting the data. Some of these will be described in more detail below. As part of these methods, and for general information, there are various methods of plotting time along the horizontal axis, listed in Table 16.3a, below:Description Time, general Data Time Pump Time Shut in Time Delta Time Square Root Time Horner Time

Symbol t tdata tp ts ∆t

Nolte Time or Dimensionless Time Delta Nolte Time Nolte G Time or G Function

tD ∆tD G or G(∆tD)

Equal to Usually time since start of pumping Time since data collection started Length of time spent pumping Time since ISIP t - tp 0.5 t tp + ts log10 ts t tp t - tp t = t -1 tp p Dimensionless function of ∆tD (see below)

Table 16.3a – Table illustrating the various ways of calculating and using time during pressure decline analysis.

Square Root Time Plots According to Equation 2.9, the volume of fluid leaked off into the formation (and hence the fracture volume) is proportional to the square root of time that the fracture has been open. However, once the fracture is closed, the fluid is no longer leaking off from the fracture faces, and is now leaking off according to Darcy’s radial flow law:q

Page 126

=

k h ∆P

µ ln (re/rw)

...................................................................... (16.1)


BJ Services’ Frac Manual 16. The Minifrac where q is the leakoff rate, k is the permeability of the formation, h is the net height of the formation, µ is the fluid viscosity, re is the radial extent of the formation, rw is the wellbore radius and ∆P is the pressure differential between the formation and the wellbore. Therefore, if a plot is made, showing BHP as the y-axis and square root time as the x-axis, the period before fracture closes should have the pressure decline as a straight line. The point at which the fracture closes is defined as the point at which the straight line starts to curve, as illustrated in Figure 16.3b:BHTP

BHP

ISIP

Closure Pressure

Linear Flow

Radial Flow

Square Root Time Figure 16.3b – Use of a square root time plot to determine closure pressure.

Square root time plots are both the easiest to use, and the easiest to understand, of all the pressure decline curve plots. However, their usefulness is limited by the ease with which multiple fractures and tortuosity can mask and obstruct the point at which the flow regime changes. The method is also dependent upon the reliability of Equation 2.9, which itself is an approximation, assuming that leakoff is independent of pressure. However, because of its ease of use, the square root time plot is usually the first stop in an often rather involved process.

Horner Plots Horner plots are taken directly from well test theory, and can very useful in helping to determine closure pressure. However, these plots must always be used in conjunction with other methods, as the Horner plot will only determine the lowest possible pressure at which closure could have existed. In other words, it will give a lower boundary, above which the closure must be found. Remember that the step rate test (step up variety – see Section 15.1) will give an upper boundary, so that using these two methods in conjunction will provide a region within which the closure pressure lies. Horner plots work by plotting BHP on the y-axis and the Horner time on the x-axis. Horner time is defined as follows:tHorner =

log10

tp + ts ................................................................. (16.2) ts

According to Horner’s theory, on a plot of pressure against Horner time, pseudo-radial flow (which, for our purposes, means flow when the fracture is closed) produces a straight line on the plot, and non-pseudo-radial flow (i.e. the fracture is open) produces a curve. A typical minifrac pressure decline Horner plot is shown in Figure 16.3c:-

Page 127



BHP

Closure Pressure

Pres 0

Horner Time Figure 16.3c – Typical minifrac pressure decline Horner plot

As the minifrac pressure decline progresses, the BHP will eventually reach the reservoir pressure, Pres. So, for Equation 16.2, as ts tends to infinity, the right hand side of the Equation tends to zero. This means that if the pressure decline is extrapolated back to the point where tHorner equals zero, the average reservoir pressure can be determined (also referred to as P*).

Nolte G Time Analysis Below is a summary of Nolte’s work on G Time and minifrac analysis. A full derivation of the method is beyond the scope of this manual, and the reader is referred to the references. Nolte derived the following relationships for the decline curve:3/2

g(∆tD)=

3/2

4/3(1 + ∆tD) - ∆tD -1 -1/2 1/2 ................................. (16.3) (1 +∆tD)sin (1 + ∆tD) + ∆tD

where the upper part of the RHS represents the upper boundary and the lower part of the RHS is the lower boundary. In practice, to find the actual value of g(∆tD), both values are calculated, and an extrapolation is made based on the power law exponent of the fracturing fluid (n’) and the fracture geometry (radial, PKN or KZD). Remember that when calculating from the lower boundary, the trigonometrical function works in radians, not degrees. The extrapolation is performed between two values of the variable α. At the lower boundary α = 0.5 and at the upper boundary α = 1. The actual value for α is given as follows:α α α

= = =

(2n’ + 2)/(2n’ + 3) (n’ + 1)/(n’ + 2) (4n’ + 4)/(3n’ + 6)

- PKN ......................................... (16.4) - GDK ......................................... (16.5) - radial ........................................ (16.6)

The actual value of α used for the extrapolation is dependent upon the fluid efficiency and n’. Values tend to be almost always in the region of 0.5 to 0.7, and in practice 0.6 is often used. Also, given the fact that n’ is often variable, a quicker method is just to take the average of the upper and lower expressions for g(∆tD). As shown in Figure 16.3d, as ∆tD increases, the difference between the upper and lower boundaries becomes smaller and eventually becomes negligible compared to the accuracy of the rest of the system:-

Page 128



5

4

3

g (∆ t D)

-1

-1/2

Lower (α = 0.5, g (∆t D) = (1 + ∆t D)sin (1 + ∆t D)

2

1

1/2

+ ∆t D )

g (∆t D = 0) = π/2

Upper (α = 1, g (∆t D) = 4/3[(1 + ∆t D)

g (∆t D = 0) = 4/3

0 0.01

0.1

3/2

3/2

- ∆t D ])

1

10

Dimensionless Time, ∆t D

Figure 16.3d – Graph showing the variation of g(∆ ∆tD) with ∆tD.

Nolte G time is then a function of tD such that:G(∆tD) =

g(∆tD) – g(∆tD = 0) .......................................................... (16.7)

Note that for α = 1 and α = 0.5, g(∆tD = 0) is equal to 4/3 and π/2 respectively. A typical plot of a pressure decline against Nolte G time is shown below in Figure 16.4e.

Additional Fracture Extension

Closure Pressure

BHP

“Ideal” ISIP

0

G(∆ ∆tD)

Figure 16.3e – Typical Nolte G time pressure decline plot. The match pressure is the gradient of the straight line section in the middle of the decline, before closure.

Figure 16.3e illustrates three important points. First, the ISIP recorded using field data may be artificially high, due to the effects of fracture storage and fluid friction. Second, that there is a period of constant gradient before the fracture closes, which is often referred to as the match Page 129


BJ Services’ Frac Manual 16. The Minifrac pressure (Pm) and has pressure units (as G time is dimensionless). This is an important parameter in Nolte’s minifrac pressure decline analysis. Finally, closure occurs when the decline pressure deviates from this constant gradient. At this point G(∆tD) = Gc. It should be noted that if the closure time equals the pump time, then Gc = 1. From the g(∆tD) time at closure [ = g(∆tcD)], the fluid efficiency can be determined as follows:-

η

=

g(∆tcD) - g(∆tD = 0) g(∆tcD)

1 - vprop ................................ (16.8) 1 - vprop/η

where vprop is the fraction of the total fracture volume occupied by proppant. For a minifrac, vprop will be equal to zero. Therefore:-

η

=

This can be simplified to:-

η

≈

g(∆tcD) - g(∆tD = 0) ..................................................... (16.9) g(∆tcD) Gc 2 + Gc .......................................................................... (16.10)

which is a quick and easy method for determining fluid efficiency. Most modern real time data monitoring systems can plot G-Function real time, so if the closure pressure can be determined, the fluid efficiency can be easily calculated from Equation (16.10). The fluid loss coefficient can be calculated as follows:Ceff =

Pmβs X ..................................................................... (16.11) rp tp E'

where Pm is the match pressure (see Figure 16.3e), βs a geometry-dependent factor (see below), rp is the ratio of fracture area in permeable formation over total fracture area (i.e. net to gross area ratio for the fracture), E’ is the plane strain Young’s modulus (see below) and X is a factor dependent upon which geometry model is being used, such that for KZD, X = 2xf, 2 for PKN, X = hf and for radial, X = (32R/ 3π ).

βs

≈

(2n’+2)/(2n’+3+a) 0.9 2 (3π /32)

PKN KZD ........................................ (16.12) Radial

where n’ is the power law exponent for the fluid and a is a variable describing how constant the viscosity of the frac fluid is in the fracture, such that for a constant viscosity, a = 1 and for a falling viscosity a < 1. Usually, a is assumed to be 1. Finally, the plane strain Young’s modulus can be easily calculated:E’

=

E 2 ........................................................................... (16.13) 1-ν

Thus, not only is Nolte G time a useful tool for finding the “ideal” ISIP and the closure pressure, it can also be used to find fluid efficiency and fluid leakoff (provided a 2-dimensional fracture geometry is assumed). Finally, Nolte G time can be used to find the fracture dimensions:Af

=

(1 - η)Vi 2 g(∆tD = 0)Ceffrp

tp

................................................... (16.14)

Where Af is the area of one fracture wing and Vi is the total volume of fluid injected. Page 130


BJ Services’ Frac Manual 16. The Minifrac Given that for the 2-D models:Af

=

2xfhf 2xfhf 2 πR

PKN KZD ............................................ (16.15) Radial

then the fracture length or fracture radius can be easily found. Average fracture width can also be obtained:w ¯

=

2 g(∆tD = 0)Ceffrp (1 - η)

tpη

.................................................. (16.16)

Derivative Plots When carrying out pressure decline analysis, a lot of time is spent trying to find various changes in gradient on the curve, or points were the pressure decline changes from a straight line to a curve (or visa versa). Therefore, it is often easier to spot these changes in gradient by actually plotting the gradient - or derivative – itself. On a derivative plot, a horizontal line (i.e. constant gradient) indicates a straight line on the parent plot (not necessarily horizontal, however). Changes in gradient on the parent plot, produce rapid changes in value on the derivative plot. An example is shown below in Figure 16.3f. All of the main types of plots - and their derivatives - can be plotted by most modern fracture simulators with the minimum of effort. Often, these plots can be displayed real-time by the data acquisition systems. Consequently, there is always a temptation to stop recording data too early – the Frac Engineer notices a change in gradient and assumes the fracture is closed. This is not necessarily the case, and so it is important to keep recording data for as long as feasible. It takes relatively little effort to record the data for an extra 10 minutes and a lot of embarrassment can be avoided.

BHP

Derivative

0

d(BHP) dtHorner

Closure Pressure

tHorner Figure 16.3f – Example derivative plot based on a Horner Plot

16.4

Pressure Matching

Another method for analysing minifrac data, is to carry out a process known as a pressure match (also referred to as a history match). In this process, the fracture simulator is “tuned” Page 131


BJ Services’ Frac Manual 16. The Minifrac until the simulator’s predicted net pressure matches the actual net pressure. This “tuning” process is carried out by adjusting variables such as Young’s modulus, Poisson’s ratio, fracture toughness, stress gradient, near wellbore friction, leakoff rate, and spurt loss. Pressure matching, which is discussed in more detail in Section 19, is a very powerful tool, providing the user is aware of the limitations. The user is actually adjusting the computer model to produce the same pressure response as the formation. Once the model has been adjusted (the pressures have been “matched”), any potential treatment schedule can be run on the simulator, and its effects assessed. This means that once the match has been made, the Frac Engineer can very quickly adjust the treatment schedule to produce a fracture of the required geometry.

Limitations of Pressure Matching 1.

2.

3.

Complexity. Because so many variables are adjusted, in so many different rock strata, a Frac Engineer may often have to keep track of 20 or more variables. Each of these variables can affect the overall outcome of the simulation. Therefore, a Frac Engineer must remain aware of what variable changes and values are realistic and what are not. Non-Unique Solution. Because there are so many variables to adjust, it is quite possible for 2 Frac Engineers to produce good pressure matches, using different values. Often, these solutions will only produce similar net pressure responses for the particular data set being analysed, so that when a different treatment schedule is simulated (such as the actual treatment schedule to be pumped), two significantly different fractures are generated. Which one is closest to the truth? Data Quality and Model Inaccuracies. As with any type of computer analysis, the results are only as good as the raw data (garbage in = garbage out). In particular, errors generated by the use of surface pressure data to calculate BHTP can sometimes render pressure matching almost ineffective. For instance, it is quite possible to interpret a gradual rise in STP as good fracture containment, whereas in reality it may have been caused by variations in fluid properties. Even if the quality of data is good, the final result is only as good as the model itself. Just because a model predicts a fracture that is 150 ft long and 100 ft high, doesn’t mean that this is what happens in the ground. In fact, two different fracture simulators will almost always produce different fractures, when fed the same input data. Again, which one is closest to the truth?

The study of the theory of how the fracture models work will only get a Frac Engineer so far in trying to solve these conundrums, especially as the companies responsible for the most widely used fracture models do not publish significant parts of their theory. Unfortunately, in this case there is no substitute for experience.

16.5

Near Wellbore Effects and Multiple Fractures

Most of the time, minifrac analysis is not simple. Often, it is not possible to find closure pressure, or obtain a pressure match. More often that not, this is due to the effects of tortuosity and/or multiple fractures. Both of these concepts were explained in Section 10. However, it is worth discussing the particular effects that these phenomena can have on minifrac analysis.

Tortuosity As previously discussed in Section 10, tortuosity consists of a number a small, restricted, flow channels in the near wellbore area, connecting the perforations to the fracture(s). Generally, this phenomenon is detected by the pressure drop it produces whilst the frac fluid is being Page 132


BJ Services’ Frac Manual 16. The Minifrac pumped. Obviously, it is much easier to detect and quantify tortuosity from bottom hole pressure data, as pipe friction effects can easily mask its effect if surface pressure data is used. Tortuosity can confuse the results of a minifrac analysis in two ways:1.

The pressure drop produced by tortuosity represents a loss of energy from the fracturing fluid. This means that the frac fluid does not have as much energy as the pressure data indicates. The pressure inside the fracture remains the same, irrespective of whether or not tortuosity exists, as the fluid flow rate out of the near wellbore area is the same as the flow rate into it. As the inlet rate and fluid properties going into the fracture are unchanged, so the fracture dimensions remain unchanged and hence the net pressure remains unchanged. Tortuosity does not produce a lower than normal pressure in the fracture – it produces a higher than normal pressure in the wellbore. However, the effect of this is to lead the Frac Engineer into believing that the pressure in the fracture is higher than it actually is. Consequently, the Frac Engineer is led to believe that the fracture is significantly bigger than it really is, and can be tempted to plan a treatment with larger volumes of proppant than can actually be pumped into the fracture.

2.

Tortuosity can also cloud the interpretation of the minifrac pressure decline. The channels which form the tortuosity are always significantly narrower than the main fracture (otherwise they wouldn’t produce a pressure drop), and so can often close entirely before the main fracture(s) itself closes. This means that the main fracture is no longer hydraulically connected to the wellbore and so the actual closure pressure can be very difficult to spot. In addition, the pressure at which the tortuosity closes can itself cause a change in gradient on the pressure decline plot, causing a false value to be selected for closure pressure, at a higher pressure.

The only way to allow for these effects is to be fully aware of the existence of tortuosity, and to have some idea of its magnitude. The main ways of obtaining this information is to use bottom hole pressure data, and to pump a step down test (see Section 15).

Multiple Fractures The existence and causes of multiple fractures have already been discussed in some detail in Section 10. Sufficient to say that under the right circumstances multiple fractures are not only possible, they are likely. Of course, the classic way to identify multiple fractures is to see two or more closure pressures on a minifrac pressure decline curve. However, in reality this very rarely happens. In order for multiple closures to be apparent on a decline curve, there must be significant differences in the actual closure pressures of each individual fracture, otherwise they will merge into one closure on the plot. Usually, the multiple fractures all exist in the same formation(s) and so will close at approximately the same pressure. The main problems for minifrac analysis associated with multiple fractures are as follows:1.

Although the multiple fractures will close at approximately the same pressure, they will almost never close at exactly the same pressure. Variations in depth and bisected formations will cause a variation that could be as much as 20 psi or more. This means that when the fractures close, instead of a nice, easy-to-spot, change in gradient on the pressure decline curve, there is a significant region where the gradient gradually changes between the open fracture environment and the Darcy flow wellbore leakoff environment. This “smudging” of the closure pressure can make it very hard to identify.

2.

As discussed above, multiple fractures will usually close at around the same pressure, allowing for the effects of variations in depth. However, they probably will

Page 133


BJ Services’ Frac Manual 16. The Minifrac not have similar fracture geometries, and so some fractures will be much larger than others. As stated previously, the leakoff rate is proportional to the fracture area, whilst the time taken for the fracture to close depends upon the fracture volume. For most fractures, the area of the fracture faces is proportional to the square of the length, whilst the volume of the fracture is proportional to the cube of its length. So a fracture which is twice the length of another fracture will leakoff at four times the rate, but will have eight times as much volume to lose before closure, so that the fracture takes twice as long to close. Therefore the bigger fractures tend to take longer to close than smaller fractures. However, all of our fractures are connected hydraulically via the wellbore. We know that our fractures will tend to close at the same time, because they will all have similar closure pressures. Therefore, in order to prevent the smaller fractures closing significantly before the larger fractures, there must be fluid flow from the larger fractures to the smaller fractures, at a rate equal to the difference in leakoff rates. This means that the smaller fractures have an artificially long closure time and the large fractures have an artificially short closure time. In a situation where there are several fractures, the flow dynamics can get very complex indeed. This flow of fluids from one fracture to another, as well as the pulling in of extra fluid from the wellbore, can produce complex shapes on the pressure decline curve. This can make analysis very difficult.

16.6

Minifrac Example 1 - 2D Minifrac Analysis

The following minifrac treatment was pumped into an oil-bearing formation, located in South Kalimantan, in the Indonesian part of island of Borneo. Bottom hole memory gauge data was available. This example will demonstrate the use of the Nolte G Function analysis technique to obtain the leakoff coefficient, closure pressure, and fracture geometry. Well and Formation Data Reservoir Type: Reservoir Temperature: Reservoir Pressure: Perforations: Deviation at Perforations: Liner: Treating String: Packer set at: End of Tubing: Top of Formation: Bottom of Formation: Permeability: Porosity: Young’s modulus: Poisson’s ratio:

Oil 145 F unknown 986 m (3235 ft) to 1032 m (3386 ft) Vertical 7”, 23# 3.5”, 9.3# tbg 940 m (3084 ft) 950 m (3117 ft) 986 m (3235 ft) 1032 m (3386 ft) 6 mD 18% 500,000 psi (assumed) 0.25 (assumed)

Treatment Data Wellbore Fluid: Treatment Fluid: Treatment Volume: Displacement Fluid: Displacement Volume: Treatment Rate:

Page 134

Slick water (from step rate test) Crosslinked gel (SpectraFrac G 4500, n’ at BH = 0.65) 3 50 m (314 bbls) Slick water 3 5.3 m (33.3 bbls) 3 3 m /min (18.8 bpm)



4,000

40

Gauge BHTP

30

STP

2,000

20 Slurry Rate

1,000

Slurry Rate, bpm

Pressure, psi

3,000

10

0

0 0

10

20

30

40

50

60

Elapsed Time, mins

Figure 16.6a – Minifrac example 1 job plot.

We can see from Figure 16.6a that the treatment was well executed, with the rate staying constant. The pressure was monitored for a significant length of time, probably longer than necessary. However, it is better to record too much data than too little. Figure 16.6a actually shows merged bottom hole gauge and surface data. This plot would not have been available whilst the treatment was being performed. Figure 16.6b shows the gauge BHTP pressure decline in more detail, whilst Figure 16.6c shows the pressure against the square root of elapsed time.

3,300

Gauge BHTP, psi

3,100

2,900

2,700

2,500

2,300 10

20

30

40

50

60

Elapsed Time

Figure 16.6b – BH gauge pressure decline against elapsed time. Possible closure pressure at +/2770 psi (where the two red lines cross, marking a change in gradient). Note the sudden drop of about 50 psi as the pumps shut down at t = +/- 13 mins.

Page 135



3,300

Gauge BHTP, psi

3,100

2,900

2,700

2,500

2,300 3.0

4.0

5.0

6.0

7.0

8.0

Square Root Time, mins1/2

Figure 16.6c – BH gauge pressure decline against the square root of elapsed time. Possible closure pressure at +/- 2790 psi (where the two red lines cross, marking a change from straight line to curve).

These two plots are basically in agreement – closure pressure at about 2780 psi. On both plots we see a sudden drop of about 50 psi, as soon as the pumps shut down. This is almost certainly due to near wellbore friction. This drop in pressure makes the true ISIP (the treating pressure inside the fracture) difficult to spot exactly. However, 50 psi is quite low and is unlikely to cause any problems as far as pumping the treatment is concerned (see later example number 3 for a case where near wellbore friction did effect the treatment). Figure 16.6d shows the G Function plot, which should enable a “true” ISIP to be determined, by extrapolating the straight line back to the y-axis:3,200

3,100

Gauge BHTP, psi

3,000

2,900

2,800

2,700

2,600 0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

2.0

G Function

Figure 16.6d – G function plot. The “true” ISIP is at +/- 3150 psi, whilst the closure pressure appears to be at +/- 2780 psi (where the two red lines cross). This gives a Gc of 1.30.

Page 136



3,500

Gauge BHTP

3,200

2,900

2,600

2,300

2,000 0.0

0.5

1.0

1.5

2.0

2.5

3.0

Horner Time

Figure 16.6e – Horner plot. The results from this plot are ambiguous and do not help in the analysis.

Figure 16.6e shows the Horner plot for the minifrac pressure decline. As can be seen, there is no clear change in gradient from linear flow to pseudo-radial – several different points could be picked. Therefore this plot is not much help in the analysis. This is a common phenomenon in minifrac analysis – one plot being ambiguous, whilst others show clearer results. This is one reason why the Frac Engineer must be familiar with the various types of plots that exist. Most fracture monitoring and analysis software packages allow the user to easily display several different types of decline curve. Results of Graphical Analysis There is a high degree of agreement between the pressure decline plot, the square root time plot, and the G function plot. In fact, any experienced Frac Engineer reading this example may find this data suspicious – minifracs are rarely this easy to interpret. However, this is real data – later on we shall see an example from a similar formation that is much less easy to analyse. To summarise the results of the graphical analysis:ISIP Closure pressure Closure time Pump Time G function at closure, Gc

3150 psi 2780 psi 31 mins (elapsed time) 17.5 mins (shut in time) 13.5 mins 1.30

Nolte G Function Analysis Assumptions Radial geometry, initial fracture radius estimate 50 ft, initial rp estimate 1.0 (i.e. the fracture is completely contained in the production formation). 1.

From Equation 16.10, we can find the fracture efficiency at pump shut down:-

Page 137



η 2.

1/2

0.00686 ft/min

=

2

20,147.56 ft

From Equation 16.16 we can get the average width of the fracture:w ¯

5.

=

From Equation 16.14, we can find the area of the fracture:Af

4.

39.9%

From Equation 16.11, we can find the leakoff coefficient:Ceff

3.

=

=

0.604 inches

From Equation 16.15 we can obtain a revised value for the fracture radius:R

=

80.1 ft

Obviously, this final result is significantly different from the initial fracture radius estimate of 50 ft. They both cannot be right, and are in fact both wrong. In order to find the final answer, an iterative process must be performed, bringing the initial and final values of the fracture radius closer and closer together until the difference is negligible. To start the first iterative step (in this example) steps 1 to 5 are re-worked using the average of the initial and final values for R, 65 ft. Remember that our formation height is 118 ft – and our fracture height is now 130 ft (2R). In this case of radial geometry, once the fracture height exceeds the formation height, the ratio of net to gross area (rp) must be less than 1. With a fracture radius of 6 5ft, rp can be calculated (using relatively simple geometry) as 0.967. 6.

Using a new initial R of 65ft and a rp of 0.967, we get the following result:-

η

Ceff Af w ¯ R

= = = = =

39.4 % 1/2 0.00922 ft/min 2 15,498.13 ft 0.785 inches 70.2 ft

(unchanged, as this depends upon Gc only)

The iterative process continues until the difference between the initial and final values for R are negligible. This gives the final minifrac analysis result:

η

Ceff Af w ¯ R

= = = = =

39.4 % 1/2 0.00988 ft/min 2 14,728 ft 0.826 inches 68.4 ft

These values can now be plugged into the 2-D fracture simulator as the basis for a simulated treatment with proppant. Note that in order to obtain this result, both Young’s modulus and Poisson’s ratio have to be assumed. In addition, it was also assumed that the formations above and below the zone of interest had the same rock mechanical properties as the main zone. Finally, it was assumed that each fluid that entered the formation had the same leakoff properties. These are the limitations of using a 2-D model.

Page 138



16.7

Minifrac Example 2 – 3D Pressure Matching with FracProPT

The next example was pumped in an offshore well in Vietnamese waters. The well itself was an exploration well that was extensively tested before and after the treatment – and then abandoned. The operating company wished to determine if hydraulic fracturing was a viable field development technique. Well and Formation Data Reservoir Type: Reservoir Temperature: Reservoir Pressure: Perforations: Deviation at Perforations: Casing: Treating String: Packer set at: Ported XOver Sub: Top of Formation: Bottom of Formation: Permeability: Porosity:

Gas Condensate 249 F unknown 3122 m (10,243 ft) to 3137 m (10,293 ft) Vertical 9-5/8”, 47# 3-1/2”, DST string 3094 m (10,151 ft) 3098 m (10,164 ft) 3116 m (10,222 ft) 3143 m (10,312 ft) na na

Treatment Data – Step Rate Test Wellbore Fluid: Treatment Fluid: Treatment Volume: Treatment Rate:

Seawater Slick Water (20 ppt GW-27) 3 2.4 m (15 bbls) 3 0.08 to 2.4 m /min (0.5 to 15 bpm)

Treatment Data - Minifrac Wellbore Fluid: Treatment Fluid: Treatment Volume: Displacement Fluid: Displacement Volume: Treatment Rate:

Slick water (from step rate test) Crosslinked gel (SpectraFrac G 4500) 3 18.9 m (119 bbls) Slick water 3 5.3 m (33.3 bbls) 3 2.4 m /min (15 bpm)

Step Rate Test Figure 16.7a shows the job plot for the step rate test. We can see from this Figure that this was not a particularly well executed step rate test – the rate, and hence the pressure, never really stabilises for each of the steps. Nevertheless, Figure 16.7b does show quite a marked change in gradient, indicating that the fracture extension pressure is around 8700 psi, which gives a frac gradient of 0.85 psi/ft – quite high, but not unheard of. Unfortunately, there is no step down portion for this step rate test. This is recommended in any situation where tortuosity is suspected. As this was a formation that had never been fractured before, it would have been prudent to perform the step down test.

Page 139



10,000

20

Gauge BHTP

6,000

16

12

Surface Pressure

4,000

Rate, bpm

Pressure, psi

8,000

8

2,000

4

Slurry Rate

0 0.0

2.0

4.0

6.0

8.0

10.0

12.0

0 16.0

14.0

Elapsed Time, mins

Figure 16.7a – Minifrac example 2 step rate test job plot.

However it appears that in this case there are no indications of tortuosity, as the step rate test pressure decline shows no immediate drop in bottom hole pressure as the pumps are shut down. Also note that the bottom hole pressure is taken from memory gauges mounted in the DST string. Therefore, the frac engineer on site did not have access to this data. 10,000

9,000

Gauge BHTP, psi

8,000

7,000

6,000

5,000

4,000 0.0

2.0

4.0

6.0

8.0

10.0

12.0

14.0

16.0

Slurry Rate, bpm

Figure 16.7b – Step rate test crossplot for minifrac example 2, step rate test, showing fracture extension at +/- 8700 psi.

Page 140



Minifrac Figure 16.7c shows the job plot for the minifrac. Execution for this treatment was not perfect, with significant variations in rate throughout the treatment. 10,000

20.0 Gauge BHTP

8,000

16.0

Surface Pressure

6,000

12.0

4,000

8.0

2,000

4.0

0 0.0

5.0

10.0

15.0

20.0

25.0

30.0

Rate, bpm

Pressure, psi

Slurry Rate

0.0 40.0

35.0

Elapsed Time, mins

Figure 16.7c – Minifrac example 2 job plot.

11,000

Calc BHTP

10,000

Pressure, psi

9,000 Gauge BHTP

8,000

7,000

6,000

5,000 0

5

10

15

20

25

30

35

40

Elapsed Time, mins

Figure 16.7d – Comparison between gauge and calculated BHTP for minifrac example 2. Note that whilst the calculated BHTP follows the same general trend as the gauge BHTP, the actual value is quite different. Short term variations in the trend of the calculated BHTP are caused by the variations in rate. The general offset of the data is probably caused by incorrect input data in the fracture monitoring package (in this case FracRT).

Page 141



9,000

0

8,900

-100

8,800

-200

8,700

-300

8,600

-400

8,500 14.0

15.0

16.0

17.0

18.0

19.0

20.0

Derivative dBHTP/dT

Gauge BHTP, psi

Figure 16.7d compares the gauge BHTP with the calculated BHTP. In this plot, we can see a significant variation between the actual BHTP and the calculated value. This plot is included to make an important point – be aware that any data you receive can contain errors. In this case, it looks as though the fracture monitoring software had the wrong data entered. If the calculated BHTP data had been used by itself, it would have indicated a large amount of tortuosity (note the large drop in pressure at ISIP). Remember that there is no step down test to corroborate this. In fact, as we can see from the gauge BHTP, there is very little near wellbore friction.

-500 21.0

Elapsed Time, mins

9,000

0

8,900

-1000

8,800

-2000

8,700

-3000

8,600

-4000

8,500

-5000 3.7

3.8

3.9

4.0

4.1

4.2

4.3

4.4

4.5


Figure 16.7f – Minifrac example 2 pressure decline square root time plot, with derivative.

Page 142


Derivative dP/dT0.5

Gauge BHTP, psi

Figure 16.7e – Minifrac example 2 pressure decline with derivative.

BJ Services’ Frac Manual 16. The Minifrac Figure 16.7e shows the pressure decline for minifrac example two, together with it’s derivative. There are two clear features to be noted on this plot. First, there appears to be an immediate pressure loss at ISIP of +/- 60 psi, which is probably due to tortuosity. Second, the derivative plot shows a clear change in gradient at about t = 16.1 minutes, giving a closure pressure of +/- 8725 psi (which corresponds closely with the fracture extension pressure from the SRT). Note that this closure pressure would have been very difficult to spot without the derivative plot. Figure 16.7f shows the same pressure decline, but this time against the square root of time. Once again, the derivative is included. This plot seems to indicate similar results to the previous plot (Figure 16.7e), with the fracture closure happening perhaps a little more quickly and at a slightly higher pressure.

Pressure Match The pressure match was performed using Pinnacle Technologies’ FracProPT fracture simulation software package. The first step in the process was to merge the surface data (collected in this case by FracRT) with the bottom hole data. This was performed using the data merging, conversion and editing functions of the software. Once this had been accomplished, the model was run with the “run from database data” option selected. Table 16.7a shows the initial formation data used to produce the initial fracture design and to provide a basis for the design of the minifrac. As we can see, the reservoir is very layered, with lots of thins beds of different strata. In reality it is often not necessary – or practical – to use this much definition when designing a fracture. However, it is included in this example to illustrate the detail that can be used if necessary.

Depth ft

Lithology

Stress psi

Leakoff Coefficient -05 ft/min

Young’s Modulus 6 psi x 10

Poisson’s Ratio

Fracture Toughness 0.5 psi.in

0 10220 10278 10284 10312 10320 10342 10368 10386 10420 10430 10456 10477

Shaley Sand Shaley Sand Sandstone Shale Sandstone Shale Sandstone Shale Sandstone Shaley Sand Sandstone Shale Sandstone

7120 7154 6372 7712 6393 7740 6412 7776 6439 7294 6467 7842 6496

0.0024 0.0024 0.0024 0 0.0024 0 0.0024 0 0.0024 0.0024 0.0024 0 0.0024

3.75 3.75 3.5 6.0 3.5 6.0 3.5 6.0 3.5 3.75 3.5 6.0 3.5

0.225 0.225 0.2 0.25 0.2 0.25 0.2 0.25 0.2 0.225 0.2 0.25 0.2

1500 1500 1000 2000 1000 2000 1000 2000 1000 1500 1000 2000 1000

Table 16.7a – Initial simulator data before pressure match.

Figure 16.7g shows the initial pressure match, using the original input data. As we can see, there is a large difference between the actual data and the simulated data, especially with regard to the stress data.

