Dynamic hydrocarbon trapping and implications for seal analysis

Christopher A.J. Wibberley (Total)

Vertical and lateral seals are often assumed to be static, with a fixed seal threshold determined by capillary

properties. When fault seals are considered in this context, the buoyancy pressure of the hydrocarbon column

is retained by the capillary membrane properties of the water-saturated fault zone. If trap filling continues

until the buoyancy pressure rises beyond the fault capillary threshold pressure, excess hydrocarbons are bled

off into the fault zone and may flow across and/or up the fault. In this way, the trappable hydrocarbon column

buoyancy pressure is always limited by the seal capillary threshold pressure, and estimates of fault capillary

threshold pressure are one way of evaluating trappable column heights in fault-bounded prospects.

Lookback analysis is often used to catalogue hydrocarbon column heights (or buoyancy pressures) trapped

against faults as a function of lithological-based parameters such as the shale-gouge ratio (SGR, Yielding et

al., 1997). These catalogues constrain algorithms which provide another way to predict maximum trappable

column heights in prospect evaluation. This presentation firstly reviews an example showing how lookback

analyses were used to compile algorithms for fault seal prediction in normally-pressured traps. However,

lookback analyses of nearby case studies of overpressured hydrodynamic traps do not fit these algorithms.

Furthermore, in active hydrodynamic environments vertical and lateral pressure cellularisation is common,

and significant pressure differences may exist across faults even in single-phase scenarii where capillary

trapping cannot be an issue. The key parameter here is therefore the permeability of the seal. Similarly, for

hydrocarbon columns which leak into fault zones when buoyancy pressure of the hydrocarbon column attains

the capillary threshold pressure, a continuous pathway of hydrocarbon through the membrane will lead to

relative hydrocarbon permeability close to that of single-phase flow. Continued active charge faster than the

leak rate will result in a larger column trapped against the fault, controlled largely by fault zone permeability.

The question then arises of whether the permeability of a fault zone can be sufficiently low to trap significant

column heights on the geological timescale (seal permeability is generally only considered important in

production contexts). This question is addressed by calculating how long it would take for leakage across a

fault zone to lead to a rise of 100m in free-water level (FWL) in a fault-bounded trap. Results in Figure 1

suggest that in some cases clay-rick fault zones can reduce the leak rate of Darcian flow of hydrocarbons on

sufficiently long timescales for this to be an important seal mechanism in active hydrocarbon basins.

Figure 1: Simple geometrical model for a fault-bounded trap with leak induced by an aquifer pressure potential gradient, with

results of calculations for the time taken for a 100m rise in FWL.

Examples are presented which demonstrate the use of applying such quantified hydrodynamic concepts in

predicting (1) aquifer pressure potentials in undrilled compartments, (2) charge rates into prospects across and

up faults and hence likelihood of fill-to-spill from active charging through faults, and (3) likelihood of

differences in fault sealing behavior between exploration and production timescales, important for identifying

possible compartmentalization during field development and production.

The first example, from the Caucasian region, is a direct application of hydrodynamic principals to prospect

evaluation (Figure 2). A pressure potential gradient is evidenced by two wells spaced around 10 km apart

(wells A and B in Fig. 2a), assuming laterally continuous reservoir properties. The lower-pressure well (well

B in Fig. 2a) is on one side of a faulted 4-way closure, but was water-bearing despite regional hydrocarbon

migration being thought to have followed the pressure potential gradient from the basin centre to arrive

against the fault in this 4-way closure. A prospect was defined in the same 4-way closure but on the other side

of the fault (prospect C in Figure 2a), with prospect evaluation relying heavily on aquifer pressure prediction

with a similar reasoning to that of hydrodynamic drive (Grauls et al., 2002; Grosjean et al. 2009). However,

this prospect did not have any nearby calibration wells for the reservoir aquifer pressure. The nearest well on

the same side of the fault was a further 30 km away towards the edge of the basin (well D in Figure 2a) and

close to hydrostatic pressure, suggesting either of two end-member scenarii: (1) a significant drop in reservoir

pressure potential across the fault, or (2) no or little pressure drop across the fault, but instead, a continuous

degradation in reservoir poroperm properties or reservoir continuity towards the edge of the basin.

Figure 2: (a) Sketch of the geological scenario of pressure potential (Ppot) decrease from basin centre (close to well A) to basin

edge (close to well D) including a strong pressure potential drop across the faulted 4-way closure in which well B and the

prospect are situated; (b) Darcian flow calculations for assessing the fault permeability required to explain the across-fault

pressure potential drop in scenario (1).

Prior to drilling the prospect no direct petrophysical data were available to distinguish between the two

possibilities. However, as in the first example, the question was asked “what fault permeability would be

required to explain the across-fault pressure potential drop assumed in scenario (1) ?” A simple 1-D steady-

state aquifer flow model is used (Figure 2b), assuming that the flux in (Qn) equals flux out (Qn+1), in each of

the flow units in Fig. 2b. From Darcy’s Law the problem can be reduced to one of a ratio of reservoir

permeability to fault permeability, for a given pressure potential decrease across the fault. The pressure

potential decrease across the fault in scenario (1) would require an across-fault permeability in the range

0.03µD to 0.3µD for a fault zone thickness from 1 to 10m. The impact of this calculation is two-fold. Firstly,

this range of values falls right on the range of across-fault permeability values predicted by Total’s fault

permeability algorithm for the SGR range of 30 – 50% estimated from well B, thus providing another

validation point for the algorithm. Secondly, the fact that the fault permeability values are thought to be

reasonable suggests that scenario (1) is likely to be the more appropriate one. This interpretation was

successfully tested when the prospect was drilled and found to contain a hydrocarbon column compatible with

the aquifer pressures suggested by scenario (1).

