Appendix: Fault Seal Analysis: Traptester6 Training Course

TrapTester6 Training Course
Appendix: Fault Seal Analysis
© Badley Geoscience Limited 2010

Course Version: 2010. rev 02 1
TrapTester6 Training Course Fault Seal Analysis
Appendix: Fault-Seal Analysis

The Problem …
There is a matter-of-fact tendency amongst geologists to refer to
faults either as sealing or non-sealing. This over-simplication can
result in bad drilling decisions and risk assessments.
The Solution …
Appplying quantitative fault-seal analysis methods permits better, more
informed decisions to be made about the sealing potential of faults.
Fault-seal analysis is not a discipline by itself. It is intrinsically linked to

seismic interpretation and 3D geo-modelling.
This session presents an overview of the concepts, techniques and main

uncertainties in fault-seal analysis.

In an Exploration context a Fault Seals ...

Shale IF: Reservoir is juxtaposed
against non-reservoir
Shale (Juxtaposition Seal)
Sand
Sand OR: Fault-zone processes
generate fault-rock with a
very high capillary entry
Shale
pressure.
Traditional fault-seal analysis
Sand assess the potential for
across-fault leakage
(Fault-Rock Seal)
Shale
AND: Fault has not been
reactivated.
Sand
Analysis of fault reactivation
Shale assess the potential for Dip
leakage.

1: Review of Capillary Fault Seals

Many seals in petroleum reservoirs are capillary (or membrane) seals. In
capillary seals, surface tension forces between hydrocarbon and
water hold back the hydrocarbon phase and prevent it from entering
the water-wet seal lithology.
Grain scale
Large scale
Small scale
Reservoir Fault zone Reservoir
Capillary seals fail by capillary leakage when the driving pressure (= buoyancy
pressure generated by the column) equals or exceeds the minimum displacement
pressure required for the hydrocarbons to enter the largest interconnected
pore throat of the seal.
2: Juxtaposition Analysis
The basic tool of fault-seal analysis is the juxtaposition diagram or
Allan diagram, which is a fault-plane section (fault-plane profile)
illustrating the stratigraphy brought into contact at the fault plane
(after Allan 1989).
Allan’s basic model assumed that: upthrown
(i) the fault itself has no sealing

properties;
(ii) a fault is not an open conduit;
(iii) the trapping and migration
relationships at a fault depend upon
the fault-juxtaposed stratigraphy. downthrown
In fact none of these assumptions is universally true. However, the

juxtaposition diagram remains the essential starting point for fault
seal analysis.

Adding extra stratigraphic detail

Define detailed reservoir sequence from well data and extrapolate the well-based picks &
Vshale data on to the faults to predict wall-rock stratigraphy & composition at the fault.
Vshale distribution at fault surface
Seismic-scale sequence Well-based picks
Yellow = sand
Brown = shale
Sand
Geological Layer model
Shale
Vshale cut-off

3: Fault Zone Processes & Products

Main factors influencing the nature
Clay content of the deformation products
found in fault zones:
Disaggregation zones Clay smears
Depth of burial at time of faulting
- The composition of the wall-rocks

PFFR or shaly that are slipping past each other at
gouge the fault, in particular their content
of fine-grained phyllosilicate clay
minerals.
Cataclasites Shale smears - The stress conditions at the time of

faulting, which are most strongly
controlled by the initial depth of
burial during faulting.
- The maximum temperature

experienced by the fault zone
+ post-deformation quartz cementation after faulting, mainly controlled by
at T>90o, ~Z>3km) the maximum post-faulting burial
depth.
Fault rock types shown above can be thought as
‘end members’ of a spectrum of fault rocks
Predicting Fault-Zone Composition: Algorithms

Predictive algorithms fall into two categories; smear factors & gouge ratios
1: Smear Factors
Smear Factors attempt to model the Morphology (or shape) of the clay or shale
smears in the fault zone. Two main types:
• Clay Smear Potential or CSP, (Shell)
• Shale Smear Factor or SSF, (Fault Analysis Group)
2: Gouge Ratios
Gouge Ratios attempt to model the up-scaled Composition of the fault zone.
• Shale Gouge Ratio or SGR (Badleys)

Clay Smear: Predicting Algorithm (CSP)