Page 143



Offshore Vietnam Initial Pressure Match Slurry Flow Rate (bpm) Simulated Net Pressure (psi)

20.00 5000

Observed Net (psi)

5000

16.00 4000

4000

12.00 3000

3000

8.00 2000

2000

4.00 1000

1000

0.00 0

0.00

4.00

8.00

Time (mins)

12.00

16.00

20.00

0

Figure 16.7g – Initial pressure match for minifrac example 2.

The first step is to increase the stresses in the formation to produce an approximate match at ISIP. Once this has been done, the match looks better, but is still not complete (see Figure 16.7h).

Offshore Vietnam Interim Pressure Match Slurry Flow Rate (bpm) Simulated Net Pressure (psi)

20.00 1000

Observed Net (psi)

1000

16.00 800

800

12.00 600

600

8.00 400

400

4.00 200

200

0.00 0

0.00

4.00

8.00

Time (mins)

12.00

16.00

20.00

0

Figure 16.7h – Interim pressure match after the stresses have had a first approximate adjustment. In this case, the stress gradient for the sandstone was increased from 0.62 to 0.68 psi/ft, and then 1300 psi was added to each stress. Note that the pressures are on a larger vertical scale than in Figure 16.7g.

Page 144


BJ Services’ Frac Manual 16. The Minifrac From this, there are some important points to be noted:•

• • • •

•

From the decline curve analysis, we know that the fracture closes at t = +/- 16 minutes at a bottom hole pressure of +/- 8725 psi. In Figure 16.7h, the simulated net pressure shows fracture closure at +/-16.5 minutes. This is close to reality. However, we must remember that this will change as we alter other variables. The shape of the simulated curve is close to the actual data to start with, but then deviates from the gauge data. Variables such as Young’s modulus, fracture toughness and stress will be changed for all formations in order to match this. Remember that changes in leakoff coefficient will affect the shape of the curve as well. From the decline analysis, we observed +/- 60 psi tortuosity/near wellbore friction. This should be remembered when matching the pressures. It should also be remembered that this may not be constant throughout the treatment. When pressure matching, it is essential to be able to differentiate between short term variations, and long term trends. In this example, it will be hard to adjust the model so that the pressure will rise after +/- 7 minutes, as the actual data does. This point corresponds to the time when the wellbore fluid has been completely displaced with crosslinked fluid, and this fluid now starts to enter the formation. This could be a function of tortuosity – which is very sensitive to fluid viscosity – or it could be a sign that the fracture has now started to extend at a relatively higher rate. Given that we have a decrease in pressure after +/- t = 10 minutes, it is possible that the rise and then fall in pressure is due to near wellbore effects. However, the Frac Engineer should closely examine the fluid samples and question both the blender tender and the lab technician, as this variation could be due to a change in crosslinked fluid properties (i.e. loss or reduction of crosslinker and/or buffer).

Figure 16.7i shows the final pressure match, after all the adjustments have been made to the simulator model.

Offshore Vietnam Final Pressure Match Slurry Flow Rate (bpm) Simulated Net Pressure (psi)

20.00 1000

Observed Net (psi)

1000

16.00 800

800

12.00 600

600

8.00 400

400

4.00 200

200

0.00 0

0.00

4.00

8.00

Time (mins)

12.00

16.00

20.00

0

Minifrac Example 2

Figure 16.7i – Minifrac example 2 final pressure match

Note that in Figure 16.7i, it proved very difficult to model the observed net pressure after the crosslinked fluid entered the formation. As discussed previously, this is almost certainly due to near wellbore and/or tortuosity effects. Note also that the pressure decline after shut down Page 145


BJ Services’ Frac Manual 16. The Minifrac does not have the same curve as the observed net pressure. However, it does have the same closure pressure and closure time, implying that the fluid must be leaking off at the same rate.

Depth ft

Lithology

Stress psi

Leakoff Coefficient -05 ft/min

Young’s Modulus 6 psi x 10

Poisson’s Ratio

Fracture Toughness 0.5 psi.in

0 10220 10278 10284 10312 10320 10342 10368 10386 10420 10430 10456 10477

Shaley Sand Shaley Sand Sandstone Shale Sandstone Shale Sandstone Shale Sandstone Shaley Sand Sandstone Shale Sandstone

8530 8720 8550 8780 8480 8690 8480 8730 8510 8740 8540 8790 8590

0.0015 0.0015 0.0033 0 0.0033 0 0.0033 0 0.0033 0.0015 0.0033 0 0.0033

6.0 6.0 1.5 1.0 1.5 1.0 1.5 1.0 1.5 6.0 1.5 1.0 1.5

0.225 0.225 0.2 0.25 0.2 0.25 0.2 0.25 0.2 0.225 0.2 0.25 0.2

1500 1500 500 2000 500 2000 500 2000 500 1500 500 2000 500

Table 16.7b – Final simulator data after pressure match.

In the actual pressure matching process, it became apparent that fracture only penetrated the top five formations, as described in Table 16.7b, above. Therefore, the only changes that made any difference to the simulation where those made to formations 1 through 5. In fact, as far as the simulation was concerned, the bottom 8 formations didn’t need to be in the simulator at all. Figure 16.7j shows the estimated fracture profile, as produced by the nowcalibrated fracture simulator. As we can see, the fracture grows preferentially upwards.

Fracture Profile

Stress Profile 10200

10220

10240

Depth (ft)

10260

10280

10300

10320

10340

10360

Permeability

10380 Low

10400

8000

High

8500

9000

Closure Stress (psi)

9500

10000

100

75

50 Propped Length (ft)

25

0

25

50

75

Hydraulic Length (ft)

Figure 16.7j – FracProPT estimated fracture dimensions for minifrac example 2.

Page 146


100

BJ Services’ Frac Manual 16. The Minifrac In order to get the pressure match, 60 psi of tortuosity was used. This was kept constant, throughout the simulation. However, it is possible that some of the changes in observed net pressure are due to changes in near wellbore friction (NWF), rather than the response of the formations themselves. Using a modern fracture simulator like FracProPT means that the NWF, and especially the tortuosity, can be adjusted on a continuous basis. As a result, the simulated net pressure can be made to fit any pressure match, just be adjusting NWF. This is one of the disadvantages of using these advanced models. Because they have so many factors that can be adjusted, it is possible to make the simulator match any pressure profile desired. However, it is up to the user to be able to understand which changes to the model are realistic, and which are not. A certain level of expertise, in both frac theory and in the way the model itself works, is required before the simulator can be use reliably. These are definitely not “expert” systems. After the treatment was redesigned, the job was pumped successfully and 100,000 lbs of 20/40 CarboProp was paced in the fracture. Post-treatment DST testing showed an increase in PI of between 4 and 7 times – the uncertainty being due to a leak in the DST string.

16.8

Minifrac Example 3 - Problems with Tortuosity

This well, which is located in the same field as Minifrac Example 1 (although in a different, slightly deeper formation), had a completely different response to the minifrac. Severe problems were encountered with the formation’s response to the minifrac. Although efforts were made to mitigate this, resources and expertise on location were limited, and the job eventually screened out about two thirds of the way through the treatment. Well and Formation Data Reservoir Type: Reservoir Temperature: Reservoir Pressure: Perforations: Deviation at Perforations: Liner: Treating String: Packer set at: End of Tubing: Top of Formation: Bottom of Formation: Permeability: Porosity:

Oil 145 F unknown 1121 m (3678 ft) to 1130 m (3707 ft) Vertical 7”, 23# 3.5”, 9.3# tbg 1105 m (3625 ft) 1115 m (3658 ft) 1121 m (3678 ft) 1157 m (3796 ft) 30 mD 20%

Treatment Data Wellbore Fluid: Treatment Fluid: Treatment Volume: Displacement Fluid: Displacement Volume: Treatment Rate:

Produced Fluids Crosslinked gel (SpectraFrac G 4500) 3 45 m (314 bbls) Slick water 3 5.3 m (33.3 bbls) 3 3 m /min (18.8 bpm)

Figure 16.8a shows the treatment plot for this minifrac. This is a well executed minifrac. There is a slight spike in the rate, as it is being increased initially, but this is not significant. The major point of interest, however, is the large pressure drop in the gauge BHTP just as the pumps are shut down. This is shown in Figure 16.8b, which displays more detail of the BHTP at shut down.

Page 147


BJ Services’ Frac Manual 16. The Minifrac 5,000

25

Slurry Rate

4,000

20

3,000

15

Surf. Press.

2,000

10

1,000

5

0

Rate, bpm

Pressure, psi

Gauge BHTP

0 0

10

20

30

40

50

60

Time, mins

Figure 16.8a – Minifrac example 3 treatment plot.

3,500

A

Gauge BHTP, psi

3,200

B

2,900

E D

2,600

C

2,300

2,000 10

15

20

25

30

35

40

Time, mins

Figure 16.8b – Minifrac example 3, detail of post-treatment pressure decline.

Figure 16.8b shows 5 main points of interest, labeled A to E as follows A B

Initial pump shut down. Note how the pressure drops immediately by 400 to 500 psi as soon as the pumps shut down. This is due entirely to near wellbore friction. A step down test would be required to tell for sure if this was due to perforation friction or tortuosity, but this was not performed. However, as the zone had been re-perforated just prior to the minifrac,

Page 148


BJ Services’ Frac Manual 16. The Minifrac it is likely that this pressure drop is due to tortuosity. This pressure drop is very large and is an immediate cause for concern. This point shows a clear change in gradient at +/- 2350 psi and is almost certainly fracture closure. This area of the plot is unusual. It is rare (but not unknown) to see a post treatment pressure decline of this shape – especially as the decline actually increases for a short period of time. This area of the plot is probably caused by poor communication between the fracture and the wellbore, and is potentially another sign of tortuosity. Point E, obtained by extrapolating the straight line pressure decline back until it gets to the point at which the pumps were shut down, is probably a good approximation for the true ISIP. A G function plot will be used to confirm this.

C D

E

A post treatment pressure decline like Figure 16.8b should set alarm bells ringing in the head of any experienced Frac Engineer. It is obvious that there is a severely restricted flow path in the near wellbore area. This means that the net pressure, which initially appears to be +/- 900 psi, is in fact probably less than half of this. This in turn means that the fracture is substantially smaller than it initially appears to be. In addition, the restricted flow paths between the fracture(s) and the wellbore, will make it very difficult to place even moderate concentrations of proppant. Figure 16.8c shows the square root of time pressure decline plot. This plot shows a high degree of similarity with the pressure decline plot in Figure 16.8b. On this plot, with a slightly expanded vertical scale, the closure can be seen to be around 2320 psi. 3,200

3,000

Gauge BHTP, psi

2,800

2,600

2,400

2,200

2,000 4

4.5

5

5.5

6

6.5

7


Figure 16.8c – Minifrac example 3, square root time pressure decline plot.

Figure 16.8d shows the Horner plot for the pressure decline. This plot is a little ambiguous, with potentially two or three different gradients and y-axis intercepts. Consequently, this plot will only be used if the other plots prove to be unreliable. In order to help verify both the closure pressure and the true ISIP, a G function plot is used, as shown in Figure 16.8e. Obviously, to do this we must assume a 2-D geometry. In this case, radial geometry was assumed. However, the fact that the plot is based on 2-D geometry does not detract from its ability to pick the true ISIP, and the closure pressure will also be reasonably reliable.

Page 149



2,800

Gauge BHTP, psi

2,600

2,400

2,200

2,000

1,800 0.0

0.5

1.0

1.5

2.0

2.5

3.0

Horner Time

Figure 16.8d – Horner plot for minifrac example 3. Note that several lines may be fitted to the final slope on the LHS of this plot. In fact, the reservoir pressure is substantially lower than that indicated on the plot (as the well is produced by ESP’s), so all of these lines may be unreliable.

3,000

Gauge BHP, psi

2,800

2,600

2,400

2,200

2,000 0

0.5

1

1.5

2

2.5

G Function

Figure 16.8e – G Function plot for minifrac example 3. Note the true ISIP of +/- 2730 psi, and the closure pressure of +/- 2320. These values are in agreement with the value obtained from other plots, such as the pressure decline and the square root time plots.

Page 150



Pressure Match with MFrac 3-D Fracture Simulator. After the data was imported into the MFrac 3-D fracture simulator, via the data collection programme MView, an initial run was performed to see how close the initial, pre-minifrac fracture model was. The results are shown in Figure 16.8f, above.

Bottomhole Treating Pressure

20

BHTP Measured BHTP Measured Surface Rate

BHTP (psi)

4000

16

3500

12

3000

8

2500

4

2000 0

10

20

30

Time (min)

40

50

Rate (bpm)

4500

0 60

Figure 16.8f – MFrac output showing the initial pressure match before any adjustments were made. There is very little agreement between the predicted and actual BHTP’s.

As can be seen in Figure 16.8f, to begin with there is very little agreement between the initial fracture model and the actual response of the formation. Remember also that the BHTP is from a gauge. We can see that the slope of the pressure decline is significantly different, indicating (in this example), that the actual fluid loss rate was somewhat faster than predicted. In addition, the pressures predicted whilst pumping are completely different both in magnitude and in the trend that they follow. Clearly, this model needed significant adjustment. This is why we perform minifracs. The effects of tortuosity also manifest themselves on this plot. We can see that, because of the huge pressure drop as the pumps shut down, the model predicts lower pressures whilst pumping and higher pressures during the decline. Obviously, some allowance needs to be made in the model for the tortuosity. It is at this point that experience and intuition start to take over. The fact is, tortuosity is not necessarily constant throughout the treatment. The fall in measured BHTP that we see whilst pumping could be due entirely to a continuous decrease in near wellbore tortuosity. Or it could be due to a reduction in perforation friction as more perforations are opened up. Worse still, it could be due to a combination of tortuosity, perforation friction and fracture geometry effects. However, three other factors help the Frac Engineer. Firstly, we need to remember that we have pumped no proppant and we have kept the rate constant. Changes in tortuosity are usually (but not always) associated with either a change in rate, or the action of the proppant. Secondly, changes in tortuosity (other than those associated with rate) tend to produce rapid changes in the BHTP (“spikes” and “dips”), rather than slow, smooth changes. Lastly, the zone had just been re-perforated prior to the treatment, and probably had very low perforation friction (although this cannot be guaranteed – perforating does go wrong occasionally). Therefore, it is probably a reasonable assumption that – in this case - the pressure loss due Page 151


BJ Services’ Frac Manual 16. The Minifrac to tortuosity is relatively constant. However, the Frac Engineer must be aware that this is not necessarily the case.

Bottomhole Treating Pressure

3500

20

BHTP (psi)

BHTP Measured BHTP Measured Surface Rate

15

3000

10

2500

5

2000 0

10

20

0 40

30

Time (min)

Rate (bpm)

4000

Figure 16.8g – Final MFrac output, after the model has been adjusted.

In Figure 16.8g, we can see the results of the pressure match. The match is not perfect, but is pretty close. At the beginning of the treatment, the initial pressure spike has not been matched. Later on, at the start of the pressure decline, matching the shape of the curve proved to be very difficult. In this area, the general trend has been matched, whilst the curve has not. The effects of the poor communication between the fracture(s) and the wellbore are very difficult to model mathematically. The changes made to the model are listed in Table 16.8a. Formation Property

Upper Shale

Sandstone

Lower Shale

Before

After

Before

After

Before

After

0.75

0.62

0.70

0.62

0.75

0.62

Young’s modulus, psi x 10

0.6

0.3

0.4

0.3

0.6

0.3

Poisson’s ratio

0.25

0.25

0.25

0.25

0.25

0.25

1000

1000

1000

7500

1000

1000

0.0004

0.0004

0.007

0.015

0.0001

0.0001

na

na

0

550

na

na

Stress Gradient, psi/ft 6

1/2

Fracture Toughness, psi in

-1/2

Leakoff Coefficient, ft min Tortuosity ∆P, psi

Table 16.8a – Changes made during the pressure matching process.

Page 152



Treatment Results The fact that we now have a reasonable pressure match does not alter the fact that it will be very difficult to place the treatment. The pressure drop due to tortuosity is very large. This means that it will be very difficult to place proppant inside the fracture – it will almost certainly bridge off in the near wellbore area. Therefore, the tortuosity needs to be removed. The processes for doing this were described in Section 10.1, and were originally detailed by Cleary et al in SPE 25892 and Køgsball et al in SPE 26796. In fact, the normal process to cure tortuosity – such as pumping a series of proppant slugs – were not an option in this instance. The well was drilled in a remote location and the expertise necessary for such an operation was not available on location. In addition, the operating company was not willing to go through the potentially lengthy processes needed – the economics of the situation demanded low cost treatments, in order for them to be justifiable. In the end, it was decided to place a +/- 6 ppa proppant slug in the middle of the pad, and observe what happened as it went into the formation. If a significant pressure rise was observed, the plan was to shut down and re-assess the situation. In fact, the well screened out as soon as the proppant slug hit the perforations. However, once the pressure had fallen and more fluids had been mixed, it was possible to break down the formation again and re-start the treatment. This time, the well treated at a significantly lower pressure – indicating that the proppant slug may have helped to remove some of the tortuosity. As it turns out, not all of the tortuosity was removed. The treatment screened out at 8 ppa, with 35,000 lbs of the planned 50,000 lbs placed in the formation. The rapid pressure rise associated with the screenout indicated a near wellbore event. However, the operator considered this a success – given the circumstances – and the production increase more than justified the expense of the treatment.

16.9

Minifrac Example 4 – Perforation Problems

This minifrac was carried out on a well in New Zealand, using an oil-based fracturing fluid. Oilbased fracturing fluids are harder to pressure match, as there is less data available on items such as the wall-building coefficient and tubing friction. The properties of these fluids are highly dependent upon the hydrocarbon used as the base for the fluid. Even fluids mixed with diesel show a marked variation in properties, when using with different sources of diesel. Prejob testing is essential. Luckily, on this treatment, bottom hole pressure gauges were used, allowing uncertainties due to tubing friction to be eliminated. New Zealand, as far as the fracturing industry is concerned, is a remote location and the success or failure of these treatments depended as much upon the logistics and organisation of the operations, as it did upon the formation or the skill of the crew. Well and Formation Data Reservoir Type: Reservoir Temperature: Reservoir Pressure: Perforations: Deviation at Perforations: Casing: Page 153

Gas 185 F 5200 psi 3397 m (11,145 ft) to 3407 m (11,178 ft) Vertical 7”, 23#


BJ Services’ Frac Manual 16. The Minifrac Treating String:

3.5”, 9.3# tbg to 3342.6 m (10,966 ft) 2 ”, 6.5# tbg to 3379 m (11,086 ft) 3343 m (10,968 ft) 3379 m (10,966 ft) 3397 m (11,145 ft) 3407 m (11,178 ft) 12 mD n/a

Packer set at: End of Tubing: Top of Formation: Bottom of Formation: Permeability: Porosity: Treatment Data Original Wellbore Fluid: Treatment Fluid:

Formation water and gas Crosslinked gelled diesel (Super Rheogel 500) 3 237 m (1488 bbls) Diesel + surfactant 3 15.3 m (96.1 bbls) 3 2.4 m /min (15 bpm)

Treatment Volume: Displacement Fluid: Displacement Volume: Treatment Rate:

First Step Rate Test The first step rate test was pumped using diesel with surfactant. Initially, a wellbore volume was pumped ahead, to ensure that no gas remained in the well. Then the pumps were shut down for 15 minutes, to ensure that the effects of this injection did not cloud the results of the step rate test. It should be remembered that on location, the first step rate test was followed immediately by the minifrac, and neither where analysed until later on, after the BH gauge data had been retrieved. Therefore, the results of the step rate test were not available before the minifrac was pumped. The significance of this will become apparent as we progress. Figure 16.9a shows the job plot for the first step rate test, Figure 16.9b shows the step up crossplot and Figure 16.9c shows the step down crossplot.

12000

12

10000

10

Gauge BHTP

8

6000

6 Slurry Rate

4000

4

Surface Pressure

2000

2

0

0 0

10

20

30

40

50

60

70

80

Elapsed Time, mins

Figure 16.9a – Job plot for Minifrac Example 4, Step Rate Test 1

Page 154


90

Slurry Rate, bpm

Pressure, psi

8000

BJ Services’ Frac Manual 16. The Minifrac 11000

Gauge BHTP, psi

10000

9000 Fracture Extension at +/- 9100 psi

8000

7000

6000 0

1

2

3

4

5

6

7

8

9

10

Slurry Rate, bpm

Figure 16.9b – Step up crossplot for Step Rate Test 1. Fracture extension seems to be at approximately 9100 psi.

The step rate test was executed reasonably well, except for one mishap when bringing an additional pump in line, when going for 10 bpm. The results from the analysis of the step up crossplot, indicate a fracture extension of 9100 psi. This gives an extension gradient of 0.82 psi/ft – high, but not exceptionally so. However, the real problems show themselves in Figure 16.9c – the step down crossplot. This plot clearly shows the characteristic shape of perforation friction. 10500

Gauge BHTP, psi

10200

9900

9600

9300

9000 0

1

2

3

4

5

6

7

8

9

10

Slurry Rate, bpm

Figure 16.9c – Step down crossplot. Note the concave shape of the best fit curve, indicating that the near wellbore friction is dominated by the perforations.

Page 155



Minifrac 12000

18

Gauge BHTP

15

Pressure, psi

8000

12

Surface Pressure

6000

9

Slurry Rate

4000

6

2000

Rate, bpm & Proppant Conc, ppa

10000

3 Proppant Conc

0

0 0

10

20

30

40

50

60

70

80

90

Elapsed Time, mins

Figure 16.9d – Minifrac Example 4 job plot.

The minifrac was pumped directly after the step rate test, before any analysis was carried out on the step rate test data. Initially, the minifrac was programmed at 8 bpm and without a proppant slug. However, previous experience had shown that these formations were subject to tortuosity, and so it was decided to include the proppant slug, to assess how conductive the near wellbore region was. Figure 16.9e shows what happened when the proppant slug arrived at the formation. Note that this plot shows bottom hole proppant concentration. 10000

20

16

Slurry Rate

Pressure, psi

Gauge BHTP

9200

12

+/- 400 psi Pressure Rise as Proppant Reaches Perfs

8800

8

8400


9600

4

Bottom Hole Proppant Conc

8000

0 30

31

32

33

34

35

36

37

38

39

40

Elapsed Time, mins

Figure 16.9e – Detail of job plot showing bottom hole proppant concentration, gauge BHTP and slurry rate, as the proppant slug enters the formation. Note the +/- 400 psi rise in pressure.

Page 156


BJ Services’ Frac Manual 16. The Minifrac From Figure 16.9e, we can see a +/- 400 psi pressure rise as the proppant enters the formation. This is not good, and indicates that we will not be able to get even moderate proppant concentration slurries into the formation. 9500

Gauge BHTP, psi

9200

+/- 650 psi Near Wellbore Friction

8900

8600

Fracture Closure at +/- 8350 psi (0.75 psi/ft)

8300

8000 48

50

52

54

56

58

60

62

64

66

68

Elapsed Time, mins

Figure 16.9f – Minifrac pressure decline, showing +/- 650 psi near wellbore friction and a closure pressure of +/- 8350 psi. 8800

Gauge BHTP, psi

8600

8400

Fracture Closure at +/- 8230 psi (0.74 psi/ft)

8200

8000

7800 0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

5.5

6

Square Root Time, mins0.5

Figure 16.9g – Square root of time plot for the minifrac pressure decline. This gives a slightly lower closure pressure than Figure 16.9f, at +/- 8230 psi.

Figure 16.9f show the ISIP and pressure decline after the minifrac. As we can see, this also does not look good. Immediately, we can see a +/- 650 psi pressure drop as the pumps are shut down. This can only be due to near wellbore friction, as we are using gauge bottom hole treating pressure. This result, together with the result from the step down test, indicates that Page 157


BJ Services’ Frac Manual 16. The Minifrac this well has severely restricted perforations. As a result of this analysis, the decision was made to re-perforate and run another step rate test. We can also see a closure pressure of +/8350 psi, which is slightly different from the closure seen in Figure 16.9g, the square root of time pressure decline plot. This gives a lower closure pressure of +/- 8230 psi. These closure pressures translate to gradients of 0.748 and 0.739 psi/ft (16.9 to 16.7kPa/m) respectively.

Second Step Rate Test 10000

20 Gauge BHTP

16

Surface Pressure

6000

12

4000

Rate, bpm

Pressure, psi

8000

8

Slurry Rate

2000

4

0

0 0

5

10

15

20

25

30

35

40

45

50

Elapsed Time, mins

Figure 16.9h – Job plot for second step rate test.

9400

Gauge BHTP, psi

9200

9000

8800

8600 0

2

4

6

8

10

Slurry Rate, bpm

Figure 16.9i – Step down crossplot for the second step rate test.

Page 158


12

BJ Services’ Frac Manual 16. The Minifrac The second step rate test was performed after the well was re-perforated. Originally, the intent had been to re-perforate the entire 10 meter section of (original) perforations. However, on entering the well, the wireline operators indicated that there was some kind of fill in the well, sufficient to block access for the perforating guns to the lower 7 meters of perforated interval. The decision was made on location to shot holes in the upper 3 meter section only. After perforating the upper 3 meters of the zone, the BH pressure gauges were re-run into the well, and the second step rate test was performed. Figure 16.9h shows the job plot for this, whilst Figure 16.9i shows the step down crossplot. By comparing Figures 16.9c and 16.9i, we can see that the near wellbore situation has changed dramatically:1. The slope of the best fit curve as changed from concave (perforation dominated) to convex (tortuosity dominated). 2. The overall bottom hole pressure has dropped significantly. At 8 bpm, the first step rate test shows a BHTP of +/- 9950 psi, whereas at 8 bpm in the second step rate test, the BHTP is +/- 9270 psi. So, as a result of the re-perforation, the restricted perforations have been removed (actually, by-passed) and the overall level of near wellbore friction appreciably reduced.

Minifrac Pressure Match The minifrac was performed before the well was re-perforated, and so still includes the effects of the restricted perforations. MFrac was used for this pressure match. Figure 16.9j shows the predicted and actual bottom hole treating pressures before the pressure match was performed, whilst the post pressure match pressures are shown in Figure 16.9k. This treatment was difficult to pressure match, largely due to the dynamic nature of the restricted flow path in the near wellbore. As we can see from Figure 16.9k, the early part of the treatment, at the lower rate, was not matched. In fact, the only part of the treatment that could be matched was after the proppant slug had entered the formation. Consequently, because of the unreliable nature of the data and the analysis, the final design had to be pretty cautious.

Figure 16.9j – Minifrac Example 4 BHTP plot before pressure matching.

Page 159



Figure 16.9k – Minifrac Example 4 pressure match using MFrac.

The minifrac was performed before the well was re-perforated, and so still includes the effects of the restricted perforations. MFrac was used for this pressure match. Figure 16.9j shows the predicted and actual bottom hole treating pressures before the pressure match was performed, whilst the post pressure match pressures are shown in Figure 16.9k. This treatment was difficult to pressure match, largely due to the dynamic nature of the restricted flow path in the near wellbore. As we can see from Figure 16.9k, the early part of the treatment, at the lower rate, was not matched. In fact, the only part of the treatment that could be matched was after the proppant slug had entered the formation. Consequently, because of the unreliable nature of the data and the analysis, the final design had to be pretty cautious. Stresses Young’s Modulus Leakoff Perforations

Total NWB Friction

All the formations’ stresses had to be significantly increased. All the formations’ moduli had to be significantly decreased. The only way the leakoff could be matched to allow significant fluid loss through the shale formations above and below the zone of interest. The initial model had 170 x 0.3” perforations. Obviously, with 7m of the 10 m covered by fill, this number had to be reduced. However, in order to get a pressure match, the perforations had to be modeled as 1 x 0.07”! This was the only way that the BHTP during the change in rate at t = 46 minutes could be matched. In addition to the restricted perforations, an extra 600 psi in near wellbore friction had to be added, in order to match the pressure drop as the pumps were shut down.

It is unlikely that the perforations had been reduced to the equivalent of one 0.07” diameter perforation. For one thing, the average grain diameter of 20/40 Carbolite (the proppant used in the proppant slug) is 730 microns or 0.029” (from manufacturer’s data). Thus, the perforation opening is less than 2.5 times the median grain diameter. It is probable that a 4 ppg proppant slug would have blocked this off. The simulator results also show that the overall near wellbore friction Figure is probably a result of a combination of poor perforations and tortuosity.

Page 160


BJ Services’ Frac Manual 16. The Minifrac This illustrates the two sides of using simulator. On one hand it is clear that we have very restricted perforations and that something should to be done about this. (However, we probably could have worked this out without the simulator, based on the step down test and the pressure drop at the end of the minifrac.) On the other hand, the extent to which the perforations are blocked is probably exaggerated by the simulator. Good engineering judgement, based on experience and knowledge of the underlying theories, needs to be applied in order to decide what is realistic and what is not. As a consequence of the restricted near wellbore situation, and the fact that the well was reperforated, much of the data used to produce the pressure match is not relevant to the main treatment design. Only the fluid leakoff data and – to a lesser extent – the stresses and moduli – can be used. For modeling the final treatment, the perforation data was re-set to fifty 0.3” diameter holes, and the total near wellbore friction reduced to 200 psi (based on the second step rate test). In an ideal world, where the Frac Engineer has a free hand with regard to technical issues, the minifrac should have been repeated. In reality, it was not repeated, for a variety of reasons.

Main Treatment Although the re-perforating had dramatically improved the near wellbore situation, it was clear that there were potentially still some problems with tortuosity. Without a minifrac, complete with proppant slug, it was difficult to assess just how bad this problem was. Consequently, the main treatment was designed with three proppant slugs in the pad. 1. 2. 3.

100 mesh sand at 1 ppa. 20/40 Carbolite at 4 ppa. 20/40 Carbolite at 6 ppa.

These stages were spaced out so that the effect of each one could be assessed before the next one arrived at the perforations. Based on the response of the formation to these stages, the treatment would be redesigned on the fly. Figure 16.9l shows the main treatment job plot and Figure 16.9m shows a detail of the bottom hole sand concentration as the 3 proppant slugs arrive at the formation hole sand concentration as the 3 proppant slugs arrive at the formation. 10000

25

Calculated BHTP

20

Pressure, psi

Slurry Rate

6000

15 Surface Pressure

4000

10

2000

5

Proppant Concentration

0 0

20

40

60

80

100

120

0 140

Elapsed Time, mins

Figure 16.9l – Job plot for the main treatment for Minifrac Example 4. Note the proppant concentration is measured at the surface.

Page 161



8000


8000

25

7600

20

7200

15

Slurry Rate

6800

10 Surface Pressure

6400


Pressure, psi

As we can see, there was very little response from the formation as these three stages went through the perforations. On the basis of this, it was decided to continue with the treatment as planned. As the job progressed, it became apparent that the proppant was entering the formation very easily. Originally, the treatment had been planned for 107,000 lbs of proppant, pumped at 1 to 6 ppa. After assessing the well’s response to the proppant slugs, and watching the early proppant stages, it was decided to extend the treatment. 130,000 lbs of proppant was placed, by extending the 4 and 5 ppa stages.

5 BH Proppant Concentration

6000

0 25

30

35

40

45

50

Elapsed Time, mins

Figure 16.9m – Detail of the main treatment for Minifrac Example 4, showing the formation’s response to the proppant slugs. Proppant concentration is bottom hole.

References Howard, G.C., and Fast, C.R.: Hydraulic Fracturing, Monograph Series Vol 2, SPE, Dallas, Texas (1970). Gidley , J.L., et al.: Recent Advances in Hydraulic Fracturing, Monograph Series Vol 12, SPE, Richardson, Texas (1989). Economides, M.J., and Nolte, K.G.: Reservoir Stimulation, Schlumberger Educational Services, 1987. Economides, M.J.: A Practical Companion to Reservoir Stimulation, Elsevier, 1992 Nolte, K.G.: “Determination of Fracture Parameters from Fracturing Pressure Decline”, paper SPE 8341, 1979. Dempsey, Brett.: “Competing with G Function Analysis”, BJ Services’ Engineering News, Vol. 12, No 1, Winter 2001 Nolte, K.G.: “A General Analysis of Fracture Pressure Decline With Application to Three Models”, paper SPE 12941, SPEFE, p. 571-583, 1986

Page 162


BJ Services’ Frac Manual 16. The Minifrac FracRT Version 4.6 User’s Manual, BJ Services, 1995 Cleary, M.P, et al.: ”Field Implementation of Proppant Slugs to Avoid Premature Screen-Out of Hydraulic Fractures with Adequate Proppant Concentration”, paper SPE 25892, presented at the SPE Rocky Mountain Regional/Low Permeability Reservoirs Symposium, Denver CO, April 1993. Køgsball, H.H., Pits, M.J., and Owens, K.A.: “Effects of Tortuosity in Fracture Stimulation of Horizontal Wells – A Case Study of the Dan Field”, paper SPE 26796, presented at the Offshore Europe Conference, Aberdeen, UK, Sept 1993. FracproPT Version 9.0 onwards on-line Help, Pinnacle Technologies/Gas Research Institute, July 1999 onwards. MFrac III Version 3.5 onwards on-line Help, Meyer and Associates Inc, December 1999 onwards.

