Natural Fracture Modeling

© 2020 ADIGM
All rights reserved. No part of this document may be reproduced in any form possible, stored in a retrieval system, transmitted and/or
disseminated in any form or by any means (digital, mechanical, hard copy, recording or otherwise) without the permission of the copyright owner.
Content
• Almost all rocks are fractured
• Data Sampling Bias
• Fracture Origin and Classification
• Fracture Identification
• Fracture Modeling
• Continuum Fracture Model
• Discrete Fracture Network Model
• Fracture Set and Orientation (Dip, Dip Azimuth and Concentration)
• Fracture Width or Aperture
• Fracture Length
• Fracture Height or Elongation Ratio
• Mechanical Stratigraphy
• Fracture Distribution (Fracture Driver)
• Fracture Property

2
Almost all rocks are fractured!
• Knowledge of fractures in our reservoir is important to decide
whether a well should be drilled through or away from fractures
• Fractures presence may enhance or deteriorate our reservoir
properties
• Fractures presence affect pressure drilling window (critical stress
related)
• The magnitude of fractures effect is determined by;
• Fracture density
• Fracture permeability and permeability anisotropy
• Fracture distribution

3
Never underestimate the fractures!
• When the effect of fracture is minor, we may ignore it
• But never underestimate the fracture roles in our reservoirs

“potential enhancement” “potential reduction”

4
Data Sampling Bias
Vertical and Deviated Well

• Many fractures are vertical or normal to bedding

• Consequently well deviation is an issue when collecting fracture data
• Vertical well is favorable to collect fracture height data, but
probability to intersect fractures is less
• Deviated or horizontal well is more likely to intersect fractures, thus
sometimes fracture spacing can be measured

5
 Data Sampling Bias
Vertical and Deviated Well

No fractures
seen from
this well
6
 Data Sampling Bias
Fractures and Fracture Corridors

• Fracture corridor is a zone with much higher fracture intensity and

larger scale than the normal fractured zones
• Commonly developed as fault-related fractured zone
• Wells penetrating this corridor will significantly have oversampling bias

7
Fractures Origin
Based on their geologic origin, fractures can be classified into:
• Tectonic fractures
• Fault related fractures
• Fold related fractures
• Stylolite

• Regional fractures
• Contractional fractures
• Desiccation (Mud Cracks)
• Syneresis (chemical process)
• Thermal Gradient (cooling of rocks)
• Mineral phase changes (volume reductions)

• Surface related fractures

8
Nelson’s Fracture Classification
Based on fractures impact to overall reservoir properties, Nelson
(1985, 2001) classifies the fractured reservoirs into four types:

• Type 1: Fractures provide the essential reservoir porosity and

permeability
• Type 2: Fractures provide the essential reservoir permeability
• Type 3: Fractures assist permeability in an already producible reservoir
• Type 4: Fractures provide no additional porosity or permeability but
create significant reservoir anisotropy (barriers)

9
Fracture Identification
Core & Image Log

Acoustic Electric

Goniometer
10
Fracture Identification
Drilling Operation

Hennings et al. (2012)

11
Fracture Identification
Seismic

Chopra and Marfurt (2010) 12

Fracture Modeling Approaches
Fractured reservoir can be modeled with different approaches:
• Based on number of medium:
• Single medium
• Dual medium
• Based on modeling technique
• Continuum Fracture Model
• Discrete Fracture Network

Warren and Root (1963) 13

Dual or Single Medium?
• Single medium can be used in the case of:
• Fractured reservoirs type 1 (no matrix storage)
• Fractured reservoirs type 3 (matrix is porous and permeable) but flow
exchanges between fractures and matrix are fast
• Permeability ratio of fracture and matrix is low
• Monophasic flow reservoir.

• Dual medium can be used when:

• Matrix – fractures exchanges are non instantaneous
• Permeability ratio fracture / matrix is high (>10)
• Flow anisotropy is observed

14
 Fracture Scale

Fracturing occurs at a variety of scales...

Model scale fractures

Microfractures Intermediate fractures Large scale fractures

 ϥm – mm  cm – m  dm – km
 Can only be seen from  The most common type  Can be seen even from
outcrop and core  Can be seen from outcrop, seismic
core and image log

15
Fracture Modeling Workflow

*SPE Forum, 2010 16

Continuum Fracture Model
• Matrix and fractures are modeled separately, i.e. dual porosity, dual
permeability
• However, this approach model all fractures of any origin and any
orientation in one model
• Focus only on the factors that control where fracturing occurs, not
the fractures themselves
• Fractures are modeled by integrating all scales information from
core, log, and seismic into a single fracture intensity model

17
Continuum Fracture Model
• The main objective of CFM is to model fractures distribution based
on :
• Conceptual geology derived from possible fracture distribution in
relation to lithology, porosity, bed thickness, distance to fault or
distance to fold hinge
• 2D or 3D seismic analyses, e.g. acoustic impedance, pre-stack or post
stack attributes etc.
• Multiple fracture indicators especially from seismic analyses will be
combined using statistical method or neural network