• Formulated after study of non-lithified ductile
clays mainly in the Niger Delta
• Clay Smear Factor (CSP) increases with shale
source bed thickness, but decreases with increasing
distance from source bed (as throw increases, the
shale incorporated into the fault zone becomes
thinner)
• Algorithm uses a ‘window’ defined as the throw/2
• Thickness exponent can range from 1 to 2; distance
exponent is 1. Distance is measured to center of
shale bed
• Sands are assumed to contain 0% Vshale; shales
contain 100% Vshale. Vshale cut-off values are used
to block the stratigraphy into sands (0% Vshale) and
shales (100% Vshale)
• Actual numbers from the algorithm do NOT equal
the real thickness of the shale smear
Key references: • Used predominantly by Shell in offshore Nigeria,
however, calibration studies using CSP tend to be
• Yielding et al (1997)
• Fulljames et al (1997) more qualitative (lack of detail)

Clay Smear Potential: Worked Example
Window of observation = throw / 2 = 15

0
a SF – Yielding method (n=1, m=1)
10 FW contribution (shale beds d & e in throw /2 window): (6/5)+(2/11) =1.38
b
20 HW contribution (shale beds b & c in throw /2 window): (2/2)+(2/12) =1.16
c
Throw = 30 SF = 1.38+1.16 = 2.54
d
a
e
CSP Shell method (n=2, m=1)
b
FW contribution: (62/5)+(22/11)
c
=7.56
d
HW contribution: (22/2)+(22/12)
e
=2.33
CSP = 7.56m (footwall contribution is largest)
CSP contribution (n=2, m=1)

CSP or SF: Σ(thicknessn / distancem)
FW contribution: (62/5)+(22/11) =7.56
HW contribution: (22/2)+(22/12) =2.33

Shale Smear Factor: Predicting Algorithm (SSF)

• Models the behavior of lithified abrasion-type smears
and is based on outcrop & seismic observations
• Smear Factor corresponds to the veneer of shale that
is abraded by a sandstone bed as it slips past a shale
bed
• Shale Smear Factor (SSF) is dependent on source bed
thickness and fault throw (not smear distance)
• Smaller numbers from the SSF algorithm correspond
to greater development of smear on the fault plane.
SSF values remain constant between the offset beds
• For single shale beds and large throw faults (seismic
scale) SSF values less than 4 correspond to continuous
smears (~ sealing). SSF values greater then 6 reflect
discontinuous smears (~ non-sealing)
• Where more than one shale bed could contribute to an
SSF value on the fault, only the greatest contribution
Key references: (smallest SSF value) is taken. This SSF calculation is
essentially 1/SGR
• Lindsay et al 1993
• Not so widely used in the petroleum industry, but for
• Gibson (1994)
a case study offshore Trinidad (Gibson 1994).
• Takahashi (2003)
Shale rich fault-gouge zones

The bulk composition of fault gouges generally reflects that of the protolith (wall
rocks), though clay enrichment can sometimes occur.
Small faults developed in impure sandstone causes ‘deformation-induced mixing’ at
the grain scale. This involves the compaction and mixing of clays with framework
grains and produces a fault rock that has a more homogeneous distribution of clays
than the host sediment.
These fault-rocks are sometimes called ‘phyllosilicate-framework fault rocks’ (or
PFFR, Knipe et al 1997) or clay gouges or shaly gouges.
The predictive algorithm Shale Gouge Ratio or SGR, uses the clay distribution
through the wall rocks, together with fault displacement, to estimate an ‘average’
clay content at each part of the fault zone.

Shale-rich fault gouge: Predictive Algorithm (SGR)

• Models the up-scaled fault-zone composition; a high
SGR value is expected to correspond to more
phyllosilicates in the fault zone. Based on seismic-
scale analysis of faulted mixed clastic sequences in the
North Sea.
• Main assumption is that sand and shale material are
incorporated into the fault gouge in the same
proportions (ratio) as they occur in the wall rocks of
the slipped interval. Completely ignores detailed fault-
rock distribution.
• Basic algorithm uses a ‘window’ defined as the throw.
The window can be applied either in the footwall, or in
the hangingwall or an average between both walls (for
average SGR, window = throw/2)
• Most common algorithm used in the petroleum
industry for predicting seal and in fault-seal
Key references: calibration studies. SGR value is a ratio, therefore
applicable in depth or time datasets
• Freeman et al (1997) • On a seismically mapped fault with an SGR of 30%,
• Yielding et al (1997) the fault is almost certainly not a uniform slab of
• Yielding (2002) phyllosilicate-framework rock with 30% clay.
Shale Gouge Ratio: Worked Example