Page 163


BJ Services’ Frac Manual 17. Designing the Treatment

17.

Designing the Treatment

As previously discussed, different types of formation require different types of fracture. For instance, a high permeability formation requires more fracture conductivity than a low permeability formation. This section of the manual contains important tips on which fracture characteristics the Engineer should be designing for, and how to go about achieving them. As a quick rule-of-thumb, the following guidelines may be used:i) ii) iii) iv)

17.1

Skin bypass fracturing is for when eliminating the effect of the skin or extremely low cost are the primary goals. High permeability fracturing is when maximising fracture conductivity is the primary goal. Low permeability fracturing is when maximising fracture inflow area is the primary goal. Frac and pack fracturing is when fracture conductivity and sand control are the dual primary goals.

General

At its most basic level, every fracture is designed to do the same thing – increase the productivity (or injectivity) of the fractured interval. At the limit, all a fracture has to be is more conductive than the skin damage around the wellbore in order to do this. This is a relatively easy thing to accomplish, which is why skin bypass fracture treatments are very low cost and are also easy to perform. However, often simply bypassing the skin is not enough – bigger production gains are needed to economically justify the treatment or to efficiently develop the reservoir. In such cases, the fracture has to be significantly more conductive than the formation. When this happens, it is easier for the formation fluids to flow down the fracture, than it is to flow through the formation and into the perforations, and the productive interval will have a negative skin. True stimulation has occurred, rather than just simple damage removal or elimination. The best way to assess if the fracture is more conductive is to calculate the relative or dimensionless conductivity, CfD as previously discussed in Section 10.3:CfD =

kp w ¯ xf k .............................................................................. (10.1)

where kp is the permeability of the proppant, w ¯ is the average fracture width, xf is the fracture half length and k is the permeability of the formation. Generally, if the CfD is greater than 1, then the fracture is more conductive than the formation. This seems easy enough to calculate, but there are two important points which can often make estimates of CfD unreliable:1

The proppant permeability is often not easy to find, nor indeed is it a constant. The permeability of the proppant will vary with closure pressure. As the reservoir pressure drops (or the drawdown is increased), the closure pressure on the proppant will change, possibly producing more fines and a permanent drop in permeability. If the proppant or sand is at the upper limit of its closure stress range, a drop in reservoir pressure can produce a significant drop in fracture conductivity. In addition, high rate wells (especially gas wells) can experience non-Darcy flow through the proppant pack, which can dramatically decrease the effective permeability. Lastly, multi-phase and/or non-Darcy flow can also significantly reduce the proppant pack’s permeability. Therefore, the value used for kp needs to be an effective permeability, under a given set of production conditions. The Frac Engineer should also be aware of how these production conditions can vary over the life of the well and design for this. Recent

Page 164


BJ Services’ Frac Manual 17. Designing the Treatment information published by proppant vendors and the StimLab consortium, provides detailed information for the effective permeability of different proppant types under many different condition. 2.

The effective width has to be estimated from a fracture model or simulator. The width generated by these models (which will vary from model to model, even when the same formation and treatment parameters are used), is highly dependent upon the Young’s modulus and closure pressure of the formation. These two parameters are often unknown and may even (as in the case of Young’s modulus in certain formations) be variable.

Therefore it is important to realise that a fracture must be designed with a safety margin built into the fracture conductivity, to allow for all these uncertainties. It is therefore recommended that the Frac Engineer design for a minimum CfD of 20 to 40% greater than theoretically required (see Section 17.9) Finally, the Frac Engineer should be aware of the upper limits for fracture conductivity. As the conductivity increases, the contrast in conductivity between the formation and the fracture will increase as well. Eventually, a point will be reached at which the formation is delivering reservoir fluids to the fracture as fast as it can. Further increases in fracture conductivity (or the conductivity contrast) will therefore produce no subsequent further increase in production. This is the so-called infinite conductivity situation, where the fracture behaves as if it has an infinite conductivity compared to the formation. Making a fracture this conductive is simply a waste of proppant, as the same production increase can be achieved for a reduced propped width. Generally, therefore, it is often not cost effective to design a treatment to produce a CfD of greater than 10, unless the formation permeability is very low.

17.2

Designing for Skin Bypass

Skin bypass fractures are the easiest fractures to design. Operationally, they are simple to execute and have a relatively low probability of screening out early. This is because they are relatively insensitive to inaccuracies in formation data. Often, the two biggest factors influencing the design of the skin bypass frac, are not formation- or perforation-related. In fact, the biggest influences are the volume of fluid already in the wellbore (which acts as additional pad fluid) and the volume of fluid and proppant that can be pre-pared and pumped on often very limited or remote locations. (Remember that skin bypass fracs are very low cost treatments, and that performing workovers or similar operations - allowing the wellbore volume to be reduced or eliminated - are often unfeasible.) The volume of fluid in the wellbore is often significantly greater than the desired pad volume. This means that the size of the actual fracture created is usually out of the control of the Frac Engineer, and the only factor that can be controlled is the volume of proppant pumped into the fracture. This in turn is often limited by the available equipment or deck space. However, it should be noted that highly effective skin bypass fracs can be placed with very small volumes of proppant, provided the effective pad volume can be minimised (so that the proppant doesn’t get too dispersed in the fracture). This inability to control either minimum pad volume or maximum proppant volume, actually makes designing skin bypass fracs very simple, as the number of variables available for the Frac Engineer to alter are greatly reduced. Skin bypass fracs should really be thought of as an alternative to acidising. Consequently, they should be designed to be cost-effective, as compared to a matrix acid treatment. Relative to other types of fracturing, this means that low cost and ease of operation are the biggest single considerations. These treatments should be cheap, relatively low-tech and easy to pump.

Page 165


BJ Services’ Frac Manual 17. Designing the Treatment The following Equation describes the production increase that can be expected from a skin bypass treatment, as described in Section 12 (after SPE 56473):J Jo

-s

=

ln[re/(rw.e )] ln[4/(CfD.xfD)] ................................................................. (17.1)

This Equation provides a more realistic measure of the effectiveness of the fracture than methods based solely on assessing the CfD, as it takes the skin factor into account. Given that the dimensionless fracture half length, xfD, is defined as follows:xfD

=

xf re ................................................................................. (17.2)

Then the lower part of the RHS of Equation 17.1 can be reduced as follows:CfD. xfD = =

kp w ¯ xf k

xf re ........................................................................ (17.3)

kp w ¯ re k ............................................................................... (17.4)

For skin bypass fracturing, it seems that the production increase is largely independent of propped fracture length per se. However, it must not be forgotten that as average width is a function of fracture length, (and vice versa). In this case, w ¯ is the average propped fracture width, not the average created fracture width. This is a significant difference that helps to reinforce the concept that skin bypass fracture effectiveness is much more dependent upon average propped fracture width than it is upon fracture length. This is why skin bypass fracs can be so cost effective and easy to perform – almost any kind of pad will suffice, as long as the proppant is kept near the wellbore at a sufficient concentration.

17.3

Designing for Tip Screenout

The tip screenout (or TSO), as previously described in Section 10.4, is a technique used to artificially induce increased fracture width, whilst at the same time limiting fracture half-length and height. In order to obtain the tip screenout, proppant has to be forced into the tip of the fracture. Once sufficient proppant has been forced into the tip, the fracture fluid is no longer able to maintain a positive net pressure at the tip, and the fracture stops propagating. At this point, the fracturing fluid is still being pumped into the fracture at a rate substantially greater than the leakoff rate. This means that the fracture volume has to increase somehow. As the treatment has artificially stopped the fracture from increasing length or height, the width has to increase. In order for the width to increase, extra net pressure (i.e. energy) is required to further compress the formation either side of the fracture. This is why a TSO is characterised by a steady increase in net (and hence surface) pressure from the point at which the TSO initiates until the end of the treatment. Obviously, the TSO must not happen too early. If this happens, the fracture may not achieve the required vertical coverage of the formation. In addition, it must be remembered that the longer the fracture is, the easier it is to produce width. Therefore, if the TSO occurs early, the treatment may not be able to produce sufficient width before the maximum surface treating pressure is exceeded – a screenout. In order to generate a TSO at the correct point in the treatment, it is necessary to pump a pad, sized such that it will have leaked off completely at the point at which the TSO must occur. In order for the proppant following the pad stage to produce the TSO, all of the pad fluid has to leak away, otherwise the proppant will not get into the fracture tip.

Page 166


BJ Services’ Frac Manual 17. Designing the Treatment Therefore, in order to achieve a TSO, a formation must have a relatively high fluid leakoff rate. It is generally not possible to produce a TSO on very low permeability formations. This is generally not a problem however, as TSO’s are usually only required on high permeability formations. In order to be able to predict (and hence control) the point at which the TSO occurs, it is therefore essential to know the rate at which the pad fluid is leaking off. This can only usually be achieved if a minifrac has been pumped prior to the treatment (unless there is a considerable history of fracturing a particular formation, and the characteristics of this formation have been shown to be reliable). In addition, it is essential to retain uniform frac fluid characteristics throughout the minifrac and main treatment. If the fluid characteristics change, the leakoff rate will almost certainly change. The minifrac is also essential for determining the Young’s modulus of the formation. This has a big influence on a TSO treatment, as it determines how much net pressure is required to produce a given fracture width. It is the Young’s modulus that determines whether or not the required width can be achieved without exceeding the maximum surface treating pressure. Therefore, the two key points to designing a successful TSO treatment are the fluid leakoff and the Young’s modulus. Every effort should be made to determine accurate values for these variables.

17.4

Designing for Frac and Pack

Frac and Pack treatments contain all the elements described in Section 17.3 (above) for a TSO design, plus some extra elements specific to the completion being installed.

Frac Pack Slurry Blank Pipe

‘Packed’ Gravel or Proppant

Figure 17.4a – The diagram on the LHS illustrates the position of the slurry and the ‘pack’ at screenout – with the top of the ‘packed’ proppant at the top of perforations, and the annular space between the completion and the wellbore full of slurry, up until the crossover ports. The RHS shows the position of the pack after all the proppant has been allowed to settle.

Figure 17.4a illustrates a schematic of the frac and pack completion, complete with the setting tool (assumed to be in the squeeze position). Towards the end of the treatment – as with any TSO design – the formation will screenout, preventing the pumping of any further slurry into Page 167


BJ Services’ Frac Manual 17. Designing the Treatment the formation. However, the frac and pack treatment is designed with an extra volume of slurry on the end of the final stage. This stage is added so that the annular space between the completion and the casing can be filled with proppant. When the completion is made up, sections of “blank pipe” (usually regular P-110 tubing) are added above the screens. This produces extra distance between the crossover ports (the point at which the slurry enters the annulus) and the screens. This extra distance provides an extra volume of slurry in the annulus after the screenout, so that once all the proppant has settled down onto the pack, the height of the pack is significantly above the top of the screens. So - basically - the frac and pack treatment is a TSO treatment, designed with some extra slurry on the final stage, so that the annular space is completely packed to above the height of the screens. This is verified after the treatment by pumping a circulation test (also referred to as re-stressing). By comparing the results of these with a similar pre-frac test, the height of proppant in the annulus can be calculated, as follows:2

H

=

(Pfinal - Pinitial) kp A 2 0.45 .................................... (17.5) (1279 µ q A) + (4.63 ρ q kp )

In Equation 17.5, Pinitial is the surface pressure for the pre-frac circulation test (psi), Pfinal is the surface pressure for the post-frac circulation test (psi), kp is the proppant permeability (darcies), A is the annular capacity between the casing and the blank pipe (ft3/ft), µ is the viscosity of the fluid being circulated (cp), q is the flow rate (bpm) and ρ is the density of the fluid (ppg). H is the height of proppant above the screens in feet. Use the same fluid, pumped at the same rate, for both the pre- and the post-frac tests. The above relationship is based on the Forcheimer Equation (Equation 10.4) and so allows for inertial flow effects.

17.5

Designing for Tight Formations

In general, tight formations have low permeability, hard rock and require some form of stimulation in order to be economic. Normally, these formations require a completely different approach to the treatments described in the previous sections. These previous treatments have relied upon the bypassing of skin damage and on the conductivity of the fracture to produce the production increase. This is not true of tight formations, in which the skin factor is usually relatively low, and it is easy to obtain a fracture that is many times more conductive than the formation. In fact, for the purposes of fracture stimulation, it is possible to define a tight formation as one in which the most important fracture characteristic is not propped width, but propped length. When defining a tight formation, it is also useful to think in terms of mobility, rather than simple permeability. Mobility, m, is defined as follows:m

=

k

µ

.................................................................................. (17.6)

where k is the permeability of the formation to the produced fluid and µ is the viscosity of that fluid at reservoir conditions. This allows us to see that a tight oil formation has considerably greater permeability than a tight gas formation. In the case of a tight formation – especially a tight gas formation – it is relatively easy to produce a fracture of essentially infinite conductivity (i.e. a fracture so conductive that any further increase in fracture conductivity produces no subsequent increase in production). In such a situation, the factor limiting the potential production increase is the ability of the formation to deliver hydrocarbons to the fracture. This is controlled by the permeability of the formation and by the inflow area of the fracture. Obviously, increasing the permeability of the

Page 168


BJ Services’ Frac Manual 17. Designing the Treatment entire formation is beyond the abilities of stimulation engineering. However, maximising inflow area – by increasing the size of the fracture faces – is relatively easy. Therefore, fractures in tight formations are designed to produce maximum size, with a minimum necessary proppant concentration. Of course, there are diminishing returns on increasing fracture size – doubling the fracture length will increase the fracture area by approximately 4 times (as the height will increase at the same relative rate as the length). This means that the proppant volume (which is spread over the entire area of the fracture) is also increased by a factor of 4. As the fracture height increases, an increasingly greater proportion of the fracture will be outside the zone of interest (unless a “massive” formation is being fractured). Therefore, an increasing proportion of the proppant will be placed out of zone (i.e. it is wasted). In addition, the fluid volume required will increase by between 4 to 8 times. Thus, doubling the length – which at best can only double the production – will can increase the cost of the treatment by 4 to 6 times. Whilst fracture conductivity is not the most important consideration for tight formation fracturing, it is important to remember that some fracture conductivity is required. Remember that the proppant pack will lose permeability due to factors like residual polymers, non-Darcy flow and multi-phase flow, and also that the pack may lose permeability as the reservoir pressure depletes (i.e. as the closure pressure increases). Therefore, when designing a tight formation fracture treatment, it is important to carefully define the minimum fracture conductivity, and to ensure that the produced fracture always remains above this. Tight formations – especially tight gas formations – tend to have the following characteristics:i) ii) iii)

Low permeability and hence low fluid leakoff High Young’s modulus and hence; Low fracture toughness

Because of the often extremely low fluid leakoff, it is possible to treat these formations with very low pad volumes. Often, it is not necessary to crosslink the pad and a linear gel is used (a "hybrid" frac). In some formations, it is even possible to frac without any pad whatsoever – the formation can be fractured with the first slurry stage. Because of the very low fluid leakoff, these fractures can take a long time to close after the treatment is finished (work by Cleary et al suggests that some fractures may take 24 hours to close). Therefore, it is important to design the fracturing fluid with very good proppant transport characteristics, so that it is capable of supporting the proppant for as long as it takes the fracture to close. Another major issue for tight formations, especially tight gas formations, is fluid recovery. In many cases, extra care and attention must be paid to the design of the fracturing fluid to ensure that it does not form fluid blocks in the formation. This is usually done by adding surfactants to reduce the surface tension of the fluid system. It is also important to break the fluid to as low a viscosity as possible. Dry gas reservoirs may be sensitive to fluids or any type – water or hydrocarbon. These can cause extensive damage due to changes in relative permeability. In such formations, it is common practice to perform treatments using N2 or CO2 foams (or with binary foams), to reduce the liquid content to a minimum. Alternatively, it is also common practice to treat dry gas wells with methanol-based fluids, as these are very easily recovered after the treatment. Tight gas fracturing is probably the single most common form of hydraulic fracturing. In many areas of the world, tight gas reservoirs can only be produced economically because of hydraulic fracturing. In these places, fracturing has become the accepted method of completing wells and whole reservoirs are developed using this technique.

Page 169



17.6

Designing for Injection Wells

Injection wells are basically fractured in the same way as production wells, although there are a number of minor points which must be observed:i) ii) iii) iv) v) v)

vi)

17.7

Take careful note of the closure pressure after the treatment. When the well is placed back on injection, on no account must this pressure be exceeded, as this will open up the fracture and (potentially) allow the proppant to fall downwards. Remember that the build-up of pressure in the near wellbore area (caused by the injection) may act to reduce the local closure pressure. Clean up the well after the fracture as much as possible before placing the well on injection. Any polymer residue or proppant fines left in the proppant pack will act to block the formation permeability and will not be produced back from the formation. When fracturing existing injection wells, fluid leakoff will often be much higher than in offset producing wells, due to the higher than normal water saturation of the formation. Do not use surfactants that leave the formation water-wet. These will act to reduce the injectivity of the water. When fracturing a new well, remember that water injection – and the control of where the water goes – is an important part of reservoir management. Select the zone to be fractured carefully and always in consultation with the Reservoir Engineer. Be aware of the consequences of fracturing into high permeability and/or low pressure formations. Consider using polymer-free fracturing fluids (e.g. visco-elastic fluids or brine with LiteProp). Such fluids have very low permeability once broken and no polymer residues. Consequently, they do not have to be flowed back - simply place the well back onto water injection once the fluid has broken

Designing CBM Treatments

The vast majority of the coal bed methane fracturing that takes place in the US in 9 or 10 major basins in the US, Australia and in China. In addition, CBM fracturing also takes place in a number of locations, including the UK, the Middle East and Russia. In all of these places, each particular coal field or basin tends to be dominated by a single operating company. Each of these basins has its own particular characteristics, in terms of the age and maturity of the coal, the reservoir pressure, the fines mobility, the water production and the mechanical characteristics of the coal seams and their surrounding rock layers. As a result of this, each operating company has developed its own particular method for producing the gas, and when this involves fracturing, they have developed their own method for this as well. CBM fracturing remains to this day very difficult to simulate on a computer. Conventional models cannot be applied to the coal, due to the extensive cleat systems that exist in the seams, the extremely plastic nature of the coal and the shear decoupling that exists between the coal and the over- and under-lying rock strata. Without the aid of reliable fracture models, Engineers have developed a number of “rules of thumb” for CBM fracturing, most of which are specific to a particular basin. In short, operating companies that are successfully producing coal bed methane, are those which have been prepared to experiment, to try out a few different methods and to except a few failures along the way.

Completions i)

There are many different completions being used, from open hole to multiple perforated monobores. There is very little agreement over which is ideal, although a cemented and perforated completion is best when fracturing is being considered.

Page 170


BJ Services’ Frac Manual 17. Designing the Treatment ii) iii)

iv) v)

Formations tend to be several thin seams, rather than a single large seam. As such, most completions contain several sets of perforations. In such cases, it is essential that each set of perforations is broken down before the fracturing operation. This usually involves using a straddle packer (or packer and bridge plug) positioned over each set of perforations in turn. The breakdown is achieved by pumping small quantities of acid (usually formic) into the zone. Some companies prefer to do this with water. The breakdown can also be achieved using ball sealers, but this is less reliable. Without the breakdown of each individual zone, it is likely that most of the perforated intervals will receive no stimulation during a treatment, whilst the other zones will receive everything.

Fluid Systems i) ii)

iii)

iv)

All sorts of different fluid systems are still being used, including foams, fresh water, slick water and crosslinked gels. Slick water and fresh water have the advantages that they are very cheap and potentially non-damaging to the coal seam. Their major disadvantage is that their low viscosity makes it difficult to carry proppant deep into the cleat system. This can also lead to pre-mature screenouts. However, neutral density proppants could potentially revolutionise this type of treatment, although the cost of the proppant may be uneconomic. Foams have good proppant transport characteristics, and are very good for placing the proppant in the wider cleats and not in the narrower channels. Foam is also very good for unloading the well after the treatment. However, foam is very expensive to use, requiring a lot of additional specialised equipment on location for the treatment. The best fluid system to start out with seems to be a cheap, reasonably low polymer loading crosslinked borate guar or guar derivative. This is a standard water-based fracturing fluid, reduced to the minimum necessary to carry proppant into the cleat system. Polymer loading would be 25 to 30 lbs/mgal. It is essential that an enzyme breaker be used, as it has been shown that oxidizing breakers can seriously damage the cleat faces.

Proppant Selection i)

ii) iii) iv)

Generally, it is best to pump as large a proppant grain size as possible. This is for two main reasons: First, the larger the proppant grain, the higher the proppant permeability and the less susceptible the proppant is to embedment in the cleat faces; Second, the larger proppant grains allow the coal fines to past through, rather than collect and gradually plug up the conductivity. The recommended proppant size is 12/20 Sand, although sometimes this can be hard to obtain. Some operators like to pump a fine grain sand (such as 100 mesh) in the early stages of the treatment (in the pad). The purpose of this is to block up the narrow cleats, and force the fracture and the larger main proppant grains into the wider cleats. Proppant volume ranges from 3,000 to 10,000 lbs per vertical ft of net height. Some operators claim to be able to place 15,000 lbs/ft, but this is not confirmed. A good starting point is to aim to place 5,000 lbs per vertical ft of coal. If this is placed without any problems, the proppant volume can be gradually increased on subsequent treatments.

Fracture Geometry i) ii)

Although it is very difficult to predict the geometry of the fracture(s), it is still possible to divide the fractures into two main regimes. The first regime occurs when the fracture penetrates up and down into the over- and under-lying rock strata. Fractures tend to have an overall radial or elliptical geometry,

Page 171



iii)

although the actual shape of the fracture(s) within the coal seam will be very complex. This regime is characterised by a moderate to low frac gradient (0.5 to 0.7 psi/ft), as the fracture is pushing against the minimum horizontal stress. The second regime occurs when the coal seam shears relative to the over- and under-lying rock strata, so that the fracture does not penetrate out of the coal seam. This results in the famous T and I –shaped fractures, where the fracture grows horizontally between the coal seam and the confining rock strata. This regime produces better stimulation, as all of the proppant is placed in the coal seam. However, pressures tend to be much higher, with the frac gradient being 1.0 psi/ft or greater, as the fracture has to lift the overburden in order to propagate. Also, it is important not to confuse the high pressures of this type of frac, with the high pressures produced by near wellbore friction (see below).

Notes on Job Design i)

ii)

iii)

iv) v) vi)

17.8

In addition to breaking down each set of perforations individually, it is also worthwhile performing a full-scale minifrac. This involves pumping into the formation at the anticipated treatment rate, using the actual treating fluid but no proppant. This allows the frac engineer to assess the overall fracture geometry (from the frac gradient) and the level of near wellbore friction (from the difference between the BHTP and the ISIP). Surface facilities should be designed to cope with fines production. All treatments, regardless of the fluid used and the additives mixed into the fluid, will cause the production of coal fines. These fines should be produced back to the surface and handled there. If they are not produced back to surface, they will block up the proppant pack and cause a loss in production that will increase with time. Treatment pump rate should be 1.0 to 1.5 bpm per vertical ft of coal. Treatments using fresh or slick water are usually pumped at higher rates. This is because the fluid has no proppant transport characteristics, and so it is essential to keep the proppant moving within the cleat system. Pad volume should be 20 to 25% of the overall treatment volume, although this is an area that varies considerably – some treatments use only 5% or even less. Proppant concentration should be 6 to 8 ppg (lbs per gal) for the crosslinked fluid and foam. Slick and fresh water systems are only capable of carrying proppant up to about 2 ppg. If significant near wellbore friction is present, then it is likely that 6 to 8 ppg will cause a premature screenout. If this friction is detected, the maximum proppant concentration should be reduced to 4 ppg. If this happens, more fluid will be required to place the same volume of proppant.

Designing for Coiled Tubing Fracturing

Coiled tubing fracturing is really a method for placing the fracture treatment, rather than a specific method of treating a type of formation. Any of the types of treatment previously described can be placed with coiled tubing. The advantages of using coiled tubing have already been explained in Section 3.6. To summarize, they are as follows:i) ii) iii) iv)

Isolation of completion. Isolation of individual zones. Rapid turn around between multiple treatments. Use of the coil to gas lift the well back to production.

The main problem with fracturing through CT is the narrow diameter of the tubing itself. This means that the most important factors in designing CT fracs are the friction pressure of the fracturing fluid and the maximum allowable pressure that can be imposed on the CT. Page 172


BJ Services’ Frac Manual 17. Designing the Treatment BJ’s Circa Coiled Tubing Simulator can be used to predict the maximum allowable injection pressure, for any given CT string. However, it must be remembered that the CT string will be static when the treatments are pumped. The maximum injection pressure usually generated by CT simulators assumes that the CT is either moving in or moving out of the hole. This means that the CT is being continuously plastically deformed, as the internal pressure is applied. However, if the CT is static – and hence it is not being plastically deformed – the CT will be able to withstand much greater internal pressures. For instance, normal maximum injection pressures for CT are in the region of 5,000 to 6,000 psi. However, during static fracturing operations, treatments have been pumped at pressures up to 13,000 psi. However, in spite of this, the friction pressure of the frac fluid (and the subsequent surface treating pressure it produces) will still dominate the design of the treatment. It is often necessary to use very low friction pressure fluids (i.e. low polymer loading gels or viscoelastic surfactant-based fluids) in order to be able to maintain the desired rate. These fluids are often significantly more expensive than their conventional alternatives. Notwithstanding the increased allowable internal pressure and a possibly reduced friction pressure, even with large diameter CT strings (2” or greater), the Frac Engineer will still be rate limited to between 5 and 12 bpm. This can often significantly limit the size of treatment that can be placed in the formation. Obviously, the shorter the string, and the larger the ID, the greater the maximum rate. However, it should also be noted that generally with CT, the larger the ID, the smaller the maximum allowable pressure (unless so-called heavy-walled CT is used). Therefore, the Frac Engineer has to balance the need for rate against the desire for a cheap fluid and the maximum allowable injection pressure. Usually, the requirements of the treatment take precedence over the cost of the fluid, allowing the Frac Engineer more freedom to design a suitable pumping schedule. CT fracturing has found niche applications in a number of areas, most notably southern Alberta. However, it remains economically viable only in areas where there are relatively shallow multi-zone formations, and where the cost of a workover is expensive.

17.9

Unified Fracture Design and Proppant Number

In 2002, Economides, Oligney and Valkó, published their principles of Unified Fracture Design, and introduced the concept of a dimensionless proppant number, or Np. This was defined as follows:Np

=

Np

=

2xf re

π

CfD (radial flow system) ....................................... (17.7)

2xf 2 xe CfD (square reservoir, area = xe ) ............................ (17.8)

Rearranging and substituting in Equation 10.1 gives the following result, for a radial flow system:Np

=

2kp w ¯ ........................................................................... (17.9) re k π

According to the theory, for each value of Np there is a corresponding optimum value of CfD, which produces the maximum production increase. Therefore, this theory allows the Frac Engineer, for any given reservoir and proppant combination, the optimum balance between average proppant width (w ¯ ) and proppant fracture half length (xf), as illustrated in Figure 17.9a.

Page 173


BJ Services’ Frac Manual 17. Designing the Treatment Figure 17.9a also shows us that for Np values less than 0.1 the optimum value of CfD is 1.6. This gives the Frac Engineer a very powerful tool – for medium to high permeability fracturing (i.e. Np < 0.1) the fracture should always be designed for CfD = 1.6. For low permeability fracturing, the relationship is not so simplistic and specific values for Np have to be calculated for each proppant-reservoir-fracture combination. As can be seen from Equation 17.9, proppant number varies inversely with formation permeability. It can therefore be thought of as a measure of the effectiveness of the proppant, as a transport medium, relative to the formation. Whilst very low values of Np are easy to obtain, in practice it is hard to get values higher than 10 (as re is limited). Under most circumstances, the changeover from Np < 0.1 to Np > 0.1 occurs in the range of 0.5 to 5 mD formation permeability. Obviously, the exact value is highly dependent upon the effective proppant permeability (allowing for the effects of multi-phase and non-Darcy flow).

Dimensionless Fracture Conductivity, CfD

100

10

C fD = 1.6 for N p < 0.1

1

Medium to High Permeability

0.1 0.0001

0.001

0.01

Low Permeability

0.1

1

10

100

Proppant Number, N p

Figure 17.9a – Optimum dimensionless fracture conductivity against dimensionless proppant number (after Economides et al, 2002).

17.10 Net Present Value Analysis Net Present Value (NPV) analysis is a method for comparing one treatment to another, on a cost basis, to determine which treatment is the most cost effective. It allows one treatment to be compared to another on economic grounds. NPV takes into account the cost of the treatment, the revenue generated and the customer’s requirements. When comparing treatments, the option that produces the greatest NPV should be selected. Within the constraints of equipment, materials, completion and cost, the Frac Engineer should design for maximum NPV. A more detailed explanation of NPV analysis, together with an example, is contained in Section 13.1.

Page 174



References Economides, M.J., and Nolte, K.G.: Reservoir Stimulation, Schlumberger Educational Services, 1987. Gidley, J.L., et al: Recent Advances in Hydraulic Fracturing, Monograph Series Vol 12, SPE, Richardson, Texas (1989). Bradley, H.B. (Ed): Petroleum Engineers Handbook, SPE, Richardson, Texas (1987) Jiang, T., Shan W.W., Ding, Y.H., Wang, Y.H. and Wang, Y.L.: “Systematic Fracturing Technology and its Application in Development of Low Permeability Reservoir”, SPE 50910, presented at the SPE international Conference and Exhibition in China, Beijing, China, November 1998. Phillips, A.M. and Anderson, R.W.: “Use of Proppant Selection Models To Optimize Fracturing Treatment Designs In Low-Permeability Reservoirs”, SPE/DOE 13855, presented at the SPE/DOE 1985 Low Permeability Gas Reservoirs, Denver, Colorado, May 1985. Voneiff, V.W., and Holditch, S.A.: “A Economic Assessment of Applying Recent Advances in th Fracturing Technology to Six Tight Gas Formations”, SPE 24888, presented at the 67 Annual Technical Conference and Exhibition, Washington, DC, October 1992. Yong Fan, and Economides, M.J.: “Fracture Dimensions in Frac&pack Stimulation”, SPE 30469, presented at the SPE Annual technical Conference and Exhibition, Dallas, Texas, October 1995. Rae, P., Martin, A.N., and Sinanan, B.: “Skin Bypass Fracs: Proof that Size is Not Important”, SPE 56473, presented at the SPE Annual Technical Conference and Exhibition, Houston, October 1999. O’Driscoll, K.: Middle-East Region Coal Bed Methane Fracturing Manual, BJ Services, 1995. Gavin, W.G.: “Fracturing Through Coiled Tubing – Recent Developments and Case Histories”, SPE 60690, presented at the 2000 SPE/ICoTA Coiled Tubing Roundtable, Houston, April 2000. Cramer, D.D.: “The Unique Aspects of Fracturing Western US Coal-beds”, SPE 21592, presented at the Petroleum Society of CIM/Society of Petroleum Engineers International Technical Meeting, June 10-13, 1992, Calgary, Alberta, Canada. Nimerick, K.H., et al: “Design and Evaluation of Stimulation and Workover Treatments in Coal Seam Reservoirs”, SPE 23455, presented at the Petroleum Society of CIM/Society of Petroleum Engineers International Technical Meeting, June 10-13, 1990, Calgary, Alberta, Canada. Archer, J.S. and Wall, C.G.: Petroleum Engineering – Principles and Practices, Graham and Trotman, London (1986). Wong, G.K., Fors, R.R., Casassa, J.S., Hite, R.H., and Shlyapobersky, J.: “Design, Execution and Evaluation of Frac and Pack (F and P) Treatments in Unconsolidated Sand Formations in the Gulf of Mexico”, SPE 26563, presented at the SPE Annual Technical Conference and Exhibition, Houston TX, Oct 1993. Tiner, R.L., Ely, J.W. and Schraufnagel, R.: “Frac Packs – State of the Art”, SPE 36456, presented at the SPE Annual Technical Conference and Exhibition, Denver CO, Oct 1996. Economides, M.J., Oligney, R.E. & Valkó, P.P.: Unified Fracture Design, Orsa Press, Alvin, TX, 2002.