18
Continuum Fracture Model

Fracture density
as trend

Jenkins et al. (2009) 19

 Continuum Fracture Model
Fracture Well Data
Fracture Seismic Attributes
Acoustic Imp. Max. Curvature Fracture Porosity

Fracture Permeability
Neural
Network

Analysis based on grouping of:

- Mechanical stratigraphy
- Litho or depofacies
Tuning Freq. of Spec. Imaging Up. Bandwidth. of Spec. Imaging

(Ouenes, 2010) 20
Discrete Fracture Network
• Introduced by Dershowitz et al. in 1998
• A mathematical representation of fracture
systems in 3D geometry
• Fractures are modeled as planar objects with
orientation, approximated size, and
distribution in space,
• Every fracture set has individual properties
• DFN is geologically driven fracture model
• Requires thorough understanding of fracture
system in the reservoir, i.e. data intensive and
time consuming

21
 DFN Modeling Workflow

Fracture Characterization

Seismic

Fract. Aperture

Fract. Geometry
Upscaling
Fract. Orientation

Fract. Intensity
DFN Ф, k, σ

22
Fracture Characterization for DFN
• DFN Modeling requires a set of fracture analysis from integration of
core, log, seismic and geological interpretation
• Fracture Set and Orientation (Dip, Dip Azimuth and Concentration)
• Fracture Width or Aperture
• Fracture Length
• Fracture Height or Elongation Ratio
• Mechanical Stratigraphy
• Fracture Distribution (Fracture Driver)
• Fracture model is then upscaled to create the fracture porosity,
permeability and sigma

23
Fracture Characterization for DFN

Narr (1996)
24
Fracture Sets and Orientation
• Fracture set is a group of fractures
whose geometry and properties can
be described by a single set of
parameters

• A set of fracture is defined by its:

• Specific orientation (dip and dip
azimuth)
• Origin
• Time of development
• Fracture scale
• Fracture morphology (open or
closed)

Richard et al. (2017) 25

 Fracture Sets and Orientation

This outcrop has three fracture sets

• 2 conjugate shear (blue and yellow)
• 1 extensional fracture (red)

26
 Fracture Sets and Orientation
• Image logs are very useful to identify
fractures in the borehole
• Several features that can be
identified:
• Natural or drilling induced fractures
• Fractures orientation
• Fractures morphology

27
Prioul and Jocker (2009)
Fracture Width and Aperture
• Fracture aperture is defined as perpendicular distance separating
fracture walls – also called as kinematic aperture
• Fracture width generally refers to distance of open space
perpendicular to fracture walls
• Fracture width may increase by dissolution of fracture walls
• Fracture width may decrease by mineral precipitation on the fracture
walls
• Aperture size generally ranges from approximately 1 μm to 1 m

28
 Fracture Width and Aperture

aperture

In the photo above the fracture width is zero

Narr (1996) 29
Fracture Width from Image Log
• The most reliable fracture width measurement is by using image log.
• However, the calculation uncertainty is high because of its dependency on
how deep the mud has penetrated into the fracture
• The smaller the fracture width the less precise the calculation
• For 1 mm fracture, the uncertainty is ~50%

Nelson (1998) 30
Fracture Width and Aperture
• Opening-mode fracture apertures commonly follow power law size
distributions
• A power law forms a straight line on a log cumulative frequency
versus log aperture size plot

Marret et al. (1999) 31

Fracture Height and Length
• Fracture shape is thought of as ellipse in nature
• However, due to computational difficulty, it is modeled as
rectangular shape
• Elongation ratio is defined as ratio of length over height

32
Measuring Fracture Height
• Vertical propagation of fractures is commonly halted at the bed
boundary
• Thus fracture height in stratified rocks very often is equal to the bed
thickness
• In crystalline rocks, i.e. metamorphic and igneous rock, the
boundary can be metamorphosed stratigraphic boundaries, or
features like dykes and sills

33
 Measuring Fracture Height

Fracture height = bed thickness

Fracture height in igneous rock
Fracture height = sill thickness

Fracture height in
sedimentary rock

34
http://pesgb.org.uk/
Measuring Fracture Length
• Fracture length is usually estimated from lineament observed from seismic.
• Large scale fracture length observed from satellite images or outcrop can
also be used
• Smaller scale fracture length can be estimated using power law scaling
derived from lineament identified from seismic

Thachaparambil (2014) 35
Mechanical Stratigraphy
• Mechanical stratigraphy subdivides stratified rock into discrete
mechanical units defined by properties such as tensile strength,
elastic stiffness, brittleness, and fracture mechanics properties
(Young’s modulus, Poisson ratio, rigidity modulus)
• Mechanical stratigraphy affects the structure development including
fracture style, dimensions, distribution and mechanics
• Mechanical stratigraphy is directly related to lithology, P-T
conditions, and strain history

36
Mechanical Stratigraphy
• Simple example; joints in bedded lithology are often restricted to
bed boundary
• Mechanical unit may include several different units of lithology
• Mechanical unit thickness often controls joint shape and joint
spacing (often proportional to bed thickness).
• Fracture characterization should be carried on each mechanical
strata

37
Mechanical Stratigraphy
Lithology Controls

Stearns and Friedman (1972) 38

Mechanical Stratigraphy
Bed Thickness Controls

Fracture spacing increases as bed

thickness increases.