Window of observation = throw = 30
SGR – Footwall stratigraphy
=((2x0.08)+(6x0.5)+(2x0.01)+(2x0.45)+(5x0.02)+(2x0.9)+(8x0.01)+(2x0.4)+
0
a
(1x0.01)) / 30
0.7
0.01
10
0.4 b =6.87/30 =0.229 (22.9%)
0.01 20
0.9 c
0.02 Throw = 30
0.45 d SGR – Hangingwall stratigraphy
0.01
a 0.7
0.5 e =((1x0.15)+(2x0.4)+(8x0.05)+(2x0.7)+(5x0.1)+(2x0.25)+(2x0.01)+(6x0.3)+
0.08 b
0.15
0.4 (2x0.02)) / 30
0.05
=5.61/30 =0.187 (18.7%)
c 0.7
Vshale (Vsh) 0.1
d 0.25
0.01
e 0.3
SGR – Average (window of observation = throw / 2 = 15)
0.02 FW contribution: ((2x0.08)+(6x0.5)+(2x0.01)+(2x0.45)+(3x0.02)) / 15

=4.14/15 = 0.276
Vshale (Vsh)
HW contribution: ((1x0.15)+(2x0.4)+(8x0.05)+(2x0.7)+(2x0.1)) / 15
SGR= (Σ(Vsh.Δz) / t) x100% =2.95/15 =0.196
SGR – Average = (0.276+0.196) / 2 = 0.236 (23.6%)
Vsh = proportion of clay in the interval
Δz = thickness of the interval

Key input for Fault-Seal Analysis: Vshale data

Calculation of the SGR, CSP or SSF requires input of the phyllosilicate content
(ideally Vclay) for the faulted intervals. Vshale is a ‘general’ proxy term for the
phyllosilicate content.
The Vshale parameter is a derived product, typically from gamma ray or neutron-
density logs and is not necessarily the same as the actual volumetric clay content
(Vclay) of the rock.
Detailed analysis of thin sections by point counting or by X-Ray diffraction

analysis is required to determine to true volume clay content, which is seldom
carried out on cores from exploration wells.
Significant differences in the estimate of Vshale can exist even within the same
company.

4: Calibration & Estimating column heights

Fault-seal attributes (e.g. SGR, CSP) are not a measure of the sealing
capacity of the fault.
They are estimates of the relative likelihood of clay gouge or smear

being developed at the fault surface.
To use the attributes to predict the height of a hydrocarbon column,

the seal attributes must be calibrated in datasets where the sealing
behaviour is documented from well data. The aim of the calibration is to
derive an empirical relationship between fault-seal attribute and
pressure that can be used to estimate the ‘strength’ of the fault seal.

Calibration & Estimating column heights (cont.)

There are two commonly used approaches to calibrate fault seal
attributes:
Empirical: Calibrate SGR in-situ against known hydrocarbon columns

- Measure pressure difference (hydrocarbon buoyancy) held at fault.
- Compare with SGR calculated at the same point on the fault surface
(assume SGR is just a proxy measure, related in a vague way to fault-
zone composition.
Deterministic: Assume that SGR = clay content of fault-zone. Use

values of capillary threshold pressure measured in the lab on fault-zone
samples of known composition. However, this approach is typically based
on ‘plug-sized’ samples from microfaults with small offsets and not trap
bounding faults that are supporting a hydrocarbon column. In addition,
it necessary to covert mercury-air measurements to hydrocarbon-water
systems, thereby adding more uncertainty.