Page 175


BJ Services’ Frac Manual 18. Real-Time Monitoring & On-Site Redesign

18

Real-Time Monitoring and On-Site Redesign

Thanks to the advent of electronic data gathering systems, personal computers and efficient, reliable fracture simulators, it is now possible to actually model the fracture as the treatment progresses. This process, known as Real-Time Monitoring, allows the Frac Engineer to actually re-design the treatment on-the-fly. The more traditional form of on-site redesign is when data from a step rate test and/or minifrac is used to redesign the main treatment. This is usually carried out on-site, with the whole of the frac spread and frac crew waiting for the Frac Engineer to produce the new frac design.

18.1

Real-Time Data Gathering Pressure Transducers

Voltage

Bottom Hole Pressure Data

Flowmeters Frequency

3600 or Isoplex

ASCII Data

JobMaster

Selected ASCII Data

Frequency

Frac Model

Nuclear Densometers

Redesigned Treatment Schedule

Frac Engineer

Figure 18.1a – Process loop for real-time fracture modeling and redesign

With a modern frac spread, it is now possible to measure, record and monitor every single treatment parameter, including items such as liquid additive rates, sand screw rpm’s and annulus pressure. However, for the Frac Engineer, there are three main variables which are required:- bottom hole pressure, proppant concentration and slurry rate. It is useful and often necessary for a whole range of data to be recorded during the treatment, but it is only these three variables which will be needed for the redesign. Often it is useful for the Frac Engineer to use surface treating pressure, in order to calculate BHTP or pipe friction data. Also, is it sometimes quite helpful for the Frac Engineer to record stage number, in order to keep track of which stage is at the perforations (especially if there are tortuosity problems). Data is recorded using three basic types of measuring devices; pressure transducers, nuclear densometers and flow meters (as illustrated in Figure 18.1a).

Page 176



Flow Meters Flow Meters come in three main types:- turbine, magnetic and inertial or mass flow meters. Turbine flow meters are the most commonly used, as they are easy to employ, require no external power supply and are cheap. They can also be used in high pressure flow lines. Fluid flow is measured by a single turbine, which is positioned in the centre of the flow stream. This turbine rotates when fluid passes through the flowmeter. The faster the fluid flow, the faster the turbine spins. A magnetic pick up measures how fast the turbine is rotating, sending an output in the form of a frequency to the control centre. The turbine flowmeter has several disadvantages. It is easily obstructed or damaged by debris in the frac fluid. This means that the flowmeter needs to be checked and potentially redressed after every treatment. The turbine flowmeter requires a separate calibration factor for each different fluid type (i.e. linear gels, gelled acids, gelled oils etc). The turbine flowmeter also has a relatively high flow rate threshold, below which the turbine will not rotate. This means that for 2 or 3 different turbine flow meters are usually required for measuring over a wide range of flow rates. Turbine flow meters can only be used for liquids. Magnetic flow meters (often referred to as mag flowmeters) rely of the physics of generating electrical current. This states that it you have motion and a magnetic field, then you will get current flow, provided there is a conductive path. The magnetic flow meter provides the magnetic field, the fluid provides the motion and a current is generated. The magnitude of the current is proportional to the flow rate. These flow meters are easy to use (once they have been set up) and very reliable, requiring little maintenance (they have no moving parts or restrictions). The main disadvantages of these flow meters are that they require an external power source, they are expensive and they can only be used for measuring conductive fluids (so they cannot be used for measuring gelled hydrocarbons or gases). Inertial or mass flow meters (such as the MicroMotion flowmeter) work by using two flow loops. As the fluid enters the flow meters, it is split into two loops of equal diameter. One loop measures the density of the fluid, whilst the other loop measures the mass flow rate. Volumetric flow rate is obtained by dividing the mass flow rate by the density. Density is measured by forcing the flow around a loop that is vibrating. This vibration is produced by a calibrated agitation system, which always provides the same force, at the same frequency. A measuring system compares the known “input” agitation with the vibration of the flow loop. Generally, as the mass of the flow loop + fluid increases, the frequency with which the loop vibrates will slow down. As the mass and volume of the flow loop is known, the density of the fluid can be quickly calculated. Mass flow rate is measured by the second flow loop. This loop is offset slightly from the main direction of flow, so that the inertia of the fluid as it flows causes the loop to twist slightly. The amount of twist is measured by a number of strain gauges placed along the flow loop. The force causing the flow loop to twist (and hence the reading on the strain gauges) is directly proportional to the mass flow rate. This type of flow meter has several advantages. Most types of fluids can be measured by this method, including gases, hydrocarbons and cryogenic fluids. The flow meter can also be used to output density, eliminating the need for a separate densometer. If the pressure differential across the flow meter is carefully measured, the apparent viscosity of the fluid can also be obtained. Unfortunately, this flowmeter also has several disadvantages. Because the fluid flow is forced around two flow loops, it cannot be used for abrasive fluids (the flow loops are quickly abraded until they fail). These flow meters are quite large and heavy. They are expensive. Finally, because of the sensitive measuring apparatus inside the flow meter, these devices are also quite fragile.

Page 177


BJ Services’ Frac Manual 18. Real-Time Monitoring & On-Site Redesign The above listed three methods are all direct methods of measurement. However, it should also be noted that flow rate is often measured indirectly, by reading the rpm' s of an input shaft powering a pump (often referred to as a "stroke counter"). Basically, the computer reading the rpm' s has a calibration factor which converts rpm' s to flow rate, a quickly and easily calculation. Whilst these flowmeters are very easy to use and also very mechanically reliable, they suffer from 2 main drawbacks. First, no pump is 100% efficient, and so the stroke counter has to be calibrated to allow for this. Second, if the pump loses prime (or doesn' t have prime to start with), then the stroke counter will give a false reading. It is therefore advisable to use a direct flow rate measurement as the primary source of flow rate measurement, using stroke counters only as a back up.

Nuclear Densometers Nuclear densometers (or densimeters, or densiometers) all work on the same basic principal. A radioactive source is held on one side of the flow stream, whilst a detector on the opposite side of the flow stream measures the radioactivity that passes through the flow stream, in counts per second. Basically, the higher the density of the fluid, the lower the number of counts per second. Nuclear densometers vary in the type of output they provide. The basic densometer has no data processing capabilities, and outputs a frequency signal (the same frequency as the number of counts per second being received by the detector). A separate data processing facility (such as a PC) is required to turn the basic data into a density or a proppant concentration. It is this type of densometer that is most commonly used in the fracturing industry. More sophisticated densometers come complete with data processing, and can output density, SG, proppant concentration or acid %. 137

The radioactive source used in the densometer is usually Caesium 137 (or Ce). This metal is a medium energy beta and gamma radiation emitter, with a half-life of 30 years. This means that the radioactive source gradually gets weaker with time – after 30 years it is only half as radioactive as it initially was. Consequently, all radioactive densometers have to be regularly calibrated, to allow for the fact that the source is gradually producing less and less radiation. Therefore, the data processing facility (usually a PC, an Isoplex or a 3600, but sometimes also a box on the side of the densometer) has to have this calibration installed, in order that density can be output. Proppant concentration is easily calculated from the overall bulk density of the fluid, using the following formula:PC =

(ρsl - ρgel)

(1 - [ρsl/ρp])

................................................................... (18.1)

where PC is the proppant concentration in ppa (see below), ρsl is the slurry density in lbs per gallon (ppg), ρgel is the base fluid (usually gel) density in ppg and ρp is the proppant density, also in ppg. Proppant density is often also quoted as an absolute volume in gals/lb. This is simply the reciprocal of the density in ppg. Proppant concentration is measured in ppa or Pounds of Proppant Added. This is the number of pounds of proppant that have been added to 1 gallon of clean base fluid (which is how the blender adds the proppant – it measures the clean flow rate in gallons per minute, and calculates how many lbs per minute of proppant need to be added). Sometimes, proppant concentrations are also quoted in ppg – meaning pounds of proppant per gallon of clean fluid. The use of these units should be avoided for proppant concentrations, as they can get easily confused with fluid or slurry densities.

Pressure Transducers Pressure transducers are the simplest of the measuring devices used in fracturing. The transducer is consists of a strain gauge, that is mounted so that as pressure is applied, the Page 178


BJ Services’ Frac Manual 18. Real-Time Monitoring & On-Site Redesign strain gauge is compressed. As the strain gauge is compressed, its electrical resistance will increase slightly. The higher the applied pressure, the greater the increase in resistance. The pressure transducer is connected, via a transducer cable, to a special measuring circuit known as a Wheatstone’s Bridge. This is an electrical circuit consisting of three known electrical resistances and an unknown electrical resistance (the transducer + cable). Because of the nature of the circuit, if the potential difference (or voltage drop) across the bridge circuit is known, the values of the three known resistances can be used to calculate the value of the unknown resistance, to a high degree of precision. Therefore, if the resistance of the cable is known, the resistance of the pressure transducer can be obtained. So in order to measure the pressure applied to a transducer, the voltage drop must be measured. Pressure transducers are regularly calibrated by applying a known pressure to them, usually via a dead weight tester. This calibration produces a relationship between resistance and pressure, so that if the resistance is known, the pressure can be quickly obtained. Because a large increase in pressure produces only a relatively small change in electrical resistance, it is important to have good quality cables that are well looked after (as the circuit measures the resistance of the cable at the same time). This also means that the cables must be of a fixed length, producing a limit to how far the control cabin can be away from the pressure transducer. Transducer cables cannot be spliced, repaired or re-used if they are damaged.

Processing the Data Raw data from the transducers, flow meters and densometers is not usable by the monitoring computers. It has to be converted to a digital form by an analogue to digital converter. Once the data has passed through this, it can be processed to give the actual treatment parameters. For instance, the turbine flowmeter is actually measuring the number of times a turbine blade passes the magnetic pick up, rather than a volumetric flow rate. Every time a blade passes the pick up, electrical current is generated, reaching a peak as the blade is directly opposite the pick up. As the blade moves away from pick-up, the current drops off. This means that the output from a turbine flow meter is cyclic – the higher the frequency of these cycles, the faster the turbine blades are rotating and the faster the fluid is flowing. The cyclic analogue input is the converted to a digital output, by the analogue to digital converter. A digital output simply means that the converter is sending the computer a number – in this case the number of cycles per second, or frequency. As stated, the output from the analogue to digital converter is passed on to a computer for processing. This computer can be a PC, an Isoplex or a 3600. Whatever form it comes in, the processing computer converts (in the case of the turbine flow meter) a number of cycles per second, into a flow rate, by applying a calibration. This calibration is user input, and will vary according to the type and size of flowmeter, and the fluid being used. For pressure transducers, the analogue to digital converter measures the voltage across the bridge circuit, and outputs this as a number. For the nuclear densometer, a similar process is carried out as for the turbine flow meter, in which a frequency is converted into a number. The processing computer will contain a calibration algorithm for each of these devices, converting the numerical output from the analogue to digital converter, into psi or ppa as appropriate.

Displaying and Analysing the Data Usually the Frac Operator or an electronics technician will run the JobMaster computer. This computer will display all the parameters being monitored by the system, and is the primary source of information for the person actually running the treatment. The Frac Operator usually has the option to run several different displays, so that unprocessed data can be displayed (such as real-time pressure, rate and proppant concentration), along with parameters that Page 179


BJ Services’ Frac Manual 18. Real-Time Monitoring & On-Site Redesign have been processed on the fly, such as calculated BHTP, cumulative volumes and of course the Nolte plot. The Frac Engineer usually operates the second computer. This machine receives selected data from the first computer, almost always in ASCII format. The Frac Engineer will use a specialised treatment monitoring programme or fracture simulator to display and analyse this data. The Frac Engineer’s computer is usually capable of receiving ASCII data from more than one source. The primary source of data will almost always be the JobMaster computer. This data usually comes in via the COM 1 serial port. However, if a second COM port is fitted to the Engineer’s computer, it is possible to receive data from a second source (such as a bottom hole pressure gauge) and merge it with the primary data, real time, so that it can be displayed and analysed. Computers fitted with USB ports can use adapters to allow several COM ports to be used simultaneously. During the treatment, most of the people in the control centre will be watching the displays controlled by the JobMaster computer. Numerical Display

JobMaster Displays

Figure 18.1b – Inside of a typical frac control van, showing the numerical display and some of the displays being run by JobMaster.

Remote Data Transmission Remote data transmission is a specialised service which allows the customer and the Frac Engineer to remain in the office, whilst the treatment is carried out. Provided there is someone on location to look after the fluids, run JobMaster and handle the data transmission process (this will usually be a junior Engineer), then the only reason that the senior Engineers are required on location is for data analysis. If the data can be transmitted to a separate location, then the customer and service company Engineers do not have to be on location. They can remain back at the office. This has particular advantages when the treatment is being carried out in remote locations (such as offshore). Instead of the Engineers being tied up for (sometimes) several days, remote data transmission means that they are only directly involved in the treatment for a few hours – the time taken for the step rate test and minifrac to be pumped, for the redesign to be carried out in the office and for the main treatment to be performed. Page 180


BJ Services’ Frac Manual 18. Real-Time Monitoring & On-Site Redesign On Location Data Modem

Office Satellite or Cellular Phone Line

Data Modem

Data Link

Satellite or Cellular Phone Line

Voice Link

Figure 18.1c – Remote data transmission schematic

Data transmission is carried out real time, using software packages like JobMaster, which have been specifically designed to carry out this process as part of its capabilities. Both the transmitting and receiving computers run JobMaster, coordinated so that receiving computer is expecting the same channels that the transmitting computer is sending. The voice link is an essential part of the process, so that the data link can be properly coordinated and also so that the on-site Engineer and keep the office-based Engineer’s fully informed of developments and as they happen. Thanks to modern communications, it is a relatively easy task to transmit the data real time. Data transmission is usually pretty reliable, but interruptions can sometimes happen. In this case, the software package should be set up so that transmission can be easily resumed, and that data that is not transmitted during the break in communications is stored for transmission as soon as communication is re-established. The latest versions of the remote data transmission systems actually use internet-based communications. Each control cabin or frac van has it' s own web address, and broadcasts the treatment onto the internet. Anyone with the job-specific password can log onto to monitor a treatment, from any computer that has internet access. It is also useful to have a separate file transfer programme or internet access for e-mail installed on both computers, allowing quick and easy transmission of data files between computers.

18.2

On Site Redesign

On site redesign is the science and art of redesigning a fracture treatment after the step rate test and minifrac. Usually it is done on location (or - via remote data transition – back in the office) whilst the frac crew and equipment are waiting. Consequently, there is usually a reasonable amount of pressure on the Frac Engineer during this process, which may take several hours. For example, in the offshore environment, were the rig may be costing over $300,000 per day, every hour spent redesigning (which is usually down time for the rig), costs the customer over $12,000. However, this down time is usually money well spent. The Frac Engineer should take as much time as is necessary and must not produce a hasty, poorly designed treatment. The object of the fracturing exercise is to maximise production increase, after all. Minimising rig time is obviously highly desirable, but it should not take precedence over the main objective.

Page 181



Minifrac Analyser Fracture Simulator Closure ISIP Pressure Match Fracture Model

CONVENTIONAL ANALYSIS

PRESSURE MATCHING

Raw Data

Formation Properties

Revised Treatment Schedule

Required Fracture Properties

Materials & Equipment on Location

Fracture Simulator

Fracture Meets Requirements?

No

Yes

Customer Approval?

No

Yes

Final Treatment Schedule

Figure 18.2a – On-site redesign process flowchart

Page 182


BJ Services’ Frac Manual 18. Real-Time Monitoring & On-Site Redesign One way to minimise down time is to restrict the number of Engineer’s involved in the redesign process. It is not exaggeration to state that the time taken for the redesign is proportional to the square of the number of Engineer’s involved. On site redesign is something that improves with practice and experience. Figure 18.2a shows a process flow chart for the redesign process. The process starts with the collection of the raw data. This includes not only the minifrac and step rate data (collected either real-time or from separate files), but also other items such as wireline logs, completion diagrams, tracer surveys, temperature logs, BHTP gauge data, data from previous fracs on offset wells and so on. Once all this has been collected, the Frac Engineer can start to analyse the data from the step rate test(s) (see Section 15) and the minifrac (see Section 16). As discussed previously, this process can often take some time and can sometimes be carried out under quite stressful conditions. Nevertheless, once this process has been completed, the Frac Engineer should have been able to tune the fracture model, so that what is in the computer is a reasonable representation of what is in the formation. Once this has been achieved, the hardest part of the redesign process has been completed. However, the Frac Engineer still has to produce the final treatment design. In order to do this, the Engineer has to design a treatment schedule based on two important parameters:1.

The objectives of the treatment. Usually, the objective of the treatment is to place a frac in the formation, with a certain geometry, and relative conductivity. These objectives are usually set up before arriving on location. Usually, these objectives will remain unchanged after the calibration tests (step rate test and minifrac). However, the results of these tests may change the specifics of how this is achieved. For instance, if the minifrac shows the permeability of the formation to be significantly different from that anticipated, the optimum fracture geometry will have to be altered in order to meet the CfD requirements.

2.

The available equipment and materials. Usually, the Frac Engineer has to work within the limitations for the equipment available for the treatment, in terms of tank volumes, maximum pumping rates etc., so that the Engineer is producing the optimum treatment design the frac spread is capable of pumping. In remote locations (where materials cannot be “hot-shotted” out to location), the Engineer can also be restricted by the quantity of materials available on location (volume of gel that can be mixed and the volume of proppant).

Working within these restrictions (and also remembering the maximum allowable pumping pressure), the Frac Engineer must produce the optimum possible frac design. This is not just a question of producing a production increase – for a lot formations, this is relatively easy to do. The Frac Engineer must also maximise the production increase, to meet or exceed the economic criteria for the treatment, as there is usually a significant cost associated with fracturing, and a small production increase may not be sufficient.

18.3

Real-Time Fracture Modeling

Some fracture simulators, such as MFrac, FracPro and FracproPT, have a facility that enables the fracture to be modeled real-time. This is a very powerful tool that - under the right conditions - can enable the Frac Engineer to redesign the main treatment on-the-fly, as it is being pumped. The modeling computer is set up to receive data from either the data processing computer or the Frac Engineer’s computer, usually in ASCII format. The user then runs the fracture model, selecting the “real-time data input” option. The user enters the relevant formation data and treatment schedule, which can be loaded from a previously created data file. The treatment starts, and the computer starts to collect the data. As the treatment progresses, the simulator models the created fracture. The model will take fluid, proppant, formation and wellbore characteristics from the input model, and will take the pump rate, pressure and proppant Page 183


BJ Services’ Frac Manual 18. Real-Time Monitoring & On-Site Redesign concentration from the real time data. Using this data, the simulator will model the fracture that has already been created, constantly updating as more data is collected. This enables the user to perform two separate operations:1. 2.

The Frac Engineer can perform a pressure match with the data that has already been collected by the simulator, until the net pressure predicted by the computer matches up with the actual net pressure. The Frac Engineer can instruct the simulator to run the job until completion, predicting the characteristics of the fracture, based on the ongoing pressure match. For the treatment schedule, the simulator will use the actual treatment data as far as possible, and then project forward until the end of the job using the remaining input treatment schedule. This allows the Engineer to predict the fracture characteristics, based on the most accurate data possible. This process can be taken one step further, as the Engineer can alter the remaining treatment schedule, and predict the fracture characteristics based on this revised schedule. Thus the Engineer can redesign the treatment schedule on-the-fly. This capability is limited to FracPro and FracproPT.

Limitations of Real-Time Modeling This ability to redesign on the fly – whilst usually not very popular with the frac crew – is a very powerful tool, provided the Frac Engineer is aware of the following:1.

2. 3.

4.

Do not over-react to short term trends. All fracture simulators treat formations as homogenous materials with uniform rock mechanical properties throughout. In reality this is usually not the case. The fracture is constantly propagating through rock with varying properties, producing unpredictable variations in the net pressure plot. In fact, what the Frac Engineer should be doing is trying to find an “average” value for each of these properties, such that the simulator’s predicted net pressure curve follows the trend (and “average” value) of the job plot, but does not necessarily follow every minute rise and fall in pressure. However, the Frac Engineer must be able to react quickly when a short term trend has become a long term trend. When this happens, it’s time to start adjusting some of the formation properties. Real-time modeling is only effective on long treatments, where the Engineer has time to spot the long term trends, adjust the model, and still be able to make changes to the treatment schedule in time for them to have some effect. If the job is too short, the crew can be pumping the displacement before the Frac Engineer has finished the pressure match. The problem outlined in Point 3 (above) is exacerbated if the wellbore volume represents a significant part of the treatment. If this is the case, the treatment can be close to displacement before the proppant has even passed into the fracture. In such cases, there is little point in modeling the fracture real-time.

References Standard Practices Manual, BJ Services, January 2001 onwards Equipment and Technology Catalogue, BJ Services, 1990 onwards Gidley, J.L., et al: Recent Advances in Hydraulic Fracturing, Monograph Series Vol 12, SPE, Richardson, Texas (1989). Johnson, D.E., Wright, C.A., Stachel, A., Schmidt, H., and Cleary, M.P.: “On-Site Real-Time Analysis Allows Optimal Propped Fracture Stimulation of a Complex Gas Reservoir”, paper SPE 25414, presented at the SPE Production Operations Symposium, Oklahoma City, March 1993.

Page 184


BJ Services’ Frac Manual 18. Real-Time Monitoring & On-Site Redesign Crockett, A.R., Okusu, N.M., and Cleary, M.P.: “A Complete Integrated Model for Design and th Real-Time Analysis of Hydraulic Fracturing Options”, paper SPE 15069, presented at the 56 California Regional Meeting of the SPE, Oakland CA, April 1986. Meyer, B.R., Cooper, G.D., and Nelson, S.G.: “Real-Time 3-D Hydraulic Fracturing th Simulation: Theory and Field Case Histories”, paper SPE 20658, presented at the 65 SPE Annual Technical Conference and Exhibition, New Orleans LA, Sept 1990. FracPro Version 8.0+ On-Line Help, RES/Gas Research Institute, March 1998 onwards. FracproPT Version 9.0+ On-Line Help, Pinnacle Technologies/Gas Research Institute, July 1999 onwards.

Page 185


BJ Services’ Frac Manual 19. Post-Treatment Evaluation

19.

Post-Treatment Evaluation

The Frac Engineer’s job is not over once the treatment has been pumped. Aside from monitoring fluid samples and preparing a post job report, the Engineer also needs to evaluate exactly what has happened in the formation. This is essential if the operating company plans to do more than one frac in a formation. The simplest method for assessing the effectiveness of the treatment is to compare before and after production. However, this does not really tell us much. In order to increase the effectiveness of future treatments, we need some idea of the size and shape of the fracture that was actually produced. Some of the methods described below – such as pressure matching - are relatively easy for the Frac Engineer to perform. However, other methods, such as tiltmeters and microseismic, require considerable expenditure and planning by the operating company. This means that plans for post-treatment evaluation must be made when planning for the treatment itself.

19.1

Pressure Matching

Pressure matching is part science and part art. In order to perform a quick and efficient pressure match, it is essential to have a good knowledge of the fracturing process, an understanding of the various rock mechanical properties, an understanding of fracture mechanics and, ideally, a reasonable idea of how the fracture simulator works. In spite of this need for an understanding of the physics behind the fracturing process and the fracture simulation, there is still an art to pressure matching. Some Frac Engineers have a feeling for this process, and some do not. Pressure matching is a very powerful tool that allows the Frac Engineer to “tune” the fracture simulator to the formation. The idea being that once the simulator has been tuned, further fracture simulations can be performed with a high degree of accuracy.

The Process of Pressure Matching Pressure matching is all about making the simulator predict the same pressure response as the reaction actually produced by the formation. This is illustrated in Figure 19.1a, as shown below:-

After

Net Pressure

Net Pressure

Before

Actual Net Pressure Calculated Net Pressure

Job Time

Actual Net Pressure Calculated Net Pressure

Job Time

Figure 19.1a – Pressure matching. The variables in the simulator are adjusted to make the calculated net pressure match the actual net pressure.

With reference to Figure 19.1a, before the pressure match (LHS), the net pressure predicted by the fracture simulator does not match the actual net pressure in any way. After the Page 186


BJ Services’ Frac Manual 19. Post-Treatment Evaluation pressure match has been performed (RHS), the computer predicts a very similar pressure response to that of the actual treatment data. Now - according to the theory - the simulator has been “tuned” to the formation. This allows the Frac Engineer to input any desired treatment schedule, and the simulator will be able to predict the fracture geometry with a reasonable degree of precision. There is no doubt that the advent of pressure matching has greatly improved the success rate and effectiveness of hydraulic fracturing. Modern fracture simulators equipped with this facility have gradually made the process increasingly user-friendly, helping to reduce the “black art” associated with frac engineering, as more and more Engineers feel capable of designing a fracture treatment. However, there are some definite limitations to this process:1.

Garbage In = Garbage Out. The computer model of the formation generated by this process is only as good as the data used to create it. Poor data on items such as permeability, net height, fluid properties (both formation and fracturing fluids) and perforations can make an otherwise perfect pressure match almost irrelevant. Another major source of errors is the use of surface pressure data to calculate BHTP. In order to calculate BHTP, the model first needs to calculate the fluid friction pressure, something that is notoriously difficult to do for a crosslinking fluid. Variations in fracturing fluid properties (such as those caused by problems with liquid additive systems, or varying gel properties) can also be very difficult to account for. Therefore, the Frac Engineer should do everything possible to get reliable bottom hole pressure data, such as that from a gauge or dead string.

2.

No Unique Solution. The process of pressure matching involves adjusting four major variables (Young’s modulus, stress, fracture toughness and leakoff) and many other minor variables, for each rock strata affected by the fracture, until the pressure response predicted by the model matches the actual pressure response of the formation. This means that the Frac Engineer may have 30 or 40 variables available for adjustment. It is therefore quite possible for two Frac Engineers to get good pressure matches, but with significantly different sets of variables.

3.

The Fracture Model. At the end of the day, the results of the pressure match are only as good as the fracture model itself. Without a doubt, modern fracture simulators are tremendously advanced – the product of more than 20 years of innovation, experimentation and inspiration. However, the fact remains that different fracture simulators will predict different fracture geometries for the same input data. Which one is right? Probably they are all wrong – so the correct question to ask is which one is closest to the Truth? This is difficult to say, and the subject of considerable debate in the fracturing industry. The popular conception is that one fracture simulator is good for a certain type of formation, whilst another is good for a different type. The debate continues.

It should also be remembered that in general (GOHFER excepted) the widely used frac models all predict a single eliptically-shaped fracture, either side of the wellbore, symmetrical around the wellbore. In reality, the fracture is probably much more complex than this. It is highly unlikely that the fracture - or more likely fractures - behaves in such a regular and predictable manner. What the fracture simulators do is predict a simplified fracture that behaves, on average, in a similar fashion to a much more complex reality

The Four Main Variables There are four main variables that the Frac Engineer should be adjusting in order to achieve the pressure match – that is to say, four main variables in each formation affected by the fracture. These variables are Young’s modulus, stress, fracture toughness and fluid leakoff. So even for the most basic formation lithology, the Frac Engineer will have to be able to keep track of a minimum of 12 variables (the zone of interest, plus the formations above and below)..

Page 187


BJ Services’ Frac Manual 19. Post-Treatment Evaluation Of course, each fracture simulator comes complete with a whole plethora of variables that the user can adjust. In fact there are so many, that it could be possible to vary several hundred parameters for a complex reservoir with several rock strata. This is for fracture simulator and rock mechanical experts only. Unless there is a really good reason, the Frac Engineer is advised to stick to the four variables listed below. Whilst pressure matching, the Frac Engineer should be aware of the fact that the process works the opposite way around to designing a treatment. In pressure matching, the bottom hole treating pressure is fixed, whereas it is a variable in fracture design. For example, an increase in in-situ stress will have the effect of decreasing the net pressure in the pressure matching process, whilst in the fracture design process, this net pressure will remain constant and the bottom hole treating pressure will increase. In pressure matching, the Frac Engineer adjusts unknown formation properties to match a known pressure. In treatment design, these formation parameters are (hopefully) already known, and the process instead involves seeing the effect they produce for a given treatment schedule. Fracture Toughness, K1c Strictly speaking, K1cis the critical stress intensity factor for failure mode 1 (see Section 9, Fracture Mechanics). However, it is commonly referred to as the Fracture Toughness and is a measure of how much energy it takes to propagate a fracture through a given material. In hydraulic fracturing, where the energy needed to propagate the fracture comes in the form of fluid pressure, fracture toughness tells us what proportion of the available energy is used to physically split the rock apart at the fracture tip. As pressure is essentially energy per unit volume, K1c tells the Frac Engineer how much net pressure is required to propagate the fracture. Generally speaking, soft plastic formations will have high fracture toughness, whilst hard brittle formations will have low fracture toughness. There is also an approximate inverse relationship between Young’s modulus and fracture toughness – hard formations tend to have a high E and a low K1c, and soft formations tend to be the other way around. For the Frac Engineer, increasing the value of fracture toughness will tend to make it harder for a fracture to propagate through the rock. Therefore, an increase in fracture toughness will generally make the fracture shorter and wider. However, an increase in fracture toughness for just one formation will tend to divert the fracture into an adjacent formation. For example, if the K1c is increased in the perforated interval, the fracture will grow into the adjacent formations, above and below. This has the effect of limiting the fracture length and increasing the fracture height. In soft formations, do not be afraid to use quite large values for this property, even several times the default values included in the simulator Fracture toughness is a material property and cannot be altered by anything under the control of the Frac Engineer. It is also a property that is very difficult to measure. There are several laboratory methods for determining K1c, but these are limited in their reliability, as fracture toughness is highly dependent upon down hole conditions and the overall geometry of the fracture. However, if core samples are available, fracture toughness can be estimated from laboratory measurements of yield stress and Young’s modulus, provided this is determined under tri-axial loading, at bottom hole temperature and pressure. Remember that some fracture models (e.g. FracPro and FracproPT) have moved away from the concept of Fracture Toughness and instead model non-linear elastic effects at the fracture tip as being more significant. In such models, variations in Young’s modulus and in-situ stresses are far more significant. Young’s Modulus, E In order for the fracture to propagate it must obtain width, to a greater or lesser extent. In order to do this, the rock on either side of the fracture has to be compressed. As discussed in Section 7 (Rock Mechanics), Young’s modulus defines how much energy is required to perform this compression. Rocks with a high Young’s modulus will require a lot of energy (a.k.a. net pressure) to compress. In these formations, fractures tend to be relatively thin, and the rock is referred to as “hard”. Similarly, rocks with a low Young’s modulus require relatively Page 188


BJ Services’ Frac Manual 19. Post-Treatment Evaluation little energy to produce width. In these formations, fractures tend to be relatively wide, and the rock is referred to as “soft”. Young’s modulus is a fundamental material property and, like the fracture toughness, cannot be altered by anything under the Frac Engineer’s control. It can be measured from core samples, provided these tests are carried out under tri-axial load conditions, at bottom hole temperature and pressure. In some formations (especially weak or unconsolidated rocks), Young’s modulus may not be constant. Fracture mechanics, rock mechanics and fracture simulation require the use of the static Young' s modulus. This is the Young’s modulus measured under static - or relatively static – conditions, such as those that occur whilst fracturing. Another form of Young’s modulus, the dynamic Young’s modulus (the Young’s modulus measured under dynamic conditions), can be measured by so-called “stress logs”. These logs, generated by a dipole sonic wireline tool (also called a sonic array), measure dynamic Young’s modulus and Poisson’s ratio by measuring the transit time of both shear and compression sonic waves. However, there can often be a significant difference between dynamic and static values, which renders the actual values reported on stress logs to be unreliable. This is covered in more detail in Section 7.10. However, stress logs can accurately report contrasts in Young’s modulus, which are almost as important as the absolute values themselves. An increase in Young’s modulus makes it harder for the fracturing fluid to produce width. Therefore, increasing this variable will make the fracture thinner, higher and longer, and vice versa. Increasing E only in the perforated interval will have the effect of forcing the fracture out of the zone of interest – i.e. increasing fracture height. A decrease in E has the opposite effect. In-Situ Stress, σ In-situ stress (often referred to as confining stress or horizontal stress) is the stress induced in the formation by the overburden and any tectonic activity. Put simply, it is pre-loading on the formation, the stress that has to be overcome (or pressure that has to be applied) in order to actually start pushing the formation apart. The actual bottom hole fracturing pressure is the pressure required to overcome these in-situ stresses, plus the pressure required for propagating the fracture (as a consequence of fracture toughness) and the pressure required to produce width. As previously discussed in Section 7, fractures will tend to propagate perpendicular to the minimum horizontal stress (i.e. along the line of least resistance). So the in-situ stress of a formation is the minimum horizontal stress of the formation, plus any tensile strength the rock may posses, and less any effects due to reservoir pressure. As the horizontal stress only exists because of the overburden (ignoring tectonic effects), it is highly dependent upon the Poisson’s ratio of the formation, as illustrated in Equation 7.18. At the limit, a Poisson’s ratio of zero means that the horizontal stresses are equal to zero, plus the effects of pore pressure. This is a theoretical minimum – in practice no material will ever have a Poisson’s ratio of zero. At the other limit, the maximum theoretical value for Poisson’s Ratio is 0.5 – at this value, the horizontal stresses will be equal to the overburden, plus the effects of pore pressure. So-called “stress logs” actually measure the dynamic Young’s modulus and Poisson’s ratio of the formation. Therefore, if the overburden is known (derived from a density logs and the TVD of each formation), the approximate in-situ stress can be calculated. However, the stresses generated from this procedure are derived from the dynamic (rather than static) Poisson’s ratio. Therefore, any stresses generated by this method are unreliable. The absolute value of these stresses cannot be trusted – however, stress contrasts between formations can be used as an indication of potential fracture height containment. In the pressure matching process, an increase in s means a reduction in net pressure (for a fixed BHTP). This means that the fracturing fluid has less energy available to fracture the Page 189


BJ Services’ Frac Manual 19. Post-Treatment Evaluation formation, and so the width, the height and the length of the fracture are all decreased. This in turn means that the volume of the fracture has decreased. However, the same volume of fluid has been pumped into the formation, so an increase in s also has the effect of increasing leakoff rate and decreasing fracture efficiency. The opposite effect applies for a decrease in in-situ stresses. Fluid Leakoff Rate, QL The fluid leakoff rate can be controlled by altering a number of variables, depending upon the fracture simulator being used, and the fluid leakoff model being employed:Pressure differential (fracturing fluid pressure minus pore pressure) Formation permeability Formation porosity Formation compressibility Formation fluid viscosity Fracturing fluid filtrate viscosity Fracturing fluid wall-building coefficient Spurt loss. A lot of these variables are difficult to measure or determine. However, the Frac Engineer should remember the ultimate objective of determining fluid leakoff – to calculate the volume of the fracture. To this end, the simulator has to be able to accurately calculate the volume of fluid lost through each unit area of the fracture face. Whether or not this is achieved by varying the permeability or the wall building coefficient is almost irrelevant. On top of this, fluid leakoff can be dramatically complicated by fracture fluid flow into fissures or natural fractures, the geometry of which can vary with the net pressure. In most cases, the Frac Engineer will have reasonable data for some of these values – and will have to guess at others. Therefore, a good strategy is to fix those values that have reasonable data, and vary the others, until the desired leakoff is obtained. Fluid leakoff is a loss of energy from the fracturing fluid, as the total energy available for propagating the fracture is equal to the net pressure multiplied by the fracture volume. High leakoff means low fracture volume, and vice versa. Therefore, and increase in fluid leakoff will tend to decrease width, height and length. The opposite applies for a decrease in leakoff. Summary of the Effects of the 4 Main Variables The basic effect of each of these variables – when applied to a fracture in a single formation – can be summarised in Table 19.1a, below:Effect of an Increase in Selected Variable

Variable

Height

Length

Width

Net Pressure

Fracture Toughness, K1c

Decrease

Decrease

Increase

Increase

Young’s Modulus, E

Increase

Increase

Decrease

Increase

In-Situ Stress, σ

Decrease

Decrease

Decrease

Decrease

Fluid Leakoff Rate, QL

Decrease

Decrease

Decrease

Decrease

Table 19.1a – The effects of an increase in each of the four, main pressure matching variables. Note that these are the overall effects when the change is taken in isolation (i.e. no other changes take place). It also assumes that the fracture is unaffected by boundary layers above and below.

