Petrel Methodology Methodology (DFN + IFM) Create a Fracture network model – Use Intensity upscaled logs – Use Zones and Regions – Create Stochastic and Deterministic Fracture sets
Visualize and QC the fracture model Create/update Fracture attributes – Aperture – Permeability
Building a Fracture network model Petrel methodology Petrel combines two approaches to standard fracture modeling by the use of a Hybrid model: Discrete Fracture Network – (DFN): The large/important fractures are modeled explicitly as discrete patches Implicit Fracture Model – (IFM): The residual part of the distribution (smaller fractures) is statistically represented as grid properties
Building a Fracture network model Overview
Fracture sets Create a Fracture model using single input Or using several fracture sets. Deterministic / Stochastic model Uses a direct input or models the fractures stochastically based on input statistcs Stochastic Fracture modeling Model fractures sets for the whole 3D grid, per Region or within Zones. A fracture model requires 3 basic inputs: 1. Fracture Distribution 2. Fracture Geometry 3. Fracture Orientation
Building a Fracture network model What are Fracture sets? Fracture sets Can typically be fractures analyzed from image logs, separated by genetic events, orientation or other factors.
Building a Fracture network model Deterministic Model Deterministic Fracture model: Can typically be single faults that act as fluid highways, or possible barriers if sealed by intense smearing Input data types: Fault Patches from Ant tracking Fault surfaces/polygons/points Existing Fractures (import .FAB files) •
•
•
Building a Fracture network model Stochastic Model Stochastic Fracture model: Can typically be fractures where location, size and orientation is not directly known, but can be inferred from statistics. Extent Defines Fracture network or Set representing: Zones Regions Entire 3D Grid •
•
•
Apply Property Filter
Building a Fracture network model Fracture distribution Fracture Intensity Most important fracture parameter. Related to fracture spacing (Sf).
Scale independence P11/Sf – Dependant on orientation, Independant of fracture size
2D Areal view - P22
1D Vertical view – P11
Sf Sf
Sf Sfe Modified from Dershowitz & Herda (1992)
P22 – Fracture length/plane area Random line Measurement line (e.g.borehole)
Same Fracture spacing
Same Fracture intensity
Fracture centres x 2 Fracture length x 1/2 P11 – Fracture centre/length Sf – Fracture spacing Sfe – Expected fracture spacing
Building a Fracture network model Stochastic model – 1. Fracture distribution Density distribution types • Property • Surface • Constant value
Define density distribution • Frac / Volume (P30) • Frac Length / Volume (P31) • Frac Area / Volume (P32) Equivalent to Petrel
1?, 7?, 10?
Frac Area / Volume (P32) Scale independant; most used in 3D fracture modeling
Building a Fracture network model Fracture geometry
Stratabound Restricted size scale with regular spacing. Joints (constant normal distribution). Non-stratabound Wide size range (power law), spatially clustered and vertically persistent.
SCALE / SHAPE
Joints - modeled in a Zone Faults - modeled in entire Grid
STD.DEV
MEAN
STD.DEV
Building a Fracture network model Stochastic model – 2. Fracture geometry Shape Defines how the simplified fracture plane should look like (number of sides)
Elongation ratio can be set larger for stratabound fractures
Not Modeled
Fracture length Use constant or property as parameter for distribution of fracture length in the model
20
1000
Note: you should truncate to limit the number of modeled fractures
Building a Fracture network model Stochastic model – 2. Fracture geometry Max length of implicit fractures Defines the threshold where any fractures below the value are modeled as Implicit properties (IFM), and all values above are modeled as Discrete fractures (DFN).
Max length cutoff
Building a Fracture network model Fracture orientation Orientation of fracture dip angle/azimuth Directions for normal to fracture planes are scattered around a mean dip and azimuth based on a given concentration. Common parameters: Mean dip – from initial fracture analysis Mean dip azimuth – direction of fracture plane dip (not strike!) Concentration (Kappa factor) – ’angular’ Standard deviation •
•
•
Focussed distribution – High Kappa factor Medium focussing – Here: for near vertical fractures Scattered/uncertain distribution – Low Kappa factor; here for near horizontal fractures
Building a Fracture network model Stochastic model – 3. Fracture orientation Regional Trends Defined orientation of the modeled fracture dip/orientation will be relative to a 3D Grid / Surface.
Method Petrel uses 3 different models to describe the distribution of angles: Fisher, Kent and Bingham.
Concentration = 10 No trend
Regional trend
Concentration = 80
Building a Fracture network model Result (DFN) Fracture model output (discrete planes) A separate Discrete fractures folder will be created showing fractures as ’flat’ planes.
Building a Fracture network model Result (IFM) Fracture model output (continuous properties) A separate Implicit fractures folder will be created showing fractures as attribute properties.
Note: The Fracture sets filter also works for the Implicit fractures
QC Fracture Model Statistics QC Discrete fractures in Histogram/Stereonet tab Inspect output distributions with input parameteres. Use interactive filters (pink selection) in both tabs for all attributes. New attributes can be generated and filtered on as well. Filtered property name is marked in red.
Filtered fracture property
Interactive filter selection
Stereonet filter options
Gray region will be filtering the poles to the fracture planes
Update Fracture Attributes Fracture attributes Aperture Aperture (a) is a highly uncertain parameter, but can be ’measured’ from image logs at the well scale. Alternatively, as a first run, a constant value can be assigned, e.g. per fracture set. Aperture may also be determined mechanically; related to fracture size/length. Aperture (FVA)
Aperture (a) Spacing (Sf) ) H A V F ( e r u t r e p a c i l u a r d y H
Update Fracture Attributes Fracture attributes Flow/Permeability Flow can generally be related the ’Cubic Law’ if no other measures are available. Permeability is assumed to increase exponentially with interconnection of fractures / sets. It can also be calibrated to production data / flow measurements / well tests. Sheet model Match sticks model
Sugar cube model
Calculated Fracture Permeability (Kf) As a simple approach we assume a Sheet model (parallel sets of fractures with equal spacing). We relate individual Fracture Permeability (Kf) to the Square of Aperture (Flow to Cube of aperture)
Modified from Reiss (1980)
Isolated fractures
Modified from Mattner (2003)
Isolated fracture zones/corridors
Interconnected fractures
Update Fracture Attributes Discrete fracture attributes – Calculator Units & Functions Discrete fractures are created with fracture attributes; some are already set like Dip and Azimuth; some can be calculated/updated like Aperture and Permeability.
Aperture (ft) = Length (ft) * Normal (Mean, Std.dev)(ft)
Perm (mD) = 1/12 * Aperture 2 (ft) * C
Derived from Surface area (ft 2) of Fracture plane Fracture length = sqrt (area*elongation ratio) Be aware of Units when using the Calculator!
Derived from Length (ft); must be converted to m
Conversion factor: 1 mD = 10-15 m2
Update Fracture Attributes Discrete fracture attributes – Calculator Calculator & Spreadsheet Calculate or update existing attributes. Calculations are updated in the Attribute spreadsheet .
Attributes used in Upscaling
Pre-calculated values (not updateable)
Process-calculated values (updateable)
QC Fracture Attributes Statistics Histogram Statistics QC the calculated attributes Aperture and Permeability. Check if the distribution, mean and dimension are as expected.