Fault Seal – Leak: Observations

Vertical lines represent range of SGR
100 values on Brent/Brent overlaps on
individual faults.
Seal
90
Faults are characterised as sealing (red)
Shale Gouge Ratio (%)
80
Leak or leaking (green) depending upon whether
70
Range of SGR values at
there is a change in hydrocarbon contact
60 sand-sand juxtapositions across the fault.
50
40
Brent Province
30 (red line)
20
15-20% SGR
10
0
Offshore Trinidad
'A'
Eider(P)
Oseberg S(F97)
Hutton(P)
Gullfaks S(Y97)
Oseberg S(F97)
Oseberg S(F97)
Oseberg S(F97)
NW Hutton(P)
Don(P)
Corm IV(P)
NW Hutton(P)
Penguin(P)
Brage(P)
Penguin(P)
'A'
'A'
Gullfaks(Y99)
Gullfaks(Y99)
Strathspey(H00)
Osprey(P)
Oseberg S(F97)
Brent S(P)
NW Hutton(P)
Gullfaks(Y99)
Oseberg S(F97)
'A'
'A'
Osprey(P)
(Gibson & Bentham 2003)
Data from Brent Province, North Sea (Yielding 2002).

SGR ~ 0.2 (20%) threshold
© Badley Geoscience Limited 2010 between leak & seal
In-situ pressure calibration

Across-fault pressure calibration
Across-fault pressure is the
Well 1
Pore Pressure
difference in pressure
Well 2 measured at the same depth
Hydrocarbons Hydrocarbon across the fault.
A B B A pressure trend
in well 1 Buoyancy pressure is the
Water difference in pressure
Water
between the hydrocarbon and
Water pressure trend water phases measured on one
Well 1 and well 2
Depth side of the fault.
Buoyancy pressure calibration
Well 1 A
Pore Pressure Buoyancy Pressure
Hydrocarbons Hydrocarbon
A pressure trend Buoyancy pressure trend.
in well 1 The buoyancy pressure
Water increases upwards.
Water
Water pressure
Depth trend in well 1 Depth

Empirical Calibration: Seismic scale

Across-fault pressure calibration Buoyancy pressure calibration
From Bretan From Yielding

et al (2003) et al (2010)
Pc = 10 ((SGR/27) – C) Pc = (17.5 * SGR) + C
Pc is capillary entry pressure (in bars) Pc is capillary entry pressure (in bars)
C is 0.5 for burial depths <3.0 km (< ca. 10,000 ft) C is -3.5 for burial depths <3.0 km (< ca. 10,000 ft)
C is 0.25 for burial depths 3.0 – 3.5km (ca.10,000-11,500 ft) C is -1.05 for burial depths 3.0 – 3.5km (ca.10,000-11,500 ft)
C is 0 for burial depths > 3.5km (> ca. 11,500 ft) C is1.5 for burial depths > 3.5km (> ca. 11,500 ft)
SGR input as percent (0-100%) SGR input as percent (0-100%)
Calibration represents the maximum pressure a specific SGR value can support
Deterministic calibration: Core Scale

Clay Fraction in Core samples
(Sperrevik et al 2002)
1: Predict fault-zone permeability (kf in mD) from

core samples
kf = 80000.exp-[19.4Vf + 0.00403zmax + (0.0055zf - 12.5)(1 - Vf)7]
2: Convert fault-zone permeability to mercury/air

threshold pressure (psi)
Comparison of mercury-air injection tests on core samples Pf = 31.84.kf-0.3848
(symbols) and predicted capillary entry pressures (solid lines)
Most samples used in the analysis have Vclay < 0.4
3: Convert mercury/air threshold pressure to equivalent

hydrocarbon/water threshold pressure (psi)
Pc = Pf x 0.14 (gas) or 0.08 (oil)
Best-fit polynomial trends through the 4: Convert hydrocarbon/water threshold pressure to bars
(optional step)
scatter of data points represent the
Zf = Initial burial depth (during faulting)
average pressure a specific clay content Zmax = Maximum burial depth
(~SGR) value can support Vf = Clay fraction of fault rock (as a fraction between 0-1)

Predicting Threshold Pressure
For burial depths of ca. 3-3.5km and SGR values of

30% and above, the equations predict very similar
threshold pressures. However, as burial depth
increases and/or SGR decreases, the equations
predict significantly different threshold
pressures.
When plotted against depth, the Sperrevik

equation (thin solid lines) predicts decreasing
threshold pressure with decreasing depth. The
Bretan equation predicts relatively constant values
(thick solid lines). The Yielding calibration
predicts relatively constant values (thick dashed
lines) that are lower then the Bretan calibration.
Thin diagonal solid = Sperrevik equation;