Page 190


BJ Services’ Frac Manual 19. Post-Treatment Evaluation This table should be used with caution, as it applies only when the fracture is confined within a single formation. If the fracture propagates into separate formations above and below the productive interval, then an increase in (for instance) fracture toughness will make it harder for the fracture to propagate through the main formation, forcing the fracture up and down. So, in this instance, an isolated increase in a property in just one formation, can actually increase the fracture height. The Effect of Poisson’s Ratio Poisson’s ratio is at the same time both important and largely irrelevant to pressure matching. It is important, because it has a major effect of defining the horizontal stresses in a formation. However, in most cases, the Frac Engineer will be determining these stresses form pressure data, not from Poisson’s ratio data. In most fracture simulators, Poisson’s ratio is used in the 2 2 form (1 - ν ) to modify Young’s modulus (i.e. E/(1 - ν ) – the plane strain Young’s modulus). This means that a large change in Poisson’s ratio, say from 0.25 to 0.35, only produces a 2 change in (1 - ν ) from 0.9375 to 0.8775 (so that a 40% increase in n produces only a 6.4% 2 decrease in(1 - ν ). Therefore, the Frac Engineer should not spend too much time varying Poisson’s ratio during the pressure match. Input what seem to be reasonable values, and then ignore it.

Tips for Pressure Matching 1.

“The Fundamental Interconnectedness of Everything” *

The Frac Engineer must be aware that the fracture is a continuous, dynamic entity. It is not composed of a number of discrete pieces, functioning independently of each other. This means that any change to any single variable will affect the whole of the fracture, to a greater or lesser extent. This can sometimes be very discouraging for the Frac Engineer, as a change to match one part of the pressure curve can alter a match already achieved in another part of the curve. However, remember that in reality, a pressure match that only matches a limited part of the plot is not really a pressure match at all. This means that the Frac Engineer should try to change only one variable for each simulator run. This can be time consuming, but is essential if the Engineer intends to keep track of how individual changes affect the overall simulation. * - Acknowledgement to Douglas Adams, Dirk Gently’s Holistic Detective Agency

2.

Ignore Short-Term Trends

Fracture simulators model a formation as a homogenous material, whose properties are uniform throughout the material. In reality, this is not the case. Real rock formations will tend to have variations – large and small – throughout their structure. These will produce any number of short-term pressure spikes and drops during the treatment. Do not even attempt to model them. In practice during the pressure match, try to use average values for formation properties that will produce a good “overall” value. Thus, the ideal pressure match will produce a relatively smooth curve that closely follows the trend of the real data, but does not match every single variation (see Figure 19.1a). In particular, most treatments see a “break down” pressure, right at the beginning of the treatment. This generally means a large pressure spike, followed by a lower, more stable pressure. This pressure spike is caused by near wellbore effects and should not be matched.

Page 191


BJ Services’ Frac Manual 19. Post-Treatment Evaluation 3.

Watch Out for Tortuosity

Tortuosity can seriously affect the bottom hole treating pressure. Remember that even if the Frac Engineer has access to bottom hole gauge pressure data, this data will be from inside the wellbore, not inside the fracture. Tortuosity can often vary significantly during a treatment. In particular, an increase in proppant concentration can often produce a pressure rise if tortuosity is present. The various methods for identifying and quantifying tortuosity have been discussed in earlier sections of this manual. If these methods are used it is possible – to a certain extent – to allow for these effects. Another thing to be careful of is an overuse of the tortuosity tables in the fracture simulator. The latest versions of the main models allow different pressure drops to be entered for different periods during the pumping. By putting enough detail and enough stages into these tables, it is possible to get the simulator to predict virtually any net pressure profile imaginable. The Frac Engineer must have a grasp of what is realistic and what is not. This mainly comes with experience. 4.

Start with the Pressure Decline

The best place to start the pressure match is with the post-treatment pressure decline. This is because the fluids are stationary and effects such as pipe friction, perforation friction and tortuosity are eliminated. It is also often possible to identify the closure pressure on the decline curve. This value is equal to the in-situ stress for the formation next to the perforations. Once this value has been obtained, the end of treatment net pressure is defined (the difference between the ISIP and the closure pressure). The four main variables should be adjusted to produce this net pressure and to match the shape and length of the pressure decline between ISIP and closure. This gives the Frac Engineer a good starting point. Obviously, as the pressure match continues and the Engineer alters variables to match the rest of the treatment, the pressure match for the pressure decline will be altered. So further changes have to be made to bring this back into match. Which in turn will affect the rest of the pressure match, and so on. This is a part of pressure matching – the process of gradually making smaller and smaller changes to the variables until all the seemingly contradictory requirements are met. 5.

Early Time vs Late Time

At the start of the treatment the fracture is relatively small, and will be confined to the formations at, or near to, the perforations. Therefore, during this “Early Time” period, there is little point in altering the properties of formations that are away from this area. However, as the treatment progresses into “Late Time”, the fracture will become increasingly influenced by the properties of formations vertically further away from the perforations. Therefore, if the first stage in the pressure match process is to match the pressure decline, the next stage is to match the Early Time fracture. At this point, there will be fewer variables to alter. Once an Early Time match has been obtained, match the Late Time section. Then keep repeating until the match has been achieved. 6.

Remember Nolte Analysis

In spite of the fact that Nolte analysis is based on PKN 2-D fracture modelling, the basic principals can be very helpful when pressure matching. For instance, a gradual rise in the next pressure plot indicates fracture containment, whilst a decline probably means a preferential height growth and possibly a radial fracture (or GDK fracture geometry, sometimes found when fracturing coal seams).

Page 192



19.2

Well Testing for Fracture Evaluation

Well testing is sometimes used both to assess the effect of the treatment and to help determine the size and shape of the fracture. To do this, well tests have to be performed both before and after the treatment. Data that is collected before the treatment is used to help evaluate the fracture geometry afterwards. Both pre- and post-treatment tests should be performed at constant rate (or as near as possible), rather than at constant drawdown, and should be followed by a shut in (or pressure build-up) period, lasting for at least as long as the flow time. In practice, it is possible to monitor the pressure build-up real-time and see when the build-up can be terminated. The post-treatment well test can take some time, as treatment fluids must be recovered first, and the well must reach some kind of relatively steady flow. Figure 19.2a illustrates the basic anatomy of a drawdown / build-up well test.

Pi Pws(tp+∆ ∆t)

BHP

∆Pdrawdown

∆Pbuild-up Pwf(t) Pwf(tp ) = Pws(∆ ∆t = 0) CONSTANT RATE DRAWDOWN

0

t

BUILD-UP

tp

∆t

Figure 19.2a – Anatomy of a drawdown / build-up well test (after Agarwal, 1980)

In Figure 19.2a, there are several variables that are often referred to in well test analysis, which can be broken down into two groups – time and pressure. All pressures refer to BH pressure. At the start of the well test t = 0, and the BHP = Pi, which ideally will be the reservoir static pressure. Sometimes this is not the case, if the well has not been left static for a long enough period. However, this can be allowed for in well test analysis. As the well is produced at a constant rate, the length of time the well has produced for is called t and the flowing BHP is referred to as Pwf. Because this variable is dependent upon t (the longer the well is produced, the lower the BHP), it is said to be a function of t, and so the notation Pwf(t) is used to denote this. The difference between the initial pressure (Pi) and the actual flowing pressure (Pwf(t)) is referred to at the drawdown, or ∆Pdrawdown. The well is flowed at a constant rate (which may require the varying of chokes) until it is shut in. At this point, t is said to equal the producing time, tp (which is a constant, for any given test). After the well is shut in, the nomenclature ∆t is used to describe the shut in time, such that at the point of shut in t = tp and ∆t = 0. Thereafter, time is described as tp + ∆t, with tp fixed and ∆t increasing as the build-up progresses. At the point of shut in, the BHP pressure is referred to as Pwf(tp) – well flowing wellbore pressure at tp – or as Pws(∆t = 0), the static wellbore pressure at ∆t = 0. These two pressures are identical. After shut in, during the pressure build-up, the now static BHP is referred to as Pws(tp+∆t) – this means that the wellbore static pressure (Pws) is a function of shut in time (tp+∆t). Finally, the difference between the shut in pressure (Pwf(tp) or Pws(∆t=0)) and the wellbore static pressure (Pws(tp+∆t)) during the build-up is referred to as the build-up pressure, or ∆Pbuild-up.

Page 193


BJ Services’ Frac Manual 19. Post-Treatment Evaluation Figure 19.2a illustrates the most basic type of well test, the constant rate drawdown and static build-up test. This type of test can be applied to the well both before and after the treatment, as discussed below. There are many other and more complex types of well test performed, designed to get more accurate data under specific circumstances. A full discussion on the current state of well testing and well test data analysis is beyond the scope of this manual and the reader is invited to consult the references for further information.

Infinite or Finite Conductivity Fractures From a pressure transient analysis perspective, propped hydraulic fractures fall into two categories, infinite conductivity and finite conductivity. The pressure transient behaviour of these two types of fracture is significantly different. Infinite Conductivity fractures have no significant pressure drop as the fluid passes down the fracture. Therefore, pressure transients happen outside of the fracture, either in the wellbore or in the formation. With this type of fracture, the productivity of the well-fractureformation system is limited by the ability of the formation to deliver formation fluids to the fracture, rather than by the ability of the fracture to transport the fluids. This type of fracture is typical of low permeability and/or tight gas fracturing. Finite Conductivity fractures have a significant pressure drop as the fluid passes down the fracture. Therefore pressure transients occur inside the fracture, as well as in the wellbore and the formation. With this type of fracture, the productivity of the well-fracture-formation system is limited by the ability of the fracture to transport formation fluids to the wellbore, i.e. by the fracture conductivity. This type of fracture is typical of high permeability fracturing.

Pressure Transient Analysis When a well is flowing, it can be in one of three states – Steady, Pseudo or Transient. The difference between these three states was discussed in Section 12 of this manual. During well testing, the flow is usually in the transient state, which is the most complex of all to analyse, and occasionally in pseudo-steady state. Put basically, steady state flow behaves as per Darcy’s Equation, with a constant re and Pi, whilst transient flow behaves like there is no outer boundary to the reservoir (so that the radius of investigation, rd, is continually increasing). Pseudo-steady state is halfway between the two, with a constant re and reservoir pressure that declines with production (i.e. a bounded reservoir). In reality, steady state Darcy radial flow very rarely exists. The difference between transient and pseudo-steady state can be seen from constant rate drawdown and pressure build-up tests, as shown in Figure 19.2b. The basic Equation for pressure transient analysis is relatively easy to comprehend (although its derivation is very complex).

Pi – Pr, t

=

q Bo µ 4π kh

φµcr2 1-e

4kt

............................................ (19.1)

where Pi is the static reservoir pressure, Pr, t is the pressure at radius r after time t, q is the stabilised flow rate, Bo is the oil formation volume factor (a factor used to correct surface volumes to bottom hole volumes) in rbbls/stb, µ is the viscosity of the produced fluid or fluids (at bottom hole conditions), k is the permeability, h is the net height, φ is the porosity of the formation and c is the overall compressibility of the formation and fluids (also called ct).

Page 194


BJ Services’ Frac Manual 19. Post-Treatment Evaluation Constant Rate Drawdown

Pressure Build Up

Transient

BHP

BHP

Pseudo-Steady State

Transient

Pseudo-Steady State

t

t

Figure 19.2b – Graphs illustrating the deviation from transient flow caused by a reservoir boundary (i.e. pseudo-steady state flow)

Under constant rate drawdown Equation 19.1 can be simplified to:Pi - Pwf

=

q Bo µ 4π kh

loge

kt

φµcrw2

+ 0.809 ................................... (19.2)

In field units (pressure in psi, flow rate in bbls per day, viscosity in cp, distances in ft, -1 compressibility in psi , permeability in mD, time in hours and porosity expressed as a fraction):Pi - Pwf

=

162.6 q Boµ kh

log10

kt

φµcrw2

- 3.23 ........................... (19.3)

So for transient flow, during a constant rate drawdown, a plot of Pwf against log t will produce a straight line of gradient equal to (162.6qµ /kh). From this, if the flow rate, viscosity, volume factor and net height are all known, the permeability can be evaluated by measuring the gradient of the straight line portion of the curve, as shown in Figure 19.2c. This is a very common method for evaluating permeability, but is dependent upon the well being produced at a constant – or nearly constant - rate. The permeability value produced by the test is much more useful for Frac Engineers than permeability derived from log or core analysis. First, this value is an "average" Figure for the whole net height being produced. Second, well test analysis is the only investigative method that penetrates deep into the reservoir and so the results are not influenced by irrelevant near wellbore effects. The radius of investigation of the test, rd, can be evaluated using Equation 19.4. This allows the distance at which a boundary is observed (see Figure 19.2b) to be estimated, by setting the value of t (in hours) to be when the drawdown semi-log plot starts to deviate from the straight line. rd

Page 195

=

2

0.00105 k t ........................................................... (19.4) φµ c



162.6 q Bo µ kh

Pwf

slope =

log10 t Figure 19.2c – Constant rate drawdown semi-log plot. The straight line section can be used to evaluate the permeability. The deviation from the straight line at late time, is due to boundary effects of the reservoir, as the transient flow changes to pseudo-steady state flow.

rd

=

2

0.00105 k t ........................................................... (19.4) φµ c

After shut-in, the pressure in the well starts to build up, as illustrated in Figure 19.1a. The Equation defining the behaviour of the pressure is as follows, in field units:-

P* (Transient)

slope = m = Pws(tp+∆ ∆t)

P* (PseudoSteady State)

162.6 q Bo µ kh

INCREASING ∆t

0

log10

tp + ∆ t ∆t

Figure 19.2d – Example Horner plot, showing extrapolation of the straight line portion to obtain P*, the average static reservoir pressure. Once again, deviation from the straight line is caused by a change from transient flow to pseudo-steady state flow.

Page 196


BJ Services’ Frac Manual 19. Post-Treatment Evaluation Pi – Pws(tp+∆t)

=

162.6 q Boµ tp+∆t log10 ....................................... (19.5) kh ∆t

Therefore, a plot of Pws(tp+∆t) against log10[(tp+∆t)/∆t] will produce a straight line for transient flow, which will have a slope equal to (162.6qµ /kh). This plot is often referred to as a Horner plot and is a widely used tool in pressure transient analysis. An example is illustrated in Figure 19.2d. Note that as ∆t tends to infinity, (tp+∆t)/∆t tends to 1 and hence log10[(tp+∆t)/∆t] tends to 0. Therefore, by extrapolating the Horner plot back to where the x-axis equals zero, an estimate for the static reservoir pressure can be made. This means that the pressure buildup portion of the well test can be more useful than the drawdown phase, and that the well does not need to be shut in for a long time prior to the well test, in order to get an accurate Figure for Pi. So from the drawdown test, we can get reliable data for the average reservoir pressure and k (or often, kh, as the net height may be unknown). We can also get the skin factor S (see Section 2.5) from the build-up data by applying the API Skin Factor Equation, in field units:S

= 1.151

Pws(∆t = 1)-Pwf(tp) k - log10 + 3.23 ..... (19.6) m φµcrw2

where Pws(∆t = 1) is the static wellbore pressure 1 hour after the well is shut in, and m is the slope from the Horner plot, as shown in Figure 19.2d. Once the skin is known, the pressure drop due to the skin, ∆Pskin can easily be calculated:∆Pskin

= 141.2

q Boµ S 2πkh

(field units) ................................... (19.7)

Which can be worked back into Equation 19.3:Pi - Pwf

=

162.6 q Boµ kh

log10

kt

φµcrw2

- 3.23 + S ................. (19.8)

Diagnostic Plots Diagnostic plots are standard plots used to determine the characteristics of the reservoir. Usually, the diagnostic plot will consist of a log-log plot of the change in wellbore pressure, ∆P, against shut in time, ∆t, for the pressure build-up. Sometimes, semi-log plots (∆P plotted against log ∆t) are also used. In addition, the derivative of the pressure build-up, ∆P’, is also plotted alongside the pressure data. This is a slightly different derivative than that used in minifrac pressure decline analysis, and is generally calculated as follows:∆P’

= ∆t

∆P .................................................................... (19.9) ∆t

Usually, this will produce a very noisy derivative plot, and it is common practice to use some kind of smoothing or averaging algorithm to produce a clear derivative trend. Modern computer-based analysis makes this easy. Example diagnostic plots for fractured and non-fractured wells are shown in Figures 19.2e to 19.2h, below (after Economides et al, 1994):-

Page 197



∆P

∆P

∆P'

∆P'

log10 t

log10 t Figure 19.2e – Log-log diagnostic plot with derivative for the pressure build-up of an infinite-acting reservoir (i.e. no boundaries and no pseudo-steady state flow).

Figure 19.2f – Log-log diagnostic plot with derivative for the pressure build-up of reservoir with a partial boundary (e.g. a sealing fault).

∆P ∆P ∆P'

∆P'

log10 t

log10 t Figure 19.2g – Log-log diagnostic plot with derivative for the pressure buildup of an infinite conductivity fracture

Figure 19.2h – Log-log diagnostic plot with derivative for the pressure buildup of a finite conductivity fracture

Gas Well Testing Gas well testing is an order of magnitude more complex than oil/water well testing, due to the fact that the above theory assumes that the produced fluid is incompressible. Obviously, this is not the case for gas wells. To compensate for this, Equation 19.8 is modified as follows:2

Pi - Pwf

2

=

1639 Q µiziTBg kh

log10

kt

φµicrw2

- 0.351 + 0.87S

(19.10)

This Equation is in field units, were Q is the gas flow rate in scf/d, µi is the gas viscosity at static reservoir conditions, zi is the z-factor at static reservoir conditions (the z-factor is a dimensionless factor used to correct the ideal gas Equation to allow for real gas behaviour, and is calculated or measured for each reservoir), T is the reservoir absolute temperature in rankine and Bg is the gas formation volume factor (this is a factor used to correct surface volumes to reservoir volumes, with units of cuft/scf). Usually, this Equation is rearranged so that it is more conveniently used:m(Pi) – m(Pwf) Page 198

=

1639 QTBg kh

(log10tD – 0.351 + 0.87S) .................. (19.11)


BJ Services’ Frac Manual 19. Post-Treatment Evaluation where 2

m(Pi)

=

Pi ........................................................................... (19.12) µi zi

m(Pwf)

=

Pwf ........................................................................... (19.13) µi zi

tD

=

2.634 x 10 kt ........................................................... (19.14) φµicrw2

2

and

-4

Note that in Equation 19.14, rw is in feet. m(P) is referred to as the gas pseudo-pressure. By using Equation 19.11 for the drawdown, similar techniques can be used as for oil well testing. However, this time m(Pwf ) is plotted on the y-axis, and the slope of the straight line portion is equal to 1639 QTBg/kh. For the pressure build-up, the transformation from incompressible to compressible is similar:m(Pi) – m(Pws)

1639 QTBg kh

=

log10

tp+∆t ..................................... (19.15) ∆t

For the Horner plot, the x-axis remains unchanged, but the y-axis plots m(Pws). The straight line portion can be extrapolated back to where the x-axis = 0, to give m(Pi).

Type Curve Matching 10 2

S = 20 S = 10 S =5 10

CD = 0

S =0

10 4

10 3 = D

C

D

=

C

D

10 5

=

C

= D

C

1

10 2

PwD

S = -5

10 -1 10 2

10 3

10 4

10 5

10 6

10 7

10 8

tD

Figure 19.2i – Type curves for a single well in an infinite reservoir, with wellbore storage and skin damage (after Agarwal, Al-Hussainy and Ramey, 1970).

Page 199


BJ Services’ Frac Manual 19. Post-Treatment Evaluation Type curve matching is a technique that involves matching field generated data curves to experimentally or numerically derived type curves. The field data is moved over the type curve, until the type curve that most closely matches the field data is found. This technique has become very popular recently, as it is very easily performed by computers. Usually, there will be several different type curves available, for wells with different characteristics such as skin, wellbore storage (see below), number of wells and reservoir boundaries. A typical set of type curves is shown in Figure 19.2i, which is for a single well in an infinite system (i.e. no boundaries), with wellbore storage and skin damage. Where PwD is dimensionless pressure:PwD

=

kh (Pi - Pwf) 141.2 q Boµ

(Oil Wells) ................................. (19.16)

PwD

=

kh [m(Pi)-m(Pwf)] 1424 Q BgT

(Gas Wells)............................... (19.17)

and tD is as defined in Equation 19.14. Wellbore Storage Wellbore storage is a measure of how much the volume of liquid and gas contained in the wellbore effects the flow of the well. For instance, the pressure transient theory outlined above, assumes that there is an instantaneous change from flowing to not flowing, when the well is shut in and vice versa. This assumption is probably valid for a drill stem test (DST), where the valve being opened and closed is located downhole. However, for most well tests, the controlling valve will be at the surface. When the well is shut in, there will be some flow from the reservoir into the wellbore, otherwise the pressure in the wellbore would not rise. The only way this can happen is if the wellbore expands. Similarly, when the well is opened, and the pressure drops, the wellbore contracts. This storage effect is greatest when the wellbore volume is largest (i.e. flow through casing) or when the fluids are compressible (i.e. gas wells or wells with significant associated gas production). The wellbore storage coefficient, C, is defined as follows, with volume measured in bbls and pressure in psi:C

=

∆V ............................................................................ (19.18) ∆P

So that C is a measure of how much change in volume is produced for a given change in pressure. Dimensionless wellbore storage, CD, as used in the type curves, is defined as follows:CD

=

5.6146C 2 ................................................................... (19.19) 2πφchrw

Type Curve Matching Type curve matching is performed as follows. 1. Select a set of type curves which most closely suit the well and reservoir situation, based on items such as reservoir boundaries, skin factor, wellbore storage, number of wells and whether or not the well has a hydraulic fracture. 2. Produce a log-log plot of ∆t against ∆P (or ∆m(P) for gas wells) for the well test data. This can be for a constant flow rate drawdown, a constant drawdown flow or for a build-up. An Page 200


BJ Services’ Frac Manual 19. Post-Treatment Evaluation example is illustrated in Figure 19.2j. When constructing this plot, make sure the axes are of the same scale as the axes used in the type curves. This is a critical component of the type curve matching process, and is made very easy by computer-based methods. 3. Move the log-log plot of the test data over the type curves, until the data matches up with one of the type curves. This process can often be very difficult – especially if the data is noisy - as several type curves may have very similar shapes. Once the curves have been matched, the type curves will yield (as in the case of Figure 19.2i) the dimensionless wellbore storage coefficient and the skin factor. 4. The final step is to obtain the match pressure and match time. With the test data curve still positioned at its curve match, pick a point on the test data plot. This can be any point, and does not necessarily have to be anywhere near the data. In fact, it is often easier to pick a point where two major axes cross. Note the value of ∆t and ∆P at this point (these values are referred to as ∆tM and ∆PM). Then note the corresponding values for this point on the type curve, to give tDM and PwDM, the dimensionless match time and pressure. This process is more easily visualised if we imagine we have two hard copies of the plots. The type curve plot is on paper, whilst the test data log-log plot is on a transparency. The transparency has been moved over the type curve to obtain the match. Then, we have selected a point on the test data plot, and pushed a pin through both plots to make a small hole in each. ∆tM and ∆PM are the coordinates of the pin hole in the test data plot, and tDM and PwDM are the coordinates of the pin hole in the type curve plot. 5. The match pressures and times can now be used to obtain reservoir data. The match pressures are substituted into Equations 19.16 and 19.17, in order to determine the permeability-thickness (kh or conductivity) of the formation, as shown in Equations 19.20 and 19.21. If the net height is known, the permeability can easily be found. kh

= 141.2qBoµ

PwDM ∆PM

kh

= 1424QBgT

PwDM ∆m(P)M

(for oil wells) ........................... (19.20) (for gas wells) ........................ (19.21)

103

∆P (psi)

102

10

1 10-2

10-1

1

∆t (hours)

10

102

Figure 19.2j – Example of a log-log plot of ∆t against ∆P, used for type curve matching.

Page 201


103

BJ Services’ Frac Manual 19. Post-Treatment Evaluation 109

∆m(P) (psi2/cp)

108 1

4 Slope

107

106 10-2

10-1

1

∆t (hours)

10

102

103

Figure 19.2k – Post-treatment log-log plot of well test data for a finite conductivity fracture in a gas well. An infinite conductivity fracture would have a half slope.

Similarly, Equation 19.14 can be re-arranged to yield the porosity-compressibility product, which is another variable that will be useful in future analysis:-

φc

-3

=

2.637x10 k ∆tM tDM .................................................. (19.22) µ r w2

Post-Treatment Well Testing Once the well has been fractured, and the fracturing fluid has been recovered, the well can be tested again, to assess the performance of the fracture. Usually, this test will consist of a shut in for one hour, a constant rate flow period of about 24 hours and finally a shut-in period of about 48 hours. These numbers are typical for oil wells. Gas wells are a little more complex, as the flow is much more affected by non-Darcy flow in the fracture and wellbore storage. For gas wells, it is advisable to rely on previous experience, and where that is not available, be prepared to change the well test plan on location. Basically, for a fractured gas well, the log-log plot of the test data should look something like Figure 19.2k. With reference to this Figure (which is also applicable to oil wells, with DP as the vertical axis), flow from a fractured well should fall into 5 distinct regimes, in chronological order:• • •

Wellbore storage dominated flow. Fracture linear flow, in which the flow is dominated by liner flow down the fracture to the wellbore. This is usually characterised by a half slope on the log-log test data plot. This period of flow will also not last very long. Bilinear flow, in which fluid flow along the fracture and through the formation perpendicular to the fracture faces are both significant. This flow regime is characterised by a quarter slope on the log-log plot and should be a prominent feature of the test, if the treatment has been successful. This is the most useful portion of the test data.

Page 202


BJ Services’ Frac Manual 19. Post-Treatment Evaluation •

Formation linear flow, where the test data is dominated by linear flow through the formation, perpendicular to the fracture faces. Pseudo-radial flow. This flow regime comes last and is characterised by a combination of linear flow perpendicular to the fracture faces, and by radial flow from the formation beyond the fracture tip.

•

For a gas well, the test should continue until significant data has been obtained for the bilinear flow data, whether for drawdown or build-up. The best way to assess this is to plot the log-log plot real time and watch for the quarter slope. A typical set of post-treatment type curves is shown in Figure 19.2l. In Figure 19.2l, the x-axis variable is the fractured well dimensionless time, defined as follows:-4

tDx

=

f

2.634x10 kt ........................................................... (19.23) φµcxf2

Where xf is the fracture half-length. It should be noted that this style of type curve is only valid if the fracture half-length is significantly less than the reservoir’s radial extent. Matching the test data log-log plot to the type curve is easier than for the pre-fracture test. When the match is performed, the quarter slope portion of the test data log-log plot is matched within the shaded area of Figure 19.2l. As all the variables for the dimensionless pressure are already known, the type curve match is used to obtain the dimensionless fracture conductivity, CfD, and the match times. The match times are used to find the fracture half length, as follows in Equation 19.24. 10

1

PwD

CfD = 0.1 0.5

1.0

10-1

5.0

Region of Bilinear Flow

10

50 100 500

10-2 10-5

10-4

10-3

10-2

10-1

1

tDxf Figure 19.2l – Type curves for a well with a finite conductivity, vertical fracture (after Agarwal et al, 1979 and Economides et al, 1987).

Page 203


BJ Services’ Frac Manual 19. Post-Treatment Evaluation xf

=

2

-4

∆tM tDx M f

2.634x10 k µφc

................................... (19.24)

As k and φc where obtained in the pre-fracture well test, the fracture half-length can be quickly determined. Once xf has been found, the value for dimensionless fracture conductivity can be used to obtain the fracture conductivity from Equation 10.1. Taking this one step further, if the permeability of the proppant, kp is known (remembering to allow for proppant damage, non-Darcy flow and multi-phase flow), then the average width of the fracture, w ¯ , can also be obtained.

Quarter Slope versus Half Slope According to the theory, as the well starts to flow the system passes through five distinct phases of flow:- wellbore storage; fracture linear flow; bilinear flow; formation linear flow; and finally pseudo-radial flow. In practice, wellbore storage occurs at very early time, and fracture linear flow only occurs for a very short period of time. For the majority of the well test period, the data will be dominated by bilinear flow or formation linear flow. If the fracture has a finite conductivity, the flow will spend a considerable amount of time in the bilinear flow regime, as characterised by a quarter slope on the plot of Log ∆P against Log ∆t (the log-log plot). However, if the fracture has infinite conductivity, then the flow quickly moves from bi-linear to formation linear, which has a half slope on the log-log plot. In fact, often the quarter slope section will not be detected. Spotting the difference between the two is easy. Bi-linear flow (finite fracture conductivity) 0.25 produces a straight line on a plot of ∆P (or ∆m(P)) against ∆t , whilst formation linear flow (infinite fracture conductivity) produces a straight line on a plot of ∆P (or ∆m(P)) against ∆t. Useful information regarding the fracture dimensions can be obtained from these plots, as follows:Finite conductivity fracture (field units):k pw =

44.1qBµ 2 h mbf

k pw =

444.8q zi T 2 h mbf

1

φµctk 1

(oil wells)................................. (19.25)

φµctk

(gas wells) .......................... (19.26)

Infinite conductivity fracture (field units):xf

=

4.064 q B h mlf

xf

=

40.925 q zi T h mlf

µ

kφct

(oil wells)....................................... (19.27)

µ

kφct

(gas wells) ............................... (19.28)

where mbf is the slope of the straight line portion of the bi-linear flow plot (i.e. ∆P against 0.25 ∆t ), whilst mlf is the slope of the straight line portion of the linear flow plot (i.e. ∆P against ∆t). Therefore, by using these plots, the average propped fracture width (w ¯ ) can be found for finite conductivity fractures, and propped fracture half length (xf) can be found for finite conductivity fractures, provided an accurate proppant permeability is known, under the producing conditions.