39
Mechanical Stratigraphy

Underwood et al. (2003) 40

Mechanical Stratigraphy
Bach Ho Field

Zone I
High ф & k

Zone II
Med. ф & k

Zone III
Low ф & k

After Cuong (2010) 41

Fracture Distribution
• Fracture spacing, frequency, density and intensity are used
interchangeably in the literatures
• Mauldon and Dershowitz (2000) proposed “fracture abundance”
terminology
• Scale independent
• The terminology covers fracture density, fracture intensity and porosity
• Defined in 1D, 2D and 3D

42
 Fracture Abundance
Dimension of Feature
0 1 2 3
P00 Point
Dimension of Sampling Region

0 Number of
fracture samples
per point samples
Measure
of rock mass

P10 P11
Number of Length of fracture Length In 3D fracture modeling, the
1 fracture intersects
per line length of
intersections per
line length of
Measure fracture distribution will be defined
scanline scanline from the volume measurement, i.e.
P20 P21 P22 P30, P31, and P32.
Number of Length of Area of fractures Areal
2 fractures per area fractures per
of rock area of rock
per area of rock
Measure

P30 P31 P32 P33

3 Number of
fractures per
Length of
fractures per
Area of fractures
per volume of
Volume of
fractures per
Volume
Measure
volume of rock volume of rock rock volume of rock
Density

Dershowitz and Herda (1992) 43

Fracture Driver
• Fracture driver is a 3D model of fracture abundance, e.g. P30
(fracture intensity), P31, or P32 (fracture density)
• Used as input to guide the fracture population in 3D space
• Can be generated from:
• Well data
• Structure (distance to fault or distance to fold hinge)
• Seismic data

44
 Fracture Driver - example

• Fractures are interpreted to be

concentrated along the fold hinge
• The closer to the fold hinge, the
higher the fault intensity

45
Fracture Driver from Seismic
Seismic analysis that can be used to identify fractures:
• Seismic attributes:
• Curvature
• Coherence/Variance
• Dip
• and many more such as Ant track, Chaos, etc.
• Azimuthal Anisotropy
• AVOA/AVAZ (Amplitude Variation with Offset/Angle and Azimuth)
• VVAZ (Velocity Variance with Azimuth)
• Shear wave splitting

46
Seismic Curvature

Beicip Franlab 47
 Coherence/Variance
Coherence
• Seismic data are converted into a volume
of discontinuity that reveals faults,
fractures, and stratigraphic variations
• Coherency, similarity, continuity,
semblance are similar and relate to a
measure of similarity between a number
of adjacent seismic traces (multi-trace
analysis).
• Variance on the other hand measure
dissimilarity

Chopra and Marfurt (2010) 48

Azimuthal Anisotropy

• Azimuthal anisotropy has been proven not only as a fracture detection

tool but also to infer the fracture orientation from seismic
• Based on P wave or S wave amplitude
• Common methods:
• AVOA/AVAZ
• VVAZ
• Shear wave splitting

49
Fracture Driver from Seismic
• The seismic derived fracture driver is generated through correlation
of well fracture intensity and seismic attributes
• Single attribute may not satisfy the correlation with fracture intensity
at well, thus multi attributes combination to generate a single
fracture intensity can be used.
• Common approach:
• Neural network
• Multivariate analysis

50
Building Fracture Drivers

Seismic attribute

Collocated
co-kriging

Fracture intensity

51
DFN Model Result

Hinge oblique fracture set ex.

52
Fracture Properties
• Fracture properties normally automatically calculated in commercial
software
• Porosity
• Permeability
• Sigma (connection factor defining the connectivity of matrix and
fractures, describing how fluids flow between matrix and fracture in
dual medium reservoir)
• These properties can be edited with apriori knowledge
• Later will be upscaled to grid and used as input for dynamic
simulation

53
Upscaling Fracture Properties
• Fracture properties that have been generated are part of fracture
plane attributes
• Those properties need to be upscaled to grid cells that will later be
used as input in dynamic simulation
• Two common upscaling methods:
• Oda method
• Tensor method

54
 Upscaling Fracture Properties
Oda Method

• A statistical-based method
• Estimates permeability based on the total area of fractures in every
cell
• Pro:
• Very fast and give relatively good results
• Con:
• Does not account for fractures connectivity, therefore potentially over
estimate the permeability
• Very sensitive to cell size

55
Upscaling Fracture Properties
Tensor Method

• Tensor method is flow-based method that creates a finite element

grid for each cell and simulates flow under a pressure gradient to
calculate the permeability on each particular direction (i, j, and k)
• Fault surfaces are used during the simulation, when the fractures are
connected, the flow will be simulated along these connected
fractures
• Pro:
• yield accurate results
• Con:
• very slow

56
Upscaling Fracture Properties

Upscaling output:
- Porosity
- Permeability I
- Permeability J
- Permeability K
- Sigma

57
 Upscaling Fracture Properties

Intensity Porosity

Permeability I Sigma

58
Thanks !

Connect with ADIGM at LinkedIn