Thick vertical solid = Bretan equation;
Thick vertical dashed = Yielding equation;
Conversion factor: 0.1

Column height at a point on the fault surface

High
Buoyancy Force Fault Zone Capillary Threshold Pressure (Pc)
(estimated from SGR)
High
Low
Hydrocarbon Buoyancy generated by the column:
ΔP
hmax
Low
Water Seal Threshold Pressure:
Combining & re-arranging to derive the maximum column height (hmax) that can be
supported by a specific capillary threshold pressure:
w = pore water density (kg/m3);

h = hydrocarbon density (kg/m3)
g = acceleration due to gravity (9.81 ms-2 or c.10ms-2)
hmax is then in metres if Pc is in Pascals (1 Pa = 10-5 bar)
Pc = threshold pressure estimated from SGR
Fault-Seal Analysis: Summary

• Different fault rocks (clay/shale smears, shaly gouges, cataclasites
and disaggregation zones) are created during fault movement. Seal
capacity will vary along and up/down the fault surface.
• In many faulted sequences, SGR of 15-25% corresponds to the

threshold between non-sealing and sealing behaviour.
• As a rule-of-thumb, self-juxtaposed sands often show low SGR (leaky)

whereas juxtaposition between different sands is more likely to be
sealed.
• SGR can be used to estimate fault-zone capillary threshold pressure

using seal-failure envelopes derived from subsurface calibrations or
from lab-based analysis of core sample data. Threshold pressure can
then be used to estimate potential column height.

Summary: The Fault Seal Workflow 1

Starting Point: What data are available? Well data, but structure not reliably mapped.
Interpretations; 3D models, structure maps, and well(s).

Prioritize the faults for analysis It may be useful to construct a ‘triangle diagram’,
e.g. those which form side-seals to potential prospects, i.e. a juxtaposition diagram for a synthetic fault.
or might compartmentalise the reservoir during production. This may reveal relevant displacements for seal/leak.
Juxtaposition analysis
Do you have maps/interpretation TWT Make Allan diagrams for key faults
in depth or TWT? using seismic interpretation of faults
and mapped horizons.
depth (Ensure horizons are picked right up to the faults.)
Make Allan diagrams for key faults,

using mapped faults and horizons.
(Ensure that the fault heaves are not too wide.)
Do you have a lithostratigraphy available from nearby well(s)?
Do the mapped horizons include all no no

lithostratigraphic layers? yes
A juxtaposition diagram
yes without stratigraphic
detail may be of little value!
Does the juxtaposition pattern show areas of Add the strat. sequence to the Allan diagram to make a
reservoir-reservoir overlap in critical areas for detailed juxtaposition diagram (either interpolate between
hydrocarbon trapping or migration? the mapped horizons or ‘hang’ fixed thicknesses
above or below the mapped horizons).
no yes
From well data (lithologies, core fracture studies), identify possible

Juxtaposition Seal fault-sealing mechanisms:
(n.b. dip leakage may occur if fault cataclasis, clay smear, diagenetic overprint
is actively slipping) (next slide)
Summary: The Fault Seal Workflow 2