Page 204



A Word of Caution Type curve methods for obtaining the fracture geometry from well test data are notorious for being non-unique. A glance at Figure 19.2h will show that all the curves in the bilinear flow area have similar gradients. Additionally, the analysis is very sensitive to the quality and reliability of the data obtained. Formation permeability, for instance, is not a constant. It will change with fluid saturations, so if the well shows a change in GOR, GLR or WOR between the pre-and post-treatment tests, the permeability will be suspect. Therefore, be aware of the limitations and risks associated with relying entirely upon this type of analysis.

19.3

Other Diagnostic Techniques.

Tiltmeters Tiltmeters are extremely sensitive devices for measuring changes in orientation from the vertical. Surface and downhole tiltmeters are used to measure the azimuth and geometry of the fracture versus time, as illustrated in Figure 19.3a. Surface tiltmeters measure the deflection of the earth at the surface. Usually, they are placed in 30 to 40 ft deep bore holes, and placed around the wellbore, from a distance of as little as 100ft, to as great as half a mile. These tiltmeters are used to measure the fracture azimuth, or the direction of the fracture relative to north. Because the determination of fracture azimuth can often be performed on a qualitative basis, the accuracy of the data required is less than that for determining fracture geometry. Therefore, fracture azimuth can be quite reliably determined from these devices, if used correctly. However, surface tiltmeters cannot provide any useful data regarding the fracture geometry, as they are usually too far away from the fracture and located on the wrong plane. Subsurface tiltmeters are placed in wells adjacent to the well being treated, at the same vertical depth. Because they are often much closer to the fracture than the surface tiltmeters, and they are located perpendicular to the most likely plane of fracture propagation, it is possible to obtain fracture height, width and length from them, against time.

Depth

Fracture-induced surface trough

Surface tiltmeters

Fracture Downhole tiltmeters in offset wells

Figure 19.3a – The principle of tiltmeter fracture diagnostics (after Cipolla and Wright, 2000).

The accuracy of fracture geometry determination is controlled by a number of factors. The most important factor is the number of tiltmeters used, which in turn is controlled by the number of available observation wells. Obviously, the fewer the number of tiltmeters, the less accurate the analysis is. Unfortunately, many candidate wells do not observation wells conveniently positioned, or else the operator is unwilling to shut these wells in for the duration of the set up, treatment and rig down (sometimes several days). Another major factor affecting the quantitative analysis required in order to obtain fracture geometry, is the quality of the data on the rock mechanical properties of the affected formations. The tiltmeter is Page 205


BJ Services’ Frac Manual 19. Post-Treatment Evaluation basically measuring the angular deflection of the rock at a particular point. The magnitude of this deflection will depend upon the distance between the fracture and the tiltmeters, and upon the mechanical properties of the rock between them (chiefly Young’s modulus and Poisson’s ratio). If these rock properties are unknown – or worse still if the formations are heterogeneous – then the accuracy of the measurements will be significantly reduced.

Microseismic Microseismic fracture diagnostics rely on the use of several highly accurate seismographs, similar in principle to the seismographs used to detect and measure earthquakes. Essentially, as the fracture propagates through the earth’s surface, it does not grow in a smooth, homogenous fashion. Instead, the fracture will tend to propagate in short bursts, each one of which produces a small seismic shock wave that can be detected and measured. As these microseismic events occur mostly at the fracture tip, it is possible, by using 3 or more microseismographs positioned in 3-D space around the well, to map the position of each of the microseismic events and hence the position of the fracture tip against time. It should be remembered that the fracture tip encompasses the entire perimeter of the fracture, and that the fracture (or fractures) could well be propagating all the way along this perimeter. Therefore, the technique can measure fracture height, length and the overall shape of the fracture. As with tiltmeters, this technique requires the use of measuring devices positioned in observation wells. If there are no suitable observation wells, then the technique cannot be applied. It should be remembered that any seismic device measures time, not distance. In order to convert time into distance, the velocity of the shock wave through the rock formation(s) must be known. This can often be obtained from acoustic logs, but is highly dependent upon the formation bulk density, which in turn is dependent upon the porosity and the relative saturations of liquids and gases. These factors, coupled with heterogeneity in the formation, tend to limit the resolution and accuracy of the results. However, provided enough suitably located measuring devices are used, this technique can be used to give a good overall idea of the fracture geometry and to detect multiple fractures.

Radioactive Tracers Radioactive tracers are soluble radioactive isotopes that are added to the fracturing fluids during the treatment, on the fly. After the treatment, the well is logged using a tool fitted with a Geiger-Müller detector that can identify and quantify the presence of the isotope. The idea behind this is to see where the fracturing fluid has gone. The capabilities of this technique are further enhanced by the use of three different 46 124 192 radioactive isotopes, Scandium-46 ( Sc), Antimony-124 ( Sb) and Iridium-192 ( Ir). These are run at different times during the treatment. By logging the well with a tool that can tell the difference between the isotopes, it is possible to see if different sections of the treatment went in different directions. Using radioactive tracers has a couple of drawbacks:1. Storage and transportation of the radioactive materials, even these low activity isotopes, is governed by strict regulations in most areas and can often be more trouble than this information is worth. This is especially true if the isotopes have to cross a national frontier or go offshore. 2. The rocks in the formation reduce the count measured by the detector. Therefore it is very difficult to tell the difference between a low level of radioactivity close to the wellbore, and a high level of radioactivity some distance away from the wellbore. This means that whilst the technique can be used to see where the fracture is located at the perforations, and to a certain extent which perforations took fluids at which time during the treatment, it cannot be used to detect fracture height or width. Page 206



Temperature Logs Temperature logs are used to detect where the treatment fluid has gone. By running a temperature log right after a treatment, and measuring the temperature of the well against depth, it is possible to see where the cold treating fluid cooled down the hot formation as it entered. The center of this zone is the point of fracture initiation, as illustrated in Figure 19.3b. DEPTH

Figure 19.3b – Generic temperature log illustrating that the treating fluid has entered only a small portion of the perforated interval. The fracture will have initiated in this small interval. However, this does not necessarily mean that this is the center of the fracture.

Perforated Interval

Section of Perfs that is actually taking fluid

TEMPERATURE

References Johnson, D.E., Wright, C.A., Stachel, A., Schmidt, H., and Cleary, M.P.: “On-Site Real-Time Analysis Allows Optimal Propped Fracture Stimulation of a Complex Gas Reservoir”, paper SPE 25414, presented at the SPE Production Operations Symposium, Oklahoma City, March 1993. Crockett, A.R., Okusu, N.M., and Cleary, M.P.: “A Complete Integrated Model for Design and th Real-Time Analysis of Hydraulic Fracturing Options”, paper SPE 15069, presented at the 56 California Regional Meeting of the SPE, Oakland CA, April 1986. Meyer, B.R., Cooper, G.D., and Nelson, S.G.: “Real-Time 3-D Hydraulic Fracturing th Simulation: Theory and Field Case Histories”, paper SPE 20658, presented at the 65 SPE Annual Technical Conference and Exhibition, New Orleans LA, Sept 1990. FracPro Version 8.0+ On-Line Help, RES/Gas Research Institute, March 1998. FracproPT Version 9.0+ On-Line Help, Pinnacle Technologies/Gas Research Institute, July 1999. Hagel, M.W., and Meyer, B.R.: “Utilizing Mini-Frac Data to Improve Design and Production”, Journal of Canadian Petroleum Technology, March 1994, pp. 26 – 35. MFrac III Version 3.5+ On-Line Help, Meyer and Associates Inc, December 1999.

Page 207


BJ Services’ Frac Manual 19. Post-Treatment Evaluation Cipolla, C.L. and Wright, C.A.: “Diagnostic Techniques to Understand Hydraulic Fracturing: What? Why? and How?”, paper SPE 59735, presented at the SPE/CERI Gas Technololgy Symposium, Calgary, Alberta, Canada, April 2000. Economides, M.J, Hill, A.D. and Ehlig-Economides, C.: Petroleum Production Systems, Prentice Hall, Upper Saddle River, NJ, 1994 Economides, M.J., and Nolte, K.G.: Reservoir Stimulation, Schlumberger Educational Services, 1987. Agarwal, R.G., Al-Hussainy, R. and Ramey, H.J., Jr.: “An Investigation of Wellbore Storage and Skin Effect in Unsteady Liquid Flow: I. Analytical Treatment,” Soc. Pet. Eng. J. (Sept 1970); Trans., AIME, 249. Agarwal, R.G., Carter, R.D. and Pollock, C.B.: “Evaluation and Performance Prediction of Low-Permeability Gas Wells Stimulated by Massive Hydraulic Fracturing”, paper SPE 6838, JPT, 362-372, March 1979. Agarwal, R.G., Carter, R.D. and Pollock, C.B.: “Type Curves for Evaluation and Performance Prediction of Low-Permeability Gas Wells Stimulated by Massive Hydraulic Fracturing”, paper SPE 8145, JPT, 651-656, May 1979. (Published as an accompaniment to SPE 6838, above). Bostic, J.N., Agarwal, R.G. and Carter, R.D.: “Combined Analysis of Post Fracturing and Pressure Buildup Data for Evaluating an MHF Gas Well”, paper SPE 8280, presented at the th SPE 54 Annual Technical Conference and Exhibition, Las Vegas, Nevada, September 1979. Agarwal, R.G.: “A New Method to Account for Producing Time Effects When Drawdown Type Curves are Used to Analyze Pressure Buildup and Other Test Data”, paper SPE 9289, th presented at the SPE 55 Annual Technical Conference and Exhibition, Dallas, Texas, September 1980 Crafton, J.W.: “Oil and Gas Well Evaluation Using the Reciprocal Productivity Index Method”, paper SPE 37409, presented at the SPE Production Operations Symposium, Oklahoma City, Oklahoma, March 1997. Cramer, D.D.: “Evaluating Well Performance and Completion Effectiveness in Hydraulically Fractured Low-Permeability Gas Wells”, paper SPE 84214, presented at the SPE Annual Technical Conference and Exhibition, Denver, Colorado, October 2003 Archer, J.S. and Wall, C.G.: Petroleum Engineering Principals and Practices, Graham & Trotman, London, 1986. Dake, L.P.: Fundamentals of Reservoir Engineering, Elsevier, Amsterdam, 1978 Cipolla, C.L. and Mayerhofer, M.J.: “Understanding Fracture Performance by Integrating Well Testing and Fracture Modelling”, paper 74632, SPEPF, November 2001. Arihara, N., Abbaszadeh, M., Wright, C.A. and Hyodo, M.: “Integration of Fracturing Dynamics and Pressure Transient Analysis for Hydraulic Fracture Evaluation”, paper SPE 36551, presented at the SPE Annual Technical Conference and Exhibition, Denver, Colorado, October 1996. nd

Horne, R.N., Modern Well Test Analysis – A Computer-Aided Approach, 2 Edition, Petroway Inc, Palo Alto CA, 2002. Cipolla, C.L. and Wright, C.A.: “State-of-the-Art in Hydraulic Fracture Diagnostics”, paper SPE 64434, presented at the SPE Asia Pacific Oil and Gas Conference and Exhibition, Brisbane, Australia, October 2000.

Page 208


BJ Services’ Frac Manual 20. Equipment

20.

Equipment

20.1

Horsepower Requirements

Working out horsepower requirements is a relatively easy thing to do, provided you know the expected treating pressure and slurry rate:HHP =

STP x Slurry Rate ......................................................... (20.1) 40.8

where STP is in psi and Slurry Rate is in bpm. The 40.8 is simply a conversion factor for the units (in the SI system, pumping power – in kW – is directly equal to pressure (kPa) multiplied by rate (m3/sec)). This formula will tell you how many pumps of what size you need on location. Remember to have at least 20% excess horsepower on location and - as a minimum - mobilise at least one spare pump. This excess capacity is required in case of pump failure or higher than expected treating pressures. It is also worthwhile looking at the set of curves supplied with each pump – called “pump curves”. These curves show the maximum rate and pressure that the pump can run at in each gear. Correctly speaking, these curves are showing maximum torque from the engine. Remember that it is quite possible to be limited by torque, rather than by horsepower. In such a situation, the pump may not be able to run at a given rate and pressure, even though it is within the pump’s rated horsepower. Remember also that the reduction ratios between the engine and transmission, and between the transmission and the pump, will affect the final torque available. In reality, “pump curves” are in fact “pump-transmission-engine” curves. Figure 20.1a shows an example. If a treatment is going to be close to the maximum power for a given pumping unit, it is recommended that the pump curves be consulted in order to confirm that the pump can actually do the treatment.

Figure 20.1a – Typical pump curves. This set is for a 30-16-6 frac skid, with a 16V92TA engine, a CLBT8962 transmission and a pacemaker pump with a 4.5 inch fluid end. Nominal rating of the pump skid is 700 HHP

Page 209



20.2

Flow Lines

This section is intended as a guideline only. Full details on requirements for high and low pressure flow lines can be found in the BJ Services Standard Practices Manual, and it is recommend that this should be consulted before any rig-up is designed.

Suction Hoses It is essential that sufficient suction hoses be used between the tanks and the blender. The only force available to move the fluid to the blender is the suction of the inlet pump and hydrostatic pressure from a difference in fluid levels. This is not much. In order to ensure that the suction pump receives fluid at sufficient rate, a simple rule applies; One 4” diameter 10’ suction hose will carry up to 10 bpm of gel If 20 bpm is required, then two hoses will be needed, and so on. In addition, longer hoses will carry less rate. For instance, 20’ of 4” diameter hose will only carry half as much rate, i.e. 5 bpm. So if 20 bpm were required from tanks which were 20’ away, 4 x 4” flow lines would be required. From this it is easy to see why the blender is usually placed as close as possible to the frac tanks, and why the frac tanks are often manifolded together with 8” (or larger) diameter lines. Also consider the comparative diameter of manifolds and suctions hoses. For instance:An 8” manifold has a flow are of 50.26 sq inches. This corresponds to the same flow area as 4 x 4 inch hoses (50.24 sq inches). Therefore, there is little point in building an 8” manifold and then using only 3 x 4” suction hoses. Finally, remember that there is a difference between suction and discharge hoses. Suction hoses need to be rigid, otherwise the suction pump of the blender can suck them flat. Discharge hoses, which generally do not have to carry suction, are often made from non-rigid hoses, which collapse flat when there is no fluid in them. This makes for easier storage and makes the hoses easier to carry. As a general rule-of-thumb, suction hoses can be used for the discharge (provided they have the correct pressure rating) but discharge hoses cannot be used on the suction.

Discharge Hoses The discharge hoses run from the blender to the high pressure frac pumps. Generally, one discharge hose is required from the blender to each pump. These hoses do not need to be rigid (see above comments on suction hoses) but must have sufficient pressure rating. They must also have crimped connections (similar to high pressure hydraulic hose connections) and not the old-style clamps. Discharge hoses should also be fitted with "whip-checks" at each connection. For rates below 5 bpm per pump, a single 3" discharge hose is required for each pump. At rates above 5 bpm, a single 4" hose should be used - although at very high rates (15 bpm +), more than one hose may be required.

High Pressure Flow Lines When pumping abrasive fluids – such as a frac gel with proppant – down a high pressure treating line, there is a limit to how fast it is advisable to pump. Above this pump rate, seals on

Page 210


BJ Services’ Frac Manual 20. Equipment chiksans, swivels and hammer unions start to wash out. It is generally accepted in the -1 industry that the velocity of the frac fluid should not exceed 40 ft sec . Therefore:Qmax =

2

2.33 d ............................................................................ (20.2)

where Qmax is the maximum flow rate down any single high pressure line, in bpm, and d is the inside diameter of the line, in inches. Important Points 1.

The actual inside diameter of high pressure flow lines is often significantly less than the nominal diameter. Equation 20.2 should be used with the actual diameter. This is illustrated in Figure 20.2a.

2.

HP flexible lines, such as Coflexip hoses, have separate guidelines. For these, follow the manufacturer’s instructions. Velocity Chart

Figure 1502 HP Iron 120

100

Fluid Velocity, ft sec

-1

1.5" 2"

80

3" 4"

60

40

-1

40 ft sec Max Velocity for Abrasive Fluid

20

0 0

10

20

30

40

50

60

Fluid Rate, bpm

Figure 20.2a – Chart showing fluid velocity against fluid rate for various nominal diameters of Figure 1502 high pressure iron.

20.3

High Pressure Pumps

Most high pressure pumps used in hydraulic fracturing are of the triplex variety, although quintuplex pumps are becoming more popular. Triplex means that there are three pistons acting to pump the fluid, quintuplex means that there are five. These pistons are driven by a rotating crankshaft, as illustrated in Figure 20.3a. Figures 20.3b and 20.3c show what happens whilst the pump is operating. Figure 20.3b shows the suction or inlet stroke of the cycle. As the plunger moves back towards the power end, fluid is pushed through the suction valve by the blender. The spring acting to close this valve requires 20 to 40 psi just to lift it up, so the blender must provide a boost pressure significantly greater than this in order to quickly fill the fluid end. Figure 20.3c shows the power or discharge stroke. As the plunger moves away from the power end, the increased pressure in the fluid end causes the suction valve to close, and once this pressure is high enough, the discharge valve to open. Page 211



Power End

Fluid End

Figure 20.3a – Schematic diagram of a generic frac pump

Discharge Valve

Plunger

Suction Valve

Figure 20.3b – Generic frac pump, suction stroke

Figure 20.3c – Generic frac pump, discharge stroke

Page 212


BJ Services’ Frac Manual 20. Equipment Frac pumps are usually powered by diesel engines, although some have been built with electric motors and even gas turbines. For diesel powered units (which includes all of BJ’s frac pumpers), there will be a transmission and a drive shaft in between the pump and the engine. The transmission allows the pump operator to select which gear the pump is in. Low gear is for high pressure/low rate, whilst high gear is for low pressure/high rate. The transmission usually includes a torque converter, which amplifies the torque coming from the engine, for a corresponding drop in rpm’s. The pump curves supplied with each pump will tell the operator what the maximum rate and pressure is for each gear. These curves include the engine/transmission gear ratio, which is the ratio for the torque converter. For instance a 2:1 engine transmission gear ration means that the torque converter reduces the input rpm’s by a factor of 2, and increases the input toque by a factor of 2. Also included on most pumpers is a lock-up device. This is a mechanism that allows slip between the engine and the transmission. In the event of the pump stalling, this can prevent serious damage to the transmission and engine. In order to make this device “lock up” (which means that there is no “slip” in the lock up device), the engine needs to be turning at a reasonable rate (usually 1700 to 1800 rpm). Below this speed, the torque converter is not locked up. The pump is still working, but there is slippage between the engine and transmission. It is possible to run a pumper out of lock up, but the transmission will quickly overheat if this is maintained for too long. Frac pumpers come in a variety of sizes, ranging from 350 HHP to 2700 HHP. Bigger pumps are more cost effective for big treatments, but are very expensive and can be difficult to move on roads and onto location. Smaller pumps may require more operators, more maintenance (per horsepower – maintenance per pump unit is not significantly effected by size) and take up more space on location. Figures 20.3d to 20.3g illustrate some of BJ’s fleet of frac pumpers.

Figure 20.3d – Skid mounted 16V 92T pump unit (700 HHP). Skid splits into two parts.

Figure 20.3e – Two views of a trailer-mounted Gorilla pump unit (2700 HHP)

Figure 20.3f – Body-load Kodiak pump unit (2200 HHP)

Figure 20.3g – Skid-mounted 1200 HHP pump unit

Page 213



20.4

Intensifiers

Intensifiers are devices that are used for pumping frac treatments for extended periods at high pressure and rate. They reply on conventional frac pumps to power them, and work on the principle that at constant power, high rate and low pressure is the same as low rate and high pressure. BJ Services no longer supplies intensifiers. At the power fluid end of the intensifier, the frac pumps supply power fluid at high rate and (relatively) low pressure. This acts to displace a large diameter piston down the power end. At the other end of this piston is a smaller diameter piston, which is mounted inside the downhole fluid end. This acts to pump the frac fluid at high pressure and (relatively) low rate, as illustrated in Figure 20.4a. Suction Stroke – Hydraulic fluid is forced behind the power fluid piston to force the piston back. This allows the downhole fluid end to fill with frac fluid from the blender. Power Stroke – The pressure on the hydraulic fluid is released. At the same time, the inlet valve from the frac pumps is opened, allowing the power end to fill with power fluid. This forces the piston down the power fluid end. At the other side of the intensifier, the frac fluid is forced out of the downhole fluid end at high pressure. 2

2

One important parameter for each intensifier is the intensification ratio. This is equal to D /d (see Figure 20.4a). This defines by how much the intensifier converts high rate-low pressure into low rate-high pressure. For instance, with an intensification ratio of 2.5, the fluid pressure going downhole will be 2.5 times the power fluid pressure, whilst the fluid rate going down hole will be 2.5 times less than the power fluid rate. Figure 20.4b shows how the intensifier is rigged up with the other equipment, whilst Figures 20.4c and 20.4d show intensifiers on location. TO POWER FLUID UNIT

OPEN

D

d

SUCTION STROKE FROM BLENDER

CLOSED HYDRAULIC FLUID IN

TO WELL

CLOSED

POWER STROKE OPEN HYDRAULIC FLUID OUT FROM FRAC PUMPS

Figure 20.4a – Schematic diagram of a generic intensifier

Page 214



BLENDER FRAC PUMP

INTENSIFIER

FRAC PUMP

FRAC PUMP

COOLER

BOOST PUMP

RESERVOIR

TO WELL

POWER FLUID UNIT Figure 20.4b – Schematic diagram of the intensifier hook-up.

E

D

A C

B

Figure 20.4c – Intensifier worksite. Each intensifier (A) is hooked up to three frac pumpers (B), which are pumping the power fluid. Power fluid is handled by the power fluid unit (C). Intensifiers are rigged into a manifold (D). Note that whilst there are three intensifiers and 9 power fluid pumpers on location, there are also an additional two frac pumpers (E) rigged up to the downhole line to provide extra horsepower.

Page 215



Figure 20.4d – Detail of an intensifier. In the foreground, on the RHS, is the downhole fluid end. In the background, on the LHS, is the power end, complete with high pressure iron rigging it to the frac pumpers.

20.5

Blending Equipment

The blender is the heart of the fracturing operation. Although modern blending equipment is often highly automated, the blender operator (or Blender Tender) still retains one of the most critical positions on any location. Figure 20.5a shows a generic schematic diagram of a frac blender.

LIQUID ADDITIVE TANKS

DRY ADD. BIN

PROPPANT SILO

TO FRAC TANKS

SLURRY SIDE FLOW METER RADIOACTIVE DENSIMETER

SUCTION PUMP CLEAN SIDE FLOW METER

DISCHARGE PUMP

DISCHARGE MANIFOLD

SUCTION MANIFOLD

FROM FRAC TANKS

BLENDER TUB

TO HIGH PRESSURE PUMPS

LA METERING PUMPS

RECIRCULATION LINE

Figure 20.5a – Generic flow diagram for a frac blender. Note that on a blender fitted with a Condor tub (such as BJ’s Cyclone I & II blenders), the functions of the blender tub and the discharge pump are combined into a single unit.

Page 216


BJ Services’ Frac Manual 20. Equipment The blender performs the following functions:i) ii) iii) iv) v)

Pre-gelling tanks. Blending liquid and dry additives on the fly. Blending proppant on the fly. Providing supercharge for the high pressure pumps. Metering and recording a variety of job critical parameters.

Figures 20.5b to 20.5e show some of BJ’s fleet of frac blenders.

Figure 20.5b – 125D Frac blender, capable of 125 bpm and 35,000 lbs/min proppant rate

Figure 20.5d – Skid mounted Cyclone blender

Figure 20.5c – Body-load mounted Cyclone II blender, capable of 25 bpm

Figure 20.5e – LFC hydration unit

When pumping a treatment the frac spread can be set up to either gel the frac tanks before the treatment - so that all the fluids are prepared beforehand – or to mix the gel on the fly.

Treatments with Pre-Gelled Tanks When carrying out a treatment with tanks that are pre-gelled, considerable time and effort has been invested into gelling a number of frac tanks filled with water. During this process, the blender will be used to circulate the tanks (via the suction manifold, suction pump, blender tub, discharge pump and recirculation line – see Figure 20.5a), whilst adding the necessary ingredients to produce the required gel. Advantages of Pre-Gelling Tanks:-

Page 217


BJ Services’ Frac Manual 20. Equipment i) ii) iii)

Intense quality control can be carried out on the gel, prior to each tank being accepted. If necessary, a tank or poor quality gel can be rejected, disposed off and then re-blended. Fewer additives need to be mixed on the fly. No need for an LFC Hydration Unit

Disadvantages of Pre-Gelling Tanks:i) Considerable time can be taken up by blending the gel. ii) Gel properties cannot be varied on the fly. iii) Approximately 5% of the gel will be wasted as tank bottoms. iv) Bactericide must be blended with the fracturing fluid to prevent sulphate-reducing bacteria from breaking down the gel.

Mixing Gel on the Fly Mixing the frac gel on the fly requires less pre-job preparation, but involves the use of more equipment and the extra cost of the LFC or XLFC (Liquid Frac Concentrate – see Section 5). LFC is an oil-based slurry of the polymer, usually mixed so that there is 4 lbs of polymer per gallon of slurry. The LFC is added to the water on the fly, allowing the gel to be prepared as it is needed. This requires an LFC hydration unit (see Figure 20.5e). This piece of equipment consists of an LFC storage tank, a metered LFC additive pump (usually progressing cavity type), a hydration tank and a boost pump. Water is supplied to the LFC hydration unit, which meters in the LFC at a controlled ratio, to provide the required gel strength. The hydrating LFC/water mix passes into the hydration tank, which is large enough so that the gel spends 3 to 4 minutes in there, before it is transferred to the blender by the boost pump. This 3 to 4 minute hydration time allows the polymer to hydrate. Some LFC Hydration units are supplied with a QC system – consisting of a viscometer and a pH probe – to provide real time gel QC information. Advantages of Mixing on the Fly:i) No wasted gel. Only the amount of gel required is blended, so that there is no wastage from tank bottoms or if the treatment ends prematurely. ii) Gel properties may be varied on the fly. iii) Less time and effort required for job preparation. iv) No need to use a bactericide. Disadvantages of Pre-Gelling Tanks:i) Extra cost of using LFC, rather than dry powder. ii) Extra cost of LFC Hydration Unit. iii) Loss of gel properties if the LFC Hydration Unit has an equipment problem.

20.6

Proppant Storage and Handling

Proppant has to be stored on location, ready for use. It has to be kept clean and dry, and must be delivered to the blender smoothly and quickly. Figure 20.6a shows frac sand being delivered to the hopper of a blender:-

Figure 20.6a – Frac sand being delivered from a Sand King to the hopper of a blender. Note that there are two blenders in this picture – one is on standby as a backup in case of equipment failure.

Page 218


BJ Services’ Frac Manual 20. Equipment There are two main methods for ensuring the smooth flow of proppant from the storage bin to the blender. The first method is to use a gravity-feed system, which relies on the proppant being stored in a bin which is higher than the blender hopper. A gate valve is used to control the sand rate. This can be done with either large vertically mounted bins (Figure 20.6b) or from a dump truck (Figure 20.6c):-

Figure 20.6b – Vertically mounted, gravity feed proppant bins

Figure 20.6c – Trailer mounted sand dumper

The second method is to use a conveyor system to move the proppant from the bin or dumper, to the blender hopper. This method is typically used on larger frac jobs, as there is usually insufficient space around the blender hopper for all the bins to be positioned. Usually, BJ’s first option for storing large volumes of proppant is the Sand King, as shown in Figure 20.6d:-

Figure 20.6d – BJ Services Sand King

The Sand King is designed to be hauled to location empty, and then filled up with proppant. BJ has two models, one with 250,000 lbs capacity and one with 400,000 lbs capacity. The proppant is held in several separate bins along the length of the Sand King. During the treatment, gates – positioned at the bottom of the hoppers – are opened to allow proppant to fall onto a conveyor. This conveyor runs along the bottom of the entire length of the Sand King, and will transport the proppant to the blender hopper. When a very large treatment is

Page 219


BJ Services’ Frac Manual 20. Equipment planned, such that several Sand Kings have to be used, a separate Sand Belt Conveyor is used, as shown in Figure 20.6e:-

Figure 20.6e – Sand belt conveyor

This device allows several Sand Kings to be placed on either side of the belt, each one feeding onto the main belts of the Sand Belt Conveyor. This, in turn, feed the proppant to the blender hopper. During the treatment, it is important that the proppant system can produce a smooth, uninterrupted flow of proppant to the blender, often at quite high rates. It must also be able to keep the proppant dry, as wet proppant can cause the blender’s proppant screws to seize up.

20.7

Treatment Monitoring

On a modern frac spread, almost every parameter can be measured, displayed and recorded. The place at which this data is displayed and recorded is the Treatment Monitoring Centre, which is usually either a van or a container, as illustrated in Figures 20.7a and 20.7b, below:-

Figure 20.7a – External view of BJ’s Stimulation Van 1800

Page 220



Figure 20.7b – External view of a treatment monitoring container

The fracturing treatment will be controlled from this facility. The Frac Supervisor, the Frac Engineer, the Pump Operator and the Company Man can sit in relative comfort and quiet, making treatment-critical decisions, based on the data that is being collected and displayed.

Figure 20.7c – Two internal views of treatment monitoring vans

Most modern treatment monitoring facilities also include the capability to transmit the treatment data real time back to a specially set up remote data monitoring computer. This can be located either in BJ’s office or in the customer’s. With this facility, Engineers no longer have to waste productive time on location or travelling to and from the location. This is especially significant offshore, where the costs of mobilising personnel can be significant. With the remote data transmission, the Engineers get the same data displayed via similar software (typically JobMaster), with only a second or two delay. Typically, there is also a voice link so that the on-site Engineer can discuss various items or pass on instructions. One other feature of most treatment monitoring containers or vans is a field lab. This will be a compact QC/QA facility, designed to ensure the quality of the fluids and proppants. On larger frac spreads this may even be a separate piece of equipment. Sometimes these are fitted with a fluid rheology and pH flow loop, allowing real time viscosity and pH data to be displayed and recorded.

20.8

The Wellhead Isolation Tool

The Wellhead Isolation Tool (WIT), often referred to as a ”Tree Saver”, is a device that allows treatments to be pumped at a STP higher than the maximum pressure rating of the wellhead. This allows treatments to be pumped at much higher rates than would normally be possible. The WIT does this by completely isolating the wellhead from the treating fluid, as illustrated in Figures 20.8a, b, c, and d. Page 221


BJ Services’ Frac Manual 20. Equipment The tool is used in the following manner:•

• • • • • • • •

Prior to the treatment, the WIT operator obtains data for the type and size of wellhead top flange connection, the distance from the top flange to the tubing hanger, the tubing size and the tubing weight. This allows the WIT operator to assemble the stinger and seal assembly to match the wellhead. The wellhead master valve is closed, and any pressure between the master valve and the top flange is bled off. The WIT is assembled to the top flange, as illustrated in Figure 20.8a. Some WIT’s are fitted with a master valve above the stinger (below the Tee section), whilst others require additional valves to be fitted. The WIT operator applies hydraulic pressure to the lower connection on the master cylinder, to ensure that the tool is fully extended, or stung out of the wellhead. The valves at the top of the WIT are closed and the tool is pressure tested. The wellhead master valve is opened and the WIT is exposed to wellhead pressure. The tool is stroked down by pumping hydraulic fluid into the top connection on the master cylinder. The stinger and the seal assembly are sized so that the seal assembly stings into the top of the tubing, at the point when the stinger is fully stroked into the well. The upper section of the WIT and the master cylinder are clamped together, so that hydraulic pressure is no longer required to keep the tool stung into the tubing.

The WIT tool can be extremely useful, as it can be operated on a live well. This then eliminates the need killing the well and replacing the wellhead. Use of the WIT on a live well is a very specialised process, requiring a trained operator. The tool can be very dangerous if not assembled or operated correctly. The WIT is generally available in two main sizes, big and small. The small size is used for stinging into most tubing sizes, from 2-3/8” up to 4” or larger. The large sized tool is used for stinging directly into casing, with no tubing in the well.

Hydraulic Valve

Master Cylinder Hydraulic Lines Seal Assembly

Wellhead Swab Valve Sub Master Valve

Master Valve

Figure 20.8a – Generic wellhead isolation tool rigged up to wellhead. The WIT is connected to the wellhead via the wellhead’s top flange. At this point the wellhead master and sub master valves are closed, maintaining control of the well and allowing the frac lines and WIT to be pressure tested.

Page 222

Wellhead Isolation Tool

Tubing Hanger



In

Clamp

Hydraulic Fluid Out

Figures 20.8b (left) and 20.8c (right) – Once the WIT has been connected to the wellhead and pressure tested (Fig 20.8a), the next stage is to close the valves of the frac lines (not shown – note that some WIT’s have their own master valves) and open the master and sub master valves on the wellhead. One the wellhead is open, the stinger is stroked down into the top of the tubing by pumping hydraulic fluid into the master cylinder.

Figure 20.8d – Wellhead isolation tool rigged up on location. Note the two 3” frac lines connected to either side, plus the remote actuated 4” plug valve.

Page 223



20.9

The Frac Spread – How it Fits Together

Treatment Monitoring

Blender

Annulus Pump

Frac Pumps

Proppant

Fluid Tanks

LFC Hydration

Frac Pumps

Low Pressure Lines High Pressure Lines Control/Data Cables

Figure 20.9a – Schematic diagram of a frac spread

Figure 20.8a illustrates how all the various components of the frac spread fit together. All frac spreads will basically look like this, although the size and number of components may vary. Some treatments will not use an LFC hydration unit, as the gel will be batch mixed prior to the treatment. Some treatments may use intensifiers, whilst some treatments (“batch” fracs, or Liquid Proppant fracs) may not have separate proppant handling equipment. However, the basic process is the same, no matter what kind of treatment is being performed. Fluid (usually water) is moved from the storage tanks and is usually blended with gelling agents to increase its viscosity. It is then blended with the proppant and pumped down the well. Figures 20.8b to 20.8f, below, show some typical frac spreads:-

Page 224



Figure 20.9b – Large scale treatment, carried out on several low permeability zones simultaneously. Note the number of Sand Kings and frac tanks on location, as well as the use of two blenders (one for backup in case of equipment failure). This frac spread features a separate mobile field lab (bottom left) and a third blender, just for gelling up the tanks and for pumping fluid from the tanks that are located a significant distance from the blender (located just above the bottom left hand row of frac tanks).

Figure 20.9c – The MV Blue Ray, a Gulf of Mexico frac boat, designed primarily for high permeability, frac and pack treatments.

Page 225



Figure 20.9d – Skin Bypass Frac spread, using the “batch” frac method. The two frac pumps are positioned opposite each other, just below the wireline mast (the small read and yellow derrick). A third pump (with “BJ” painted on its roof) is being used as an annulus pump. The two vertical stainless steel tanks on the RHS are for fluid storage. The two batch mixers (each with two round batch tanks - the blue batch mixer is 2 x 50 bbls, whilst the red one is 2 x 40 bbls), used to batch mix the proppant into the gel, are located at the bottom of the picture.

Figure 20.9e – Coiled tubing frac spread. The wellhead is positioned directly below the CT injector (center of picture), with the reel on the RHS. On the LHS are two nitrogen tankers. The main part of the frac spread is positioned behind the injector, with the sand dump truck being the most prominent feature.

Page 226



Figure 20.9f – The MV Thanh Long. This was a boat put together for a single fracturing treatment, for a customer operating offshore Vietnam. The aft deck holds the following equipment:- 4 x 1200 HHP frac pumps, Cyclone II blender, 2 x 640 cu ft proppant bins, treatment monitoring container c/w field lab, 4 x 165 bbls tanks and a 100 bbl vertical tank.

References Standard Practices Manual, BJ Services, January 2001 onwards Corporate Safety Standards and Procedures Manual, BJ Services, January 2001 onwards Equipment and Technology Catalogue, BJ Services, 1990 onwards Bradley, H.B. (Ed): Petroleum Engineers Handbook, SPE, Richardson, Texas (1987) Economides, M.J., and Nolte, K.G.: Reservoir Stimulation, Schlumberger Educational Services, 1987. Gidley, J.L., et al: Recent Advances in Hydraulic Fracturing, Monograph Series Vol 12, SPE, Richardson, Texas (1989).

Page 227


BJ Services’ Frac Manual 21. Designing Wells for Fracturing

21.

Designing Wells for Fracturing

The single biggest influence on the feasibility of the hydraulic fracturing process is the design of the well, including its completion and perforations. The influence of perforations and how they can be designed to maximise the effectiveness of hydraulic fracturing, has been discussed in Section 14. In this section, we will discuss the philosophy and impact on well planning and design of the hydraulic fracturing process. On a wider scale, we shall discuss the influence hydraulic fracturing can have on field development, whilst on the smaller scale we shall discuss how to plan individual wells for fracturing.

21.1

How Many Wells do I Need to Drill?

The answer is, not nearly as many as you think. Very few operating companies outside of North America plan a field development with stimulation in mind. Hydraulic fracturing is the most effective form of stimulation, but it is also the type is most often restricted by the design of a well. Fracturing is often perceived by Engineers who do not have first hand experience with the process, as a high risk operation. Consequently, the Engineers who design the development of a field are either not aware of the benefits of fracturing, or not aware of the chances of a successful treatment. If a well is planned with hydraulic fracturing in mind, it is relatively realistic to expect at least double the production from the treated well, compared to the untreated well. In many cases, fracturing will produce a production increase significantly greater than this. In addition, often production targets can be met at significantly lower drawdowns, which can have a tremendous impact on reservoir management and can often prevent or significantly delay the onset of water production from a WOC or gas production from a gas cap. So if an operating company can produce at least twice as much oil from a given well, what does this mean for reservoir development plans? It means that the operating company needs to drill fewer wells, which can result in tremendous cost savings - especially offshore, where the need for fewer wells may even eliminate the need for entire platforms. Obviously, in highly faulted reservoirs, each “pool” will need at least one well, but in reservoirs that would ordinarily require several wells, it is not unreasonable to expect to eliminate up to half of these. Injection wells can also be fractured very effectively. An additional benefit to fracturing is that each zone in an injection well can be individually treated, allowing a specific fracture, of a specific conductivity, to be placed in each zone. This allows the Reservoir Engineer to custom design the injectivity profile of an injection well, to meet the requirements of long term pressure maintenance. Traditionally, the only sector of the industry that has a profound understanding of what can be achieved by fracturing, is the tight gas sector. Most tight gas wells have to be fractured otherwise they would not be economic. In a lot of cases, these wells have to be fractured or they would not produce at all. In these areas, the tight gas operating companies are totally dependent upon the hydraulic fracturing process for the success or failure of their field developments. Yet companies keep drilling wells, keep developing tight gas fields and keep fracturing them – so the process must be successful. If it works for tight gas wells, why not for oil wells or even high permeability gas wells? After all, the basic process is the same, the equipment is the same, the proppants are the same and the fluids are the same. The only thing that varies from well to well is the amount of each of these items we use and the relative quantities in which they are used. Obviously, the potential percentage production increase from fracturing a tight gas well is much greater than for fracturing a high permeability oil well. However, which generates the most revenue – Page 228


BJ Services’ Frac Manual 21. Designing Wells for Fracturing increasing the production from a tight gas well from 50 mscfpd to 500 mscfpd, or increasing the production of an oil well from 5,000 bopd to 10,000 bopd? Both of these production increases are realistically achievable.

21.2

The Best Wells are also the Best Candidates for Fracturing

Too often, hydraulic fracturing is seen as a last-try-process, used because the company has a bad well and needs to do something with it. Unfortunately, in most circumstances, hydraulic fracturing cannot turn a bad well into a good well, unless the only reason for the low production is a large skin factor. In all cases, the reservoir must have some potential in order for the full benefits of the fracturing process to be realised. In the late 1980’s, a company operating in the Danish sector of the North Sea, began developing a new field. The oil was held in the highly same highly productive chalk formations, which were responsible for the huge Ekofisk development, just across the border in the Norwegian sector. The traditional way to develop these reservoirs was to drill deviated or S-shaped wells through the chalks and then perform an acid frac. However, the operating company – and its partners (which included some major US operating companies) - realised that this may not be the best method. Over a series of wells and a number of years, the operating company perfected a method for developing their reservoirs that involved drilling long horizontal wells, each of which would have between 8 and 15 fracs placed along its length, depending upon the length of the productive section. Each of the horizontal liners was cemented in place – a bold new approach in itself – and selectively perforated to control the point of fracture initiation (see Section 13). These wells were also fitted with a special completion, which allowed individual access to each of these perforated intervals. Then, over a period of 4 to 8 weeks, each of these zones would be hydraulically fractured. As time progressed and the technology improved, this time decreased, but still took weeks, rather than days, to frac each well. In one well, the company successfully pumped over 13 million lbs of proppant, a record for a well that has only recently been passed. How much did this cost? A lot. Each well drilled and completed in this fashion typically cost 3 times what a conventional well would cost, in a part of the world where drilling costs were already huge. However, each well was also producing between 4 and 6 times what the typical conventional well was producing. In addition, the conventional acid fractured wells had to have the acid fracture repeated every 18 months to 2 years, as the highly plastic chalk formations slowly deformed into the fractures. However, this was not the case with the propped fractures, resulting in greatly reduced future expenditure. The point of this story is that good wells are the best candidates for fracturing. The industry should not be limited to remedial and low-productivity applications. When selecting candidates for fracturing look for the good wells first.

21.3

Designing Wells for Fracturing

The best time to fracture a well is right after it has been drilled and cased – before the completion has been run. This is another reason why it is important to consider the implications of fracturing whilst planning the well. In general, completions act to restrict what can be done with a treatment, and can often eliminate the fracturing option entirely. Completions can limit fracturing operations for the following reasons:i)

Pressure limitations. Fractures are created by pressure, and as a result abnormally high pressures can be generated by the treatment. Often, completions are not designed to withstand this loading. Although it is often possible to reduce this effect

Page 229


BJ Services’ Frac Manual 21. Designing Wells for Fracturing by placing pressure on the annulus, many are completed with two or more packers, eliminating the effectiveness of annulus pressure. ii)

Temperature limitations. The pumping of a cool frac fluid will cause the completion to shrink. Sometimes, the completion can shrink so much that the tubing can sting out of packers. The effect of the extra pressure acts to make this effect even worse.

iii)

Completion jewelry. Items such as sub-surface safety valves, gas lift mandrels and sliding side doors can often take significantly less differential pressure than the actual completion itself.

It should be noted that the above three limitations can be eliminated by the use of coiled tubing in the fracturing process. iv)

Multiple zones. Often, wells are completed with multiple sets of perforations. Whilst it is possible to treat multiple zones at the same time, it is generally a much more complex process, which requires more equipment and more materials (treating two identical zones requires twice the pump rate, and twice the volume of proppants and fluids. It may require significantly more than twice the hydraulic horsepower, as the friction pressure will rise by significantly more than this factor).

In short, if the well can be fractured before it is completed, all the limitations imposed by the completion can be eliminated. However, doing this requires a degree of forward planning, faith in the fracturing process and increased up-front expenditure. Fracturing before completion allows the perforate-stimulate-isolate method to be employed:1. Perforate

The individual zone is perforated, allowing each zone to be fractured with the optimum treatment. By carefully positioning the perforations, the point of fracture initiation can be controlled..

2. Stimulate

The fracture treatment is pumped either down the casing or through a frac string.

3. Isolate

The zone is isolated by setting either a sand fill or a bridge plug.

Repeat steps 1 to 3 as often as necessary, moving from the bottom to the top of the well. Obviously, this process can take a lot longer than the conventional drill and complete process. However, the extra cost is more than offset by the substantially increased production from these wells. If the well cannot be fractured prior to completion, then the completion should be designed with fracturing as a potential scenario. Packers and tubing jewelry should be designed to withstand the pressures of fracturing. Seal assemblies should be long enough to cope with the cooldown. Zones should be as isolated as possible. Of course, all this requires substantial extra investment, which has to be justified purely on the basis of faith in the fracturing process. However, in case after case, field development after field development, this initial expenditure has proved its worth.

References Nagel, W.B, et al.: “An Integrated Team Approach for Improving Company-Wide Stimulation Design and Quality Control”, paper SPE 26142, presented at the SPE Gas Technology Symposium, Calgary, June 1993.

Page 230


BJ Services’ Frac Manual 21. Designing Wells for Fracturing Cipolla, C.L., Bernsten, B.A., Moos, H., Ginty, W.R., and Jensen, L.: “Case Study of Hydraulic Fracture Completions in Horizontal Wells, South Arne Field Danish North Sea”, paper SPE 64383, presented at the SPE Asia-Pacific Oil and Gas Conference and Exhibition, Brisbane, October 2000. Owens, K.A., Pitts, M.J., Klampferer, H.J., and Kreuger, S.B.: “Practical Considerations for Well Fracturing in the ‘Danish Chalk’”, paper SPE 25058, presented at the SPE European Petroleum Conference, Cannes, France, November 1992. Schubarth, S.K., Yeager, R.R., and Murphy, D.W.: “Advanced Fracturing and Reservoir Description Techniques Improves Fracturing in......”, paper SPE 39777, presented at the SPE Permian Oil Basin Oil and Gas Recovery Conference, Midland TX, 1998 Voneiff, G.W., and Holditch, S.A.: “An Economic Assessment of Applying of Applying Recent Advances in Fracturing Technology to Six Tight Gas Formations”, paper SPE 24888, presented at the SPE Annual Technical Conference and Exhibition, Washington DC, October 1992. Stewart, B.R, et al.: “Economic Justification for Fracturing Moderate to High Permeability Formations in Sand Control Environments”, paper SPE 30470, presented at the SPE Annual Technical Conference and Exhibition, Dallas, October 1995. Conway, M.W., et al.: “Expanding Recoverable Reserves Through Refracturing”, paper SPE 14376, presented at the SPE Annual Technical Conference and Exhibition, Las Vegas, October 1985. Church, D.C., and Peters, B.A.: “Improved Fracturing Technique Yields Increased Production Potential”, paper SPE 17045, presented at the SPE Eastern Regional Meeting, Pittsburgh, October 1997

Page 231


BJ Services’ Frac Manual 22. The Fracture Treatment

22.

The Fracture Treatment: From Start to Finish

22.1

Frac Job Flow Chart Obtain Well data: Logs, DST’s, Mud Logs, Production History (if any), PVT Data, Completion Diagram, Previous Treatments

Use Nodal Analysis or Similar

History Match Production Data

IN OFFICE

Establish Base Case Production

Input Speculative Fracture Geometry into Production Simulator

Run Production Simulation with Fracture

Optimum Fracture Geometry?

No

Yes Design Treatment for Optimum Fracture Geometry Using Fracture Simulator

Preliminary Treatment Schedule Page 232

1

SRT Schedule Minifrac Schedule



1

Mobilise Equipment, Materials and Personnel

Rig Up, Mix Fluids, Pressure Test

Pre-Job Safety Meeting

ON LOCATION

Pump Step Rate Test (Step Up and Step Down)

Real Time Data Modelling

Analyze SRT Data Fracture Extension Pressure, Near Wellbore Friction

Is NWF Significant? No

Pump Minifrac

Yes

Pump Minifrac with Proppant Slugs


No

Pressure Rise due to Prop. Slugs?

Yes

Pump Proppant Slugs as per SPE 25892 2

Page 233



2

Pressure Match Simulator Output to Minifrac Data

Re-Design Treatment Final Treatment Design

E, ν, Klc, Cl,ll, lll Pnet, Pclosure, ηfrac

Load Proppant & Additives. Mix Fluids

ON LOCATION

Pre-Job Safety Meeting

Pump Treatment

No

Monitor Pressure until Fracture Closure


Premature Screenout? Yes

Shut in Well & Bleed Off Pressure

ON LOCATION OR IN OFFICE

Wait for Fluid Samples to Break

Flow Back Well

Rig Down

Analyze Treatment Data

Post-Job Report

Figure 22.1a – Frac job process flow diagram

Page 234


BJ Services’ Frac Manual 22. The Fracture Treatment The design and execution of a frac job can be broken down into 5 major steps:-

1.

Data Collection

Collect as much data as possible on the well, and on treatments carried out on offset wells. This data includes, but is not limited to:i)

ii)

iii) iv)

v) vi) vii) viii) ix) x)

xi)

xii)

xiii)

Wireline logs. Useful for spotting boundaries between formations, high and low permeability and porosity, and also for spotting fluid contacts. Specialised logs can also give dynamic Young’s modulus and Poisson’s ratio, stresses and the quality of the cement bond. Get summary or evaluated logs whenever possible – there is no point in doing a full log analysis when somebody else has already done this. Also – if you are not confident with logs - a good first step is to mark where the perforations are, as these will be the productive intervals. Well test data. Useful for obtaining values such as reservoir pressure, permeability and skin factor. Again, get the report with the analysis already done. No one will expect you to be an expert well test analyst. These reports may also contain calculated data for porosity, viscosity, fluid saturation and compressibility. Completion diagram. Essential, as this will contain all the details you will need on the perforations, depth and sizes of tubing and casing strings etc. Wellhead diagram. Usually, all the Frac Engineer needs from this is a description of the top connection, so that the crew can have the appropriate crossover when they rig up to the wellhead. However, if a wellhead isolation tool is being used, a detailed diagram will be required. Deviation survey. If the well is not vertical, the Frac Engineer will need to know MVD vs TVD for all formations, perforations and tubulars. Core data. If the well has been cored, this report may contain useful data on porosity, permeability and fluid saturation. In addition, the report may contain rock mechanical data and mineralogy (useful if the formation is suspected to be “water-sensitive”). Core samples. If core samples are available, get hold of them and have them tested for Young’s modulus and Poisson’s ratio. Reservoir fluid samples. It is important to carry out compatibility testing between the frac fluid and the reservoir fluids. Problems are rare, but when they do occur they can ruin a well. Production data. Production data is useful for two reasons. First, this data is the basis for post treatment production forecasts. Secondly, a qualitative analysis should be performed to check for items such as water or gas coning and fines migration. Produced sand samples. Essential if a frac and pack treatment is being designed, as a sieve analysis will be required to find the correct proppant size. However, getting a representative sample can be difficult. Surface samples tend to have a higher proportion of fines, as these are more easily carried out of the well. Bottom hole samples tend to be the other way around – high proportions of the fines have been carried away out of the well. Offset treatment data. Often, this is the most important and reliable source of data. Perform a complete analysis of these treatments, including a pressure match, if the data is available. If the data is reliable enough, this may even eliminate the need for a minifrac and step rate test. Location diagram. The Frac Engineer needs to know what size the location is, to ensure that all the equipment can be placed. If not, a smaller treatment needs to be designed. Especially important offshore, where additional factors such as crane maximum lift and deck loading must also be considered. Other information, such as production logs (i.e. spinner surveys), temperature logs, caliper logs, mud logs, stress surveys, core flow testing, workover reports and drilling records can all provide useful information.

Page 235



2.

Preliminary Design

This stage uses all the data gathered in step 1 to produce a preliminary frac design. The initial step is to analyse the reservoir and production data and derive the optimum fracture geometry required. This step is best accomplished using nodal analysis. Then the fracture simulator is used to design a treatment to produce this fracture. Often, this design has to be tempered by considerations such as cost, mobilisation and equipment availability, so that the Engineer may go back and forth between the nodal analysis and the simulator several times. Unless the Engineer has good data from offset treatments, a step rate test and a minifrac will be required. The step rate test is pretty much standard for any well and an example is included below. The minifrac needs to be designed on a well by well basis. It should be pumped at the same rate as the preliminary frac design, using the same fluid and then displaced at the same rate using slick water. The volume of the minifrac should be at least equal to the anticipated pad volume. The minifrac fluid volume should be large enough to contact every formation that the actual frac will contact. This means that for tip screen out designs, the minifrac should be the same size as the pad, whereas for tight gas fracturing it must be considerably larger. Remember – it is much better to pump too much fluid than too little. The minifrac is exactly what its name suggests – a small frac. In fact, it should be as close as possible to the actual treatment, in order to produce data as relevant as possible. Remember that if minifrac and step rate tests are being performed, there is no point in doing too detailed a design at this stage. The real design work will be done on location after these calibration tests. At this stage, what is required are reasonable estimates for the expected production increase, the quantity of materials and equipment that must be mobilised and the cost of the treatment. Preliminary design work also includes designing the frac fluid. This often involves the use of Fann 50 (or similar) HPHT rheometers in order to ensure that the frac fluid has the right combination of stability and break.

3.

Calibration Tests and Redesign

Finally, the frac spread and crew gets mobilised and is rigged up on location. The next major step in the design and execution process is to perform the calibration tests (minifrac and step rate test). It is vitally important to get good data from these. Whenever possible, get bottom hole pressure data, either from a gauge or from a dead string. For the step rate test, remember the following points:i) ii)

iii) iv)

Get as many low rate steps as possible. Ideally, this means 4 steps below 2 bpm, although this is not always easy with big frac pumps. However, the more steps that can be taken before the frac starts to initiate, the better the results will be. Don’t fiddle with the rate. When moving from one step to another, change the rate and then leave it alone. Getting a stabilised pressure is difficult enough without someone fiddling with the throttles. As long as the rate is approximately what it should be, that is good enough. Use the step rate test procedure as a guideline only, especially with regard to volumes. Getting a stabilised rate and pressure for each step is what we are after. Once this has been achieved, move on to the next step. It is important that the step rate test (step up variety) is performed on an unfractured formation. So either do the step rate test before the minifrac, or wait for a significant period of time after the minifrac is finished.

Page 236


BJ Services’ Frac Manual 22. The Fracture Treatment v) vi)

The opposite is true for the step rate test. An open fracture is needed before the step down begins, and the fracture must be open throughout the entire test. It is common to combine the two tests – step up and then step down again. Remember that the well must be full of fluid before the step rate test commences. If the well has to be filled up, do it at low rate to ensure no fracture forms.

For a minifrac, the following points are important:i) ii) iii)

iv)

Keep the rate constant, even if this means pumping at a different rate than programmed. This makes the analysis easier and more reliable. Keep the fluid quality constant, again to make the analysis easier and more reliable. If necessary, gel up a couple of frac tanks, rather than mixing on the fly. Understand the wellbore fluid. Know its fluid properties and it’s volume. Remember that this fluid will be injected into the fracture ahead of your carefully prepared fracturing fluid. So if you don’t know the wellbore fluid, the careful preparation of the frac fluid is wasted. If necessary, circulate the well to completion fluid, or something similar before pumping the minifrac. Monitor the pressure decline. During this period, don’t let the frac crew do anything, except drink coffee. It is all too easy for a silly mistake to ruin data collection. The Frac Engineer can also do his part by zero-ing out the rate on the fracturing monitoring computer – so that any fluid pumping by the blender does not show up as an erroneous downhole rate. Remember also to collect data for long enough – if data collection stops before closure (or closures) has happened, then the minifrac will have lost at least half of its value.

Finally, don’t forget the primary objective of the exercise – to produce a good frac design. Other objectives – such as minimising rig time or trying to get the job in the ground before nightfall – are desirable, but secondary. The customer should be aware of the fact that a redesign can sometimes take several hours.

4.

Job Execution

After all the planning and preparation has taken place, the actual treatment can sometimes take a surprisingly short period of time. During this period, the fate of the treatment no longer rests in the hands of the Frac Engineer. It is now up to the Supervisor and the rest of the frac crew to put the job in the ground as closely as possible to the revised treatment design. Of course, on longer treatments, real-time redesign may be performed. In which case, the Frac Engineer may still have some influence on the treatment. However, usually it is time for the Engineer to sit back and let the crew get on with their job. Some Frac Engineers like to run the monitoring computer or check the fluid samples – both these occupations are useful and need to be performed. It is also important that the Frac Engineer stays in close contact with the Supervisor, just in case something unexpected happens.

5.

Post Treatment Analysis

There is no such thing as the perfect frac job. Every job has room for improvement, however slight. This applies to the Frac Engineer’s job as well and the post treatment analysis is the way to find out what could have been done better. Post treatment analysis comes in two parts:i)

Analysing the pressure and rate data from the job. The best way to do this is with a pressure match, although don’t spend too much time on this if you have no downhole pressure data. Results obtained from this will improve the success rate of future treatments.

Page 237


BJ Services’ Frac Manual 22. The Fracture Treatment ii)

Assessing the production increase. Sometimes it is easy to loose sight of the objective of the entire process – to increase production. It is vitally important to keep track of the production of fractured wells. Remember that production over the first few days doesn’t really count – we should be looking at the stabilised production several weeks after the treatment is performed. If production does not meet or exceed expectations, then the following three questions must be satisfied; Was the well a good candidate (i.e. reserves and pressure)? Was the optimum fracture placed in the formation? And were the post treatment expectations realistic?

22.2

Example Treatment Schedules

Whilst BJ Services is not at liberty to publish confidential data, the example treatments were actually pumped and all of them produced significant production increases for our customers. These designs are included so that the reader can gain some idea of the size and scale of fracturing treatments. However, remember that each treatment must be designed individually for each well – these schedules are for guidance only and are not meant as “ready-to-use” frac designs.

Typical Step Rate Test Schedule Rate Time bpm secs 0.7 120 + 1.0 30 1.5 30 2.0 30 3.0 30 5.0 30 7.0 30 9.0 30 11.0 30 15.0 30 12.0 15 9.0 15 6.0 15 3.0 15 Total Volume (gals)

Volume gals 15 21 32 42 63 105 147 189 231 315 126 95 63 32 1476

The maximum rate can be raised if desired, but this will probably not be necessary for most treatments. However, remember to hold the maximum rate for a few minutes to ensure that the fracture is of sufficient volume. If the fracture is too small, it may close before the step down portion can be completed.

Tight Gas Fracturing Stage 1 (Pad) 2 3 4 5 6 (Flush) Page 238

Fluid Type Linear XLink XLink XLink XLink Sl/Water

Rate bpm 40 40 40 40 40 40

Clean Vol gals 20,000 20,000 40,000 40,000 100,000 8,300


Prop. Conc. ppa 0 1 2 3 4 0

BJ Services’ Frac Manual 22. The Fracture Treatment Notes

Linear XLink Sl/Water

= Linear gel (i.e. base gel with no crosslinker) = Crosslinked gel = Slick Water

Linear gel

2.5 gpt VSP 3 20,000 gals (75.7 m ) total

Crosslinked gel

Vistar 20 3 200,000 gals (757 m ) total

Slick Water

2 gpt VSP 3 8,300 gals (31.4 m ) total

Proppant

20/40 CarboLite 600,000 lbs (272 tonnes) total

Treating Pressure

5,100 psi (352 bar, 35.2 MPa)

Pumping Capacity

+/- 5,000 HHP (3,730 kW)

Frac and Pack Stage

Fluid Type XLink XLink XLink XLink XLink XLink XLink XLink Sl/Water

1 (Pad) 2 3 4 5 6 7 8 9 (Flush)

Rate bpm 15 15 15 15 15 15 15 15 15

Clean Vol gals 1,200 1,250 600 800 1,000 1,250 1,850 2,000 7,830

Crosslinked gel

35 ppt Viking 1D 3 9,950 gals (37.6 m ) total

Slick Water

4 gpt XLFC-1 in CaCl2 brine 3 7,830 gals (29.6 m ) total

Proppant

20/40 EconoProp 70,000 lbs (31.8 tonnes) total

Treating Pressure

6,300 psi (maximum) (434 bar, 43.4 MPa)

Pumping Capacity

+/- 2,300 HHP (1,716 kW)

.

Page 239


Prop. Conc. ppa 0 1 3 5 7 9 11 12 0


Skin Bypass Fracturing Stage

Fluid Type XLink XLink XLink XLink Sl/Water

1 (Pad) 2 3 4 5 (Flush)

Rate bpm 8 8 8 8 8

Clean Vol gals 4,000 1,300 1,100 1,050 2,540

Crosslinked gel

SpectraFrac G 3500 HT 3 7,500 gals (28.4 m ) total

Slick Water

4 gpt XLFC-1 3 2,540 gals (9.6 m ) total

Proppant

20/40 CarboLite 16,500 lbs (8.1 tonnes) total

Treating Pressure

900 psi (62.0 bar, 6.2 MPa)

Pumping Capacity

+/- 200 HHP (149 kW)

Prop. Conc. ppa 0 2 5 8 0

References Howard, G.C., and Fast, C.R.: Hydraulic Fracturing, Monograph Series Vol 2, SPE, Dallas, Texas (1970). Gidley , J.L., et al.: Recent Advances in Hydraulic Fracturing, Monograph Series Vol 12, SPE, Richardson, Texas (1989). Economides, M.J., and Nolte, K.G.: Reservoir Stimulation, Schlumberger Educational Services, 1987.