Cataclasis / Clay smear : Collect well information on
the distribution of clay through the sequence (e.g.
GR or Vsh logs). Use the Vsh info and displacement
to calculate Shale Gouge Ratio at the reservoir- Geomechanical analysis
reservoir overlaps. Display this on the appropriate
areas of the juxtaposition diagram or triangle
diagram. Determine local in situ stress
orientation, and pore pressure.
Shale Gouge Ratio Estimate Sv from integration
of density logs, Shmin from
leak-off test data, and
Calibration for exploration/appraisal
or production behaviour? Production attempt to constrain Shmax.
Exploration Estimate the mechanical
Use the SGR pattern on the properties of probable fault-
In your dataset, do you have any pairs of wells juxtaposition (or triangle) diagram as zone material (from core
that provide information on pressure seal an estimate of fault-rock composition. samples if possible).
across a mapped fault?(e.g. RFT profile on Convert this to fault-zone permeability
both sides of fault, differences in hydrocarbon using published data, augmented by
contacts in juxtaposed reservoirs) local analyses (from core) of fault
zone material, if
Use ‘global’ calibration of SGR available.
yes no Compute Slip Tendency and
to estimate possible pressure
seal on your fault. Convert this Fracture Stability attributes.
Combine calculated fault-zone permeability
to column height using the with: fault-zone thickness (derived from
expected hydrocarbon density. fault displacement), juxtaposed reservoir
Depth datasets only: Plot the pore- permeabilities in simulation model, grid-
pressure field on each side of the fault, cell size in reservoir simulation model, to
using the relevant wells. At each calculate transmissibility modifiers for input Convert Fracture Stability
reservoir reservoir overlap, find the to the reservoir simulation model.
to maximum sustainanble
across-fault pressure difference supported column height, using the
Cross-plot the buoyancy pressure and SGR values, expected hydrocarbon
for each point of reservoir-reservoir overlap on the fault plane. density.
Use this calibration as an estimate of the potential pressure seal
at different SGR values, on other faults in the dataset.
Convert these to column heights using the expected hydrocarbon density

Appendix: Fault Seal Analysis: Traptester6 Training Course

Uploaded by

Copyright:

Available Formats

Appendix: Fault Seal Analysis: Traptester6 Training Course

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Appendix: Fault Seal Analysis: Traptester6 Training Course

Uploaded by

Copyright:

Available Formats

TrapTester6 Training Course

Appendix: Fault Seal Analysis

© Badley Geoscience Limited 2010

Appendix: Fault-Seal Analysis

Fault-seal analysis is not a discipline by itself. It is intrinsically linked to

This session presents an overview of the concepts, techniques and main

© Badley Geoscience Limited 2010

In an Exploration context a Fault Seals ...

© Badley Geoscience Limited 2010

1: Review of Capillary Fault Seals

Reservoir Fault zone Reservoir

(i) the fault itself has no sealing

In fact none of these assumptions is universally true. However, the

© Badley Geoscience Limited 2010

Adding extra stratigraphic detail

Vshale distribution at fault surface

Seismic-scale sequence Well-based picks

© Badley Geoscience Limited 2010

3: Fault Zone Processes & Products

- The composition of the wall-rocks

Cataclasites Shale smears - The stress conditions at the time of

- The maximum temperature

Predicting Fault-Zone Composition: Algorithms

© Badley Geoscience Limited 2010

Clay Smear: Predicting Algorithm (CSP)

© Badley Geoscience Limited 2010

Clay Smear Potential: Worked Example

Window of observation = throw / 2 = 15

CSP contribution (n=2, m=1)

© Badley Geoscience Limited 2010

Shale Smear Factor: Predicting Algorithm (SSF)

Shale rich fault-gouge zones

© Badley Geoscience Limited 2010

Shale-rich fault gouge: Predictive Algorithm (SGR)

Shale Gouge Ratio: Worked Example

0.02 FW contribution: ((2x0.08)+(6x0.5)+(2x0.01)+(2x0.45)+(3x0.02)) / 15

© Badley Geoscience Limited 2010

Key input for Fault-Seal Analysis: Vshale data

Detailed analysis of thin sections by point counting or by X-Ray diffraction

© Badley Geoscience Limited 2010

4: Calibration & Estimating column heights

They are estimates of the relative likelihood of clay gouge or smear

To use the attributes to predict the height of a hydrocarbon column,

© Badley Geoscience Limited 2010

Calibration & Estimating column heights (cont.)

Empirical: Calibrate SGR in-situ against known hydrocarbon columns

Deterministic: Assume that SGR = clay content of fault-zone. Use

© Badley Geoscience Limited 2010

Fault Seal – Leak: Observations

(Gibson & Bentham 2003)

Data from Brent Province, North Sea (Yielding 2002).

In-situ pressure calibration

© Badley Geoscience Limited 2010

Empirical Calibration: Seismic scale

From Bretan From Yielding

Pc = 10 ((SGR/27) – C) Pc = (17.5 * SGR) + C

Deterministic calibration: Core Scale

1: Predict fault-zone permeability (kf in mD) from

2: Convert fault-zone permeability to mercury/air