Page 240


BJ Services’ Frac Manual Nomenclature

Nomenclature a

=

A Af

= =

AR Bg Bo BHA BHP BHTP c C CI CII CIII cb Cc CD Ceff cf CfD cr ct Cv Cw d dp DCF E E’ Ed f F Fc Fcd 9 g(∆tD) g(∆tcD) G

= = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =

G(∆tD) = Gc = Glc Gd gf GOR GLR h hf H HD HH HHP Page 241

= = = = = = = = = = =

fracture half length (Griffith crack) or variable used in Nolte G time analysis. Area, annular capacity Total area of fracture (usually both wings, but for a single wing in Nolte G-Function analysis). aspect ratio gas formation volume factor oil formation volume factor bottom hole assembly bottom hole pressure bottom hole treating pressure total reservoir compressibility (also called ct) wellbore storage coefficient viscosity controlled leakoff coefficient compressibility controlled leakoff coefficient wall building controlled leakoff coefficient bulk reservoir compressibility (i.e. with porosity) compressibility controlled leakoff coefficient dimensionless wellbore storage coefficient effective or combined leakoff coefficient fracture compliance, formation compressibility dimensionless fracture conductivity (new API notation) zero porosity reservoir compressibility (i.e. rock compressibility) total reservoir compressibility (also called c) viscosity controlled leakoff coefficient wall building controlled leakoff coefficient diameter, diameter of plastic zone at fracture tip proppant grain diameter discount factor Young’s modulus plane strain Young’s modulus dynamic Young’s modulus Fanning friction factor force fracture conductivity dimensionless fracture conductivity (old – now CfD) 2 2 acceleration due to gravity (= 9.81 m/s or 32.18 ft/s ) dimensionless loss-volume function (Nolte minifrac analysis) g(∆tD) at fracture closure shear modulus, or elastic energy release rate Nolte G time critical elastic energy release rate or Nolte G time at fracture closure critical elastic energy release rate – failure mode l dynamic shear modulus frac gradient gas oil ratio gas liquids ratio height fracture height at wellbore depth, fracture height at wellbore dimensionless fracture height hydrostatic head hydraulic horsepower Version 1.0 June 2005 Uncontrolled Copy – DO NOT DISTRIBUTE

BJ Services’ Frac Manual Nomenclature Hξ ISDP ISIP IPR J J0 Jf K K’ K’’ Kl K1c

= = = = = = = = = = = =

k Kd kf kp kr KZD L m m(P) N n’ n’’ Np NPV NRe p, P P* P’ Pb Pclosure Pext Pfinal Pfrict Pi Pinitial Pm Pnet Pnwb Pob Pperf Pr Pr, t Pres Pv Pwb PwD PwDM Pwf Pws PC PKN q Q QL Qmax R rd

= = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =

Page 242

characteristic length (MFrac) instantaneous shut-down pressure (= ISIP) instantaneous shut-in pressure inflow performance relationship original (pre-stimulation) productivity index with skin undamaged productivity index post-fracturing productivity index bulk modulus power law consistency index Herschel-Buckley consistency index stress intensity factor, failure mode 1 critical stress intensity factor – failure mode 1, or fracture toughness permeability dynamic bulk modulus formation permeability, permeability to frac fluid filtrate proppant permeability permeability to reservoir fluid Kristianovich, Zheltov, Daneshy – 2 dimensional frac model tubing of casing length, fracture half length (also xf) mobility, gradient of curve real gas pseudo-pressure viscometer spring factor power law exponent Herschel-Buckley exponent dimensionless proppant number (or simply proppant number) net present value Reynold’s number pressure average reservoir pressure from well test analysis pressure derivative breakdown pressure closure pressure or Pc fracture extension pressure post-frac surface circulation pressure (frac and packs) friction pressure (usually ∆Pfrict) static reservoir pressure pre-frac surface circulation pressure (frac and packs) match pressure (Nolte minifrac analysis) net pressure near wellbore friction pressure (usually ∆Pnwb) pressure due to overburden perforation friction pressure (usually ∆Pperf) pressure at a distance r from the wellbore pressure at a distance r from the wellbore, after a time t. reservoir pressure (also Pi) plastic viscosity (Bingham plastic fluids) wellbore pressure (usually bottom hole) dimensionless wellbore pressure dimensionless wellbore match pressure flowing wellbore pressure static wellbore pressure proppant concentration Perkins, Kern, Nordgren – 2 dimensional frac model pump rate, average pump rate, liquid flow rate pump rate, average pump rate, gas flow rate fluid leakoff rate maximum pump rate frac radius (esp. radial model) radius of investigation or disturbed radius Version 1.0 June 2005 Uncontrolled Copy – DO NOT DISTRIBUTE

BJ Services’ Frac Manual Nomenclature re rp

= =

rw r w’ S Sp SG SGf SGp STP t tD tDM tDx

= = = = = = = = = = = =

reservoir radial extent radius of plastic zone at fracture tip or ratio of fracture area in permeable formation over total fracture area (i.e. net to gross fracture area ratio) for 2-D fracture models. wellbore radius effective wellbore radius skin factor spurt loss coefficient specific gravity specific gravity, fluid specific gravity, proppant surface treating pressure time dimensionless time dimensionless match time fractured well dimensionless time

tDx M

=

fractured well dimensionless match time

tHorner tma tp tsma ts T TVD U Ufluid U ˙ v vprop V Vi Vs W, w W ¯ ,w ¯ Wmax WOR x, y, z xe xf xfD Yp z zi

= = = = = = = = = = = = = = = = = = = = = = = = = =

Horner time rock matrix compression wave transit time pump time, producing time, compression wave transit time rock matrix shear wave transit time shut in time (also ∆t), shear wave transit time tensile strength, or temperature true vertical depth energy energy in the fracturing fluid energy per unit time, work, horsepower velocity fraction of fracture volume occupied by proppant volume total volume injected into fracture spurt loss volume fracture width average fracture width maximum fracture width water oil ratio mutually perpendicular directions, distances 2 length and width of a square reservoir (such that area = xe ) fracture half length dimensionless fracture half length ( = xf/re) yield point (Bingham plastic fluids) gas z-factor gas z-factor at static reservoir conditions

α β

= =

βs γ

= =

poroelastic constant (Biot), Nolte analysis boundary variables flow capacity factor (Forcheimer Equation) or shape factor (LEFM) ratio of average to wellbore net pressures (Nolte minifrac analysis) shear rate, proppant specific gravity, or shape factor (MFrac) real gas pseudo-pressure differential real gas pseudo-pressure differential match pressure differential, drawdown drawdown (= Pi – Pwf) build-up pressure ( = Pi – Pws) pressure drop due to fluid tubing friction test data log-log plot match pressure

f f

∆m(P) = ∆m(P)M = ∆P = ∆Pdrawdown = ∆Pbuild-up = ∆Pfrict = ∆PM = Page 243


BJ Services’ Frac Manual Nomenclature ∆Pnwb ∆Pperf ∆Pskin ∆t ∆tD ∆tcD ∆tM

ε ε1 ε2,3 εx, y, z η θ µ µi µapp µf µr ν νd π ρ ρb ρgel ρp ρsl σ σ1,2,3 σc σH σH, max σH, min σv σxx, yy, zz σy τ τ’o φ φp ω

Page 244

= = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =

pressure drop due to near wellbore friction pressure drop due to perforations pressure loss due to skin damage change in time, time since shut-in or shut down. change in dimensionless time a.k.a. delta Nolte time delta Nolte time at closure test data log-log plot match time strain strain in the vertical direction strains due to the principle horizontal stresses, σ2 and σ3 strain in the x-, y- and z-directions fluid efficiency, fracture efficiency angle, viscometer dial reading viscosity viscosity at static reservoir conditions apparent viscosity viscosity of frac fluid filtrate viscosity of reservoir fluid Poisson’s ratio dynamic Poisson’s ratio Pi, the ratio of a circle’s circumference to it’s radius (= 3.1415926....) density proppant bulk density, formation bulk density gel or base fluid density proppant absolute density slurry density stress principle (i.e. mutually perpendicular) stresses critical stress horizontal stress maximum horizontal stress minimum horizontal stress vertical stress principle stresses in the x-, y- and z- directions yield stress shear stress initial or threshold shear stress (Herschel-Buckley fluids) porosity proppant bulk porosity length of unwetted part of fracture (FracPro, FracproPT), angular velocity, viscometer rotor speed


BJ Services’ Frac Manual Index

Index A Absolute volume ..................................................................................................................... 178 Additives .......................................................................................... see Fluid systems, additives Always look.............................................................................................. on the bright side of life Aluminates ..........................................................................................................see Crosslinkers Aspect ratio................................................................................................................................. 7

B Bacteria..................................................................................................................................... 41 Beta-factor ................................................................................................................................ 88 Bilinear flow ............................................................................................................................ 202 Binary foam fracturing .............................................................................................................. 39 Bingham plastic fluids .......................................................................................................... 20-21 Biocides, bactericides ............................................................................................................... 41 Biot’s constant ........................................................................................................ 58, 62, 64, 67 Blenders. blending equipment ......................................................................................... 216-218 Borates ...............................................................................................................see Crosslinkers Breakers ................................................................................................................................... 41 Brines............................................................................................................................. 35-36, 51 Brittle fracture ........................................................................................................................... 75 Buffers ...................................................................................................................................... 41 Bulk modulus ............................................................................................................................ 57 Dynamic....................................................................................................................... 64

C Calibration tests ............................................................................................................... 115-120 Candidate selection ................................................................................................. 101-108, 229 Cement bond .......................................................................................................................... 108 Circulation tests ...................................................................................................................... 168 Clay control.......................................................................................................................... 43-44 Cleats................................................................................................................................ 16, 170 Closure stress........................................................................................................................... 47 CO2 fracturing .................................................................................................................. 39, 169 Coflexip high pressure hoses ................................................................................................. 210 Completions..................................................................................................... 104-106, 229, 230 Jewelry........................................................................................................ 105-106, 230 Compressibility ....................................................................................................................... 202 Average reservoir .......................................................................................... 10, 64, 202 Formation......................................................................................................... 8, 64, 190 Compression wave ................................................................................................................... 63 Conductivity Finite ........................................................................................................... 194, 203-205 Fracture .................................................................................. see Fracture, conductivity Infinite ................................................................................................. 165, 194, 203-205 Corrosion, tubulars ................................................................................................................. 108 Crack driving force.................................................................................................................... 73 Crack tip dilatency ............................................................................................................... 75-76 Crack tip plasticity................................................................................................................ 76-77 Critical energy release rate....................................................................................................... 73 Critical fracture length............................................................................................................... 73 Critical micellar concentration................................................................................................... 35 Critical stress intensity factor.................................................................................................... 74 Constant external phase........................................................................................................... 37 Page 245


BJ Services’ Frac Manual Index Constant internal phase............................................................................................................ 37 Crosslinked fluid systems ...................................... see Fluid systems, water-based, crosslinked Crosslinkers ................................................................................................................... 31-33, 41 Aluminates .............................................................................................................. 31-32 Borates ................................................................................................................... 31-32 Borates, exotic ........................................................................................................ 31-32 Titanates ................................................................................................................. 31-32 Zirconates ......................................................................................................... 31-32, 33

D Darcy’s Equation..................................................................................................... 9, 88, 95, 194 Darcy’s law ................................................................................................... 2, 88, 109, 116, 126 data collection......................................................................................................................... 235 Data frac ....................................................................................................................see minifrac Dead string ............................................................................................................................. 121 Decline curve analysis ..................................................................................................... 125-131 Density...................................................................................................................................... 19 Bulk, formation............................................................................................................. 63 Bulk, formation, log-derived......................................................................................... 63 Slurry, measurement of ..................................................................................... 177, 178 Densometers .......................................................................................................... 176, 178, 179 Mass flowmeter density measurement...................................................................... 177 Nuclear ...................................................................................................................... 178 Derivative plots, minifrac ........................................................................................................ 131 Derivative plots, well testing ............................................................................................ 197-198 Desorption ................................................................................................................................ 16 Discounted revenue................................................................................................................ 102 Dilatency ................................................................................................................................... 75 Dilatency contribution ............................................................................................................... 75 Dipole sonic logs.......................................................................................................... 63-67, 189

E Economics ....................................................................................................................... 101-104 Elastic constants.................................................................................................................. 57-58 Elastic deformation .............................................................................................................. 54-57 Elastic energy release rate ....................................................................................................... 73 ElastraFrac ............................................................................................................................... 36 Emulsifier ............................................................................................................................ 35, 43 Emulsions ........................................................................................................................... 35, 42 Energy.......................................................................................................... 5, 73, 77, 77-79, 166 Kinetic .......................................................................................................................... 89 Rate of using................................................................................................................ 82 Energy balance.................................................................................................................... 77-79 Enzyme breakers...................................................................................................................... 41 Erosion...................................................................................................................................... 61

F Failure mode............................................................................................................................. 73 FlexSand........................................................................................................................ 50-51, 88 Flow lines, high pressure................................................................................................. 209-210 Flow lines, low pressure ......................................................................................................... 209 Flowmeters ............................................................................................................. 176, 177, 179 Magnetic .................................................................................................................... 177 Mass or inertia ........................................................................................................... 177 Turbine....................................................................................................................... 177 Fluid efficiency .............................................................................................. 69, 70, 71, 128, 130 Fluid friction ......................................................................................................................... 27-28 Fluid leakoff .................................................................................. 8, 12, 122, 166, 169, 187, 190 Page 246


BJ Services’ Frac Manual Index Coefficient...................................................................................................................... 8 Compressibility-controlled coefficient ............................................................................ 8 Dynamic......................................................................................................................... 9 Harmonic ....................................................................................................................... 9 Spurt loss....................................................................................................................... 9 Viscosity-controlled coefficient ...................................................................................... 8 Wall-building coefficient...................................................................................... 8-9, 190 Fluid loss additives ................................................................................................................... 44 Fluid loss test........................................................................................................................... 8-9 Fluid mechanics................................................................................................................... 19-28 Fluid systems....................................................................................................................... 29-44 Additives ................................................................................................................. 39-44 Emulsion-based ........................................................................................................... 35 Energised................................................................................................................ 36-39 Oil-based ................................................................................................................ 33-35 Visco-elastic surfactant........................................................................................... 35-36 Water-based, crosslinked ....................................................................................... 30-33 Water-based, linear ................................................................................................ 29-30 Foam fracturing............................................................................................................ 36-39, 169 Proppant concentration ............................................................................................... 37 Stability ........................................................................................................................ 38 Quality.......................................................................................................................... 36 Viscosity....................................................................................................................... 38 Foaming agents ........................................................................................................................ 42 Forced closure .......................................................................................................................... 88 Forcheimer Equation ................................................................................................... 88-89, 168 Formation linear flow .............................................................................................................. 202 Frac and Pack...............................................................................................13-14, 167-168, 239 Frac job flowchart ............................................................................................................ 232-234 Frac spreads.................................................................................................................... 224-227 Fracture Area ........................................................................................................................... 130 Closure time......................................................................................................... 12, 127 Conductivity ....................................................7, 12, 16, 45-48, 83, 86, 95, 96, 164, 203 Dimensionless conductivity ..................12, 82-83, 96, 98, 164, 166, 173, 174, 183, 203 Efficiency ...........................................................................................see Fluid efficiency Gradient ............................................................................................................ 61-62, 67 Half length................................................ 7, 13, 69, 70, 71, 83, 164, 166, 202, 203, 205 Half length, dimensionless................................................................................... 97, 166 Height .......................................................................................................................... 68 Initiation, controlling............................................................................................ 109-111 Orientation ................................................................................................ 59-60, 80, 111 Relative conductivity.................................................................................................... 12 Fracture linear flow ................................................................................................................. 201 Fracture mechanics ............................................................................................................. 72-79 Fracture Models........................................................................................68-71, 91-94, 165, 186 2-D Models ..................................................................................................... 68-71, 128 3-D models ............................................................................................................. 91-94 FracPro ..................................................................................... 75, 91-92, 117, 183, 184 FracproPT..................................................................75, 91-92, 117, 135-147, 182, 184 Geertsma and de Klerk (GDK) ............................................................................... 69-70 GOHFER ..................................................................................................................... 93 Kristianovich and Zheltov – Daneshy (KZD) .................................. 69-70, 128, 130, 131 MFrac................................................................. 72, 92-93, 117, 151-152, 159-161, 182 MinFrac...................................................................................................................... 117 Penny-shaped......................................................................................................... 68-69 Perkins and Kern – Nordgren (PKN) .............................................. 70-71, 128, 130, 131 Radial.............................................................................................. 68-69, 128, 130, 131 Stimplan................................................................................................................. 72, 93 Fracture tip diameter ................................................................................................................ 76 Page 247


BJ Services’ Frac Manual Index Fracture toughness...................................................................... 72-75, 116, 169, 187, 188, 190 Fracturing........................................................................................................................... 1-3, 12 Batch fracs......................................................................................................... 224, 226 Coal bed methane ........................................................................................ 16, 170-172 Coiled tubing....................................................................................16-17, 172-173, 226 High permeability...................................................12, 166-167, 167-168, 225, 228, 239 Injection wells ............................................................................................. 169-170, 228 Low permeability, or tight gas................................12, 107, 168-169, 225, 228, 238-239 Multiple intervals ......................................................................................... 112-113, 230 Skin bypass ................................................. 15-16, 75, 98, 112-113, 165-166, 226, 240 Weak or unconsolidated formations .......................................................... 107, 225, 239 Friction factor (Fanning) ...................................................................................................... 27-28

G G-function ............................................................................................................... 126, 129, 130 G-function analysis .......................................................................................................... 128-131 Example.............................................................................................................. 134-138 Gas contacts........................................................................................................................... 108 Gas lift....................................................................................................................................... 17 Gas oil ratio (GOR) ................................................................................................................... 99 Gauges, pressure ................................................................................................................... 121 Gel stabilisers ........................................................................................................................... 43 Gelling agents................................................................................................................ 30, 39-40 Carboxymethyl guar (CMG)............................................................................. 30, 33, 40 Carboxymethyl hydroxyethyl cellulose (CMHEC).................................................. 30, 40 Carboxymethyl hydroxypropyl guar (CMHPG) ................................................ 30, 33, 40 Cellulose ...................................................................................................................... 30 Guar................................................................................................................. 30, 33, 40 Hydroxyethyl cellulose (HEC)................................................................................ 30, 40 Hydroxypropyl guar (HPG) .................................................................................... 30, 40 Oil-based fluids ...................................................................................................... 34, 40 Starch .......................................................................................................................... 30 Polysaccharide ............................................................................................................ 40 Xanthan ................................................................................................................. 30, 40 Xanthan, derivatives of ................................................................................................ 30 Gravel pack............................................................................................................................... 14 Griffith crack......................................................................................................................... 72-73 Griffith failure criterion............................................................................................................... 73

H Half length, fracture .................................................................................................................... 7 Hard rocks ................................................................................................................................ 80 Height ................................................................................................................................... 7, 16 Dimensionless ....................................................................................................... 15, 16 Helical screw rheometer ........................................................................................................... 24 Herschel-Buckley fluids ............................................................................................................ 22 High pressure flow lines .................................................................see Flow lines, high pressure Hooke’s law ......................................................................................................................... 58-59 Horizontal wells....................................................................................................................... 112 Horner plot, minifrac ........................................................................................................ 127-128 Horner plot, well testing .................................................................................................. 195, 198 Hugoton field......................................................................................................................... 1, 33 Hydration unit................................................................................................................... 217-218 Hydraulic horsepower ..................................................................................................... 2, 4, 209 Hysteresis ........................................................................................................................... 55, 66

I Independent Torpedo Company................................................................................................. 1 Page 248


BJ Services’ Frac Manual Index Inflow performance relationship (IPR) ...................................................................................... 99 Injection wells .................................................................................................................. 169-170 Intensifiers ....................................................................................................................... 213-216 Internal rate of return .............................................................................................................. 102 ISDP ................................................................................................................................... 6, 125 ISIP ............................................................................................................. 6, 125, 127, 130, 192

J Job design ...................................................................................................see treatment design

K K-prime or K’.................................................................................................................. 21-22, 86 Klepper No 1 well.................................................................................................................. 1, 33

L Laminar flow ....................................................................................................................... 26, 27 Leakoff ...............................................................................................................see Fluid Leakoff Lightning ................................................................................................................................... 32 Limited entry fracturing ................................................................................................ 84-85, 110 Linear elastic fracture mechanics (LEFM) ........................................................................... 72-75 Liquid frac concentrate (LFC, XLC, GLFC, VSP) ..................................................................... 33 LiteProp ............................................................................................................................... 51-52 Live annulus............................................................................................................................ 121 Logistics.................................................................................................................................. 108 Low surface tension modifiers .................................................................................................. 42

M McGuire and Sikora ............................................................................................................. 97-98 Medallion Frac .......................................................................................................................... 32 Medallion Frac HT .................................................................................................................... 33 Micelles ..................................................................................................................................... 35 Micro fibres .......................................................................................................................... 87-88 Micro sheets ............................................................................................................................. 88 Micriseismic .................................................................................................................... 186, 206 Minifracs ....................................................................... 8, 115, 121-163, 167, 181-183, 236-237 Anatomy of................................................................................................................. 124 Bottom hole data................................................................................................. 121-122 Examples ............................................................................................................ 134-162 Fluid type ................................................................................................................... 122 Planning and execution ...................................................................................... 121-123 Rate ........................................................................................................................... 122 Volume....................................................................................................................... 122 Mobility ................................................................................................................................... 168 Monobore.......................................................................................................................... 17, 170 Multi-phase flow................................................................................................ 47, 164, 169, 174 Multiple fractures ..................................................................................84-85, 109, 111, 133-134 Mutual solvents......................................................................................................................... 43

N n-prime or n’................................................................................................... 21-22, 86, 128, 130 N2 fracturing ..................................................................................................................... 38, 169 Napalm ..................................................................................................................................... 33 Near wellbore damage ............................................................................................................... 9 Net height ....................................................................................................................... 200, 201 Net present value (NPV).......................................................................................... 102-104, 174 Net pressure ..................................................................................................... see pressure, net Net revenue ............................................................................................................................ 102 Page 249


BJ Services’ Frac Manual Index Newton’s law of fluids .......................................................................................................... 19-20 Newtonian fluids ................................................................................................................. 20, 22 Nodal analysis ................................................................................................................... 99-100 Nolte analysis ................................................................................................... 82, 124, 179, 192 Nolte G-function...................................................................................................................... 126 Nolte G-function analysis................................................................................................. 128-131 Example.............................................................................................................. 134-138 Non-Darcy flow .....................................................................12, 16, 47-48, 88-89, 164, 169, 174 Non-emulsifiers......................................................................................................................... 41

O Oxidising breakers .................................................................................................................... 41

P p-wave ...................................................................................................................................... 62 Perforations ....................................................................................................... 81, 108, 109-114 Deviated wells..................................................................................................... 111-112 Friction ................................................................................................................... 6, 116 Strategy ............................................................................................................... 81, 108 Vertical wells.............................................................................................................. 111 Permeability ........................................................................................................ 4, 169, 189, 194 Formation........................................................................................... 7, 10, 13, 100, 164 Proppant pack .................................................................... 7, 12, 13, 45-48, 83, 89, 164 Regained ................................................................................................................ 46-47 Pipelining ............................................................................................................................. 86-87 Plane strain............................................................................................................................... 74 Plastic deformation .................................................................................................. 54-55, 76, 77 Plastic zone ......................................................................................................................... 76-77 Plug flow ............................................................................................................................. 26, 27 Poisson’s ratio ...................................................................55-56, 58, 59, 60, 62, 68, 69, 70, 191 Dynamic......................................................................................................... 63, 66, 189 Polished bore receptacle .......................................................................................................... 17 Poly CO2 .................................................................................................................................. 39 Polyemulsion ............................................................................................................................ 35 Poroelastic constant ............................................................................................... 58, 62, 64, 67 Porosity........................................................................................................................... 190, 202 Formation............................................................................................................. 10, 190 Power........................................................................................................................................ 77 Power law fluids................................................................................................................... 21-22 Pressure ..................................................................................................................................... 5 Bottom hole flowing ............................................................................... 10, 99, 100, 193 Bottom hole static ...................................................................................................... 193 Bottom hole treating .............................................................................. 5, 122, 124, 125 Bottom hole treating, calculated ................................................ 122, 124, 176, 179, 187 Breakdown...................................................................................................... 61-62, 191 Build-up...................................................................................................................... 193 Closure .......................................................................... 6, 115, 116, 125, 127, 164, 192 Dimensionless ........................................................................................................... 199 Dimensionless, match ............................................................................................... 200 Drawdown.......................................................................................................... 193, 228 Extension ....................................................................................................... 6, 115, 116 Fluid friction .................................................................................................... 27-28, 186 Hydrostatic..................................................................................................................... 5 Instantaneous shut-in (ISIP) .......................................................... 6, 125, 127, 130, 192 Maximum wellhead .................................................................................................... 105 Near wellbore friction............................................................................................. 6, 117 Net .................................................6, 68, 69, 70, 71, 76, 79, 82, 84, 117, 125, 166, 192 Perforation friction.......................................................................................................... 6 Pore ................................................................................................................. 61, 62, 67 Page 250


BJ Services’ Frac Manual Index Pseudo, gas....................................................................................................... 199, 201 Reservoir ........................................................................................................... 100, 128 Shut-in ....................................................................................................................... 193 Surface treating ..................................................................................................... 5, 122 Tubing.................................................................................................................... 5, 123 Tubing friction ............................................................................................. 5, 27-28, 176 Wellhead........................................................................................................................ 5 Wellhead, flowing......................................................................................................... 99 Pressure decline ..................................................................................................................... 123 Decline curve analysis ........................................................................................ 125-131 Pressure matching.............................................................................131-132, 183-184, 185-191 Examples ..............................................................................135-147, 151-152, 159-161 Limitations of...................................................................................................... 132, 184 Pressure transducers .............................................................................................. 176, 178-179 Pressure transient analysis ............................................................................................. 194-199 Production increase ........................................................................................................... 95-100 Dimensionless ............................................................................................................. 98 Pseudo-steady state............................................................................................... 96-98 Steady state............................................................................................................ 95-96 Productivity index (PI)........................................................................................... 95, 96, 98, 166 Proppant .......................................................................................................................... 4, 43-52 Average grain size ....................................................................................................... 46 Closure stress.............................................................................................................. 47 Concentration, areal ...................................................................................................... 7 Concentration, foams .................................................................................................. 37 Concentration, slurry ............................................................................................ 7, 178 Convection................................................................................................................... 85 Grain size distribution .................................................................................................. 41 Multi-phase flow........................................................................................................... 48 Non-Darcy flow ....................................................................................................... 47-48 Permeability ............................................................................ 7, 12, 45-48, 83, 164, 203 Pump more ............................................................................................... can’t go wrong Regained permeability............................................................................................ 46-47 Resin-coated.......................................................................................................... 47, 87 Roundness................................................................................................................... 46 Selection ................................................................................................................. 47-49 Settling.................................................................................................................... 85-86 Slugs.................................................................................................................... 81, 123 Sphericity ..................................................................................................................... 46 Storage and handling ......................................................................................... 218-220 Substrate ..................................................................................................................... 45 Transport ............................................................................................................... 37, 85 Volume....................................................................................................................... 130 Proppant flowback ............................................................................................................... 86-88 Causes of................................................................................................................ 86-87 Forced closure ............................................................................................................. 88 Prevention of........................................................................................................... 87-88 Proppant number ............................................................................................................. 173-174 Pseudo radial flow .................................................................................................................. 202 Pseudo-steady state flow ................................................................................................. 97, 194 Pump curves........................................................................................................................... 209 Pumps, high pressure...................................................................................................... 211-213

Q Quality, foams...................................................................................see Foam fracturing, quality

R Radial extent (of reservoir) ....................................................................................................... 16 Radioactive tracers ................................................................................................................. 207 Page 251


BJ Services’ Frac Manual Index Radius of investigation ................................................................................................... 194, 196 Rate ............................................................................................................................ is my friend Re-design, treatment, on-site .......................................................................................... 181-183 Re-design, treatment, real-time ....................................................................................... 183-184 Relative conductivity................................................................................................................. 97 Relative fracture conductivity...................................82-83, 96, 98, 164, 166, 173, 174, 183, 203 Remote data transmission............................................................................................... 180-181 Resin-coated proppant ............................................................................................................. 47 Reynold’s number.......................................................................................................... 26-27, 28 Roundness................................................................................................................................ 46

S s-wave ...................................................................................................................................... 63 Shear modulus..................................................................................................................... 56-57 Dynamic....................................................................................................................... 64 Shear rate ................................................................................................................................. 19 Shear strain .............................................................................................................................. 57 Shear stress (fluids).................................................................................................................. 19 Shear stress (solids) ................................................................................................................. 57 Shear-thickening fluids ............................................................................................................. 22 Shear-thinning fluids ................................................................................................................. 22 Shear wave............................................................................................................................... 63 Skin factor..................................................................... 9-11, 15, 16, 99, 100, 106-107, 196-197 Sliding side door (SSD) ................................................................................................... 105,106 SpectraFrac G .......................................................................................................................... 32 SpectraFrac G HT..................................................................................................................... 32 Sphericity .................................................................................................................................. 46 Spurt loss............................................................................................................................ 9, 200 Steady state flow .................................................................................................................... 194 Step down test ................................................................................................................. 116-117 Step rate test .................................................................... 115-120, 121, 181-183, 236-237, 238 Examples .............................................................. 117-119, 139-140, 154-155, 158-159 Step up test...................................................................................................................... 115-116 Strain ................................................................................................................................... 53-54 Stress........................................................................................................................................ 53 Closure ........................................................................................................................ 47 Cycling ......................................................................................................................... 86 Horizontal, maximum and minimum ......................................................58-59, 61-62, 81 Horizontal, contrasts ............................................................................................ 81, 117 In-situ ...........................................................................58-59, 60-61, 126, 187, 189, 190 Radial...................................................................................................................... 59-60 Tangential ............................................................................................................... 59-60 Vertical......................................................................................................................... 58 Wellbore-related ..................................................................................................... 59-61 Logs ................................................................................................................ 63-67, 189 Stress intensity factor .................................................................................................... 74-75, 76 Sub-surface safety valve (SSSV) .................................................................................... 105,106 Suction hoses ......................................................................................................................... 210 Surface tension......................................................................................................................... 42 Surfactants........................................................................................................................... 41-42 Amphoteric................................................................................................................... 41 Anionic ......................................................................................................................... 41 Cationic........................................................................................................................ 41 Emulsifying .................................................................................................................. 42 Foaming agents ........................................................................................................... 42 Low surface tension modifying .................................................................................... 42 Mutual solvents............................................................................................................ 42 Non-emulsifying ........................................................................................................... 41 Nonionic....................................................................................................................... 41 Page 252


BJ Services’ Frac Manual Index Super RheoGel .................................................................................................................... 34-35

T Temperature ...................................................................................................... 19, 121-122, 198 Temperature logs.................................................................................................................... 207 Tensile strength ............................................................................................................. 52, 61-62 Terminal velocity....................................................................................................................... 86 Tiltmeters ................................................................................................................. 186, 205-206 Time................................................................................................................................. 126-131 Closure .............................................................................................................. 126, 127 Data ........................................................................................................................... 126 Delta .......................................................................................................................... 126 Delta Nolte Time........................................................................................................ 126 Dimensionless, fractured well.................................................................................... 202 Dimensionless, match ............................................................................................... 200 Dimensionless, minifrac..................................................................................... 126, 128 Dimensionless, well testing ....................................................................................... 198 Horner......................................................................................................... 126, 127-128 Nolte time................................................................................................................... 126 Nolte G time....................................................................................... 126, 128, 129, 130 Producing................................................................................................................... 193 Pump ......................................................................................................................... 126 Shut in........................................................................................................................ 126 Square root time ......................................................................................... 126, 126-127 Tip screenout ..........................................................................................13, 83-84, 166-167, 167 Titanates .............................................................................................................see Crosslinkers Tortuosity ......................................................................80-81, 111, 116, 117, 123, 132-133, 192 Controlling........................................................................................................... 111-112 Curing of ...................................................................................................................... 81 Example.............................................................................................................. 147-153 Tracer logs.............................................................................................................................. 207 Transient flow ........................................................................................................... 97, 193, 194 Transit time ............................................................................................................................... 63 Compression wave (p-wave) ....................................................................................... 63 Matrix ........................................................................................................................... 64 Shear wave (s-wave) ................................................................................................... 63 Treatment design..............................................................................................164-175, 232-238 Examples ............................................................................................................ 238-240 General ....................................................................................................... 164-165, 236 On-site redesign ................................................................................................. 181-183 Real-time redesign.............................................................................................. 183-184 Treatment monitoring....................................................................................................... 176-186 Analysis and display of data ...................................................................................... 179 Data processing......................................................................................................... 179 Equipment........................................................................................................... 220-221 FracRT....................................................................................................................... 143 Isoplex ....................................................................................................................... 179 JobMaster ................................................................................................... 179-180, 181 Remote data transmission.................................................................................. 180-181 Treesaver........................................................................................... see Wellhead isolation tool Tubing cooldown..................................................................................................................... 104 Tubing expansion ................................................................................................................... 105 Turbulent flow ..................................................................................................................... 26, 28

U Unified fracture theory ..................................................................................................... 173-172

Page 253



V Viking ........................................................................................................................................ 32 Viking D .................................................................................................................................... 32 Visco-elastic surfactants (VES) ........................................................................................... 35-36 Viscometers ......................................................................................................................... 23-25 Brookfield..................................................................................................................... 25 Funnel.......................................................................................................................... 25 Model 35 ........................................................................................................... 23-24, 25 Model 50 ................................................................................................................ 24, 85 Viscosity.................................................................................................................. 10, 19, 20, 23 Apparent .............................................................................................................. 25, 177 Foams.......................................................................................................................... 38 Fracturing fluid filtrate ............................................................................................ 8, 190 Gas ............................................................................................................................ 198 Measurement of...................................................................................................... 23-25 Reservoir fluid.................................................................................................. 8, 10, 190 Vistar......................................................................................................................................... 33 von Mises’ yield criterion ..................................................................................................... 76-77

W Water contact.................................................................................................................. 108, 228 Water cut .................................................................................................................................. 99 Weak formations....................................................................................................................... 86 Well testing ...................................................................................................................... 193-204 Build-up.............................................................................................................. 193, 194 Constant rate ..................................................................................................... 193, 194 Diagnostic plots .................................................................................................. 197-198 Drawdown.................................................................................................................. 193 Fractured wells .......................................................................................................... 201 Gas well testing .................................................................................................. 198-199 Post-treatment .................................................................................................... 202-204 Pressure transient analysis ................................................................................ 194-199 Type curve matching .......................................................................................... 199-202 Wellbore deviation .................................................................................................................... 80 Wellbore fluid, effects of ......................................................................................................... 123 Wellbore orientation.................................................................................................................. 61 Wellbore radius....................................................................................................... 10, 11, 16, 98 Wellbore radius, effective ......................................................................................................... 11 Wellbore storage............................................................................................................. 200, 202 Dimensionless ................................................................................................... 200, 202 Wellhead isolation tool..................................................................................................... 221-224 Wheatstone’s bridge ............................................................................................................... 178 Width........................................................................................................................................... 7 Average ....................................................................................... 7, 68, 69, 70, 203, 205 Average propped ........................................................................................... 13, 83, 164 Maximum ......................................................................................................... 68, 69, 70 Wireline logs ........................................................................................................................ 63-67

Y Yield point ................................................................................................................................. 76 Young’s modulus .............. 13, 54-55, 68, 69, 70, 73, 80, 84, 130, 165, 167, 169, 187, 188-189, ........................................................................................................................................ 190, 191 Dynamic................................................................................................... 55, 63, 66, 189 Plane strain.................................................................................................. 55, 130, 191 Static...................................................................................................................... 55, 66

Page 254



Z z-factor .................................................................................................................................... 198 Zirconates ...........................................................................................................see Crosslinkers

Page 255


BJ Frac Manual v1.0 June 2005(A4)

Recommend Documents