A dynamic context can also be created artificially during production by reservoir pressure depletion on one

side of a fault. Lookback analyses of exploration well pressure and contact data can often give a misleading

impression of how such fault seals may behave during production. This is because the long-term context of

fluid pressure distribution and possibly stabilized hydrocarbon charge may not reveal information on how

such seals could behave due to the introduction of a short-term, transient instability. The second example, of a

Tertiary turbidite field in conventional offshore W. Africa, is a case of potential compartmentalization by

anticline crestal-collapse faults not being recognized by exploration wells because identical pressure trends

and a communal OWC were detected across the field. Despite this, the distribution of pressure depletion after

production start-up suggests that this fault is causing a production timescale compartmentalisation effect. This

effect has been confirmed as being due to the fault both by pressure data in a later well drilled across the fault,

and by the 4-D signal which shows a break at the fault itself. This change in apparent fault seal behaviour can

be explained simply by the fault transmissibility from reasonable across-fault permeability values (Figure 3).

Figure 3: A pulse in across-fault pressure gradient by instantaneous charge or drawdown on one side of the fault, and the fault

permeabilities required to re-equilibrate the gradient over a given timescale.

Figure 3 examines the range of values of across-fault permeability required to produce this effect, with the

production timescale (~ 10 years) being too short to allow pressure re-equilibration which can nevertheless

occur over a long (geological) period of time (e.g. 100,000 to 10 million years). From simple 1-D flow

calculations, across-fault permeability values in the micro to sub-microDarcy range match this scenario

(Figure 3). Such a fault permeability range lies in between published values of fault permeability for fault

zones in purely sand-rich contexts and those with a high clay content (e.g. Wibberley & Shimamoto, 2005;

Saillet & Wibberley, 2013). Given the turbiditic depositional environment of the reservoirs, an “intermediate”

proportion of clay is expected in the fault zone, even without performing a detailed shale gouge ratio

calculation. This apparent match between calculation results and expected fault permeability from published

data provides support for the hypothesis proposed of the transient sealing behaviour of the fault during field

production.

Conclusions

Advances in quantifying fault zone permeability in the last few years allow us to test hypotheses for dynamic

flow through faults over geologically short and long timescales. It is shown here how this helps in

understanding seal mechanisms in lookback analyses as well as prediction of trapped hydrocarbon columns

and aquifer pressure distributions in hydrodynamic settings. For short timescales, production well data may

appear to diverge from initial pre-prod conclusions drawn from exploration well results. However, our

knowledge of fault permeability along with the help of simple calculations are used to show how these

apparent differences in behaviour do not conflict, but rather improve, our understanding of fault seal

behaviour. These case studies illustrate Total’s ability to accelerate the application of R&D results to

exploration studies through integrating structural geology and petrophysics in basin evaluation.

Acknowledgements

The authors wish to thank TOTAL for permission to publish the work.

References

Grauls, D., Pascaud, F. and Rives, T. 2002. Quantitative fault seal assessment in hydrocarbon-

compartmentalised structures using fluid pressure data. In: Hydrocarbon Seal Quantification. A.G. Koestler

and R. Hunsdale (eds.), NPF Special Publication 11: 141 – 156.

Grosjean, Y., Zaugg, P. and Gaulier, J.-M. 2009. Burial Hydrodynamics and Subtle Hydrocarbon Trap

Evaluation: from the Mahakam Delta to the South Caspian Sea. Paper presented at the International Petroleum

Technology Conference, 7-9 December 2009, Doha.

Saillet, E. and Wibberley, C.A.J. 2013. Permeability and flow impact of faults and deformation bands in high-

porosity sand reservoirs: Southeast Basin, France, analog. AAPG Bulletin, 97: 437–464.

Wibberley, C. A. J. and Shimamoto, T. 2005. Earth-quake slip weakening and asperities explained by thermal

pressurization. Nature, 436: 689–692.

Yielding, G., Freeman, B. and Needham, T. 1997. Quantitative fault seal prediction. American Association of

Petroleum Geologists Bulletin, 81: 897–917.

© Marc Roussel / Total

DYNAMIC HYDROCARBON TRAPPING and implications for seal analysis

Christopher WIBBERLEY, TOTAL

- Fault zone seal composition

- Examples of seal lookback analyses

- Static seal threshold pressure approach

1) LOOKBACK ANALYSES: HYDROSTATIC VS. OVP CONTEXTS

assumed to = % shale within the fault zone

SHALE GOUGE RATIO (SGR): CONCEPT

Throw

%Vcl1 . ∆t1

%Vcl2 . ∆t2

%Vcl3 . ∆t3

%Vcl4 . ∆t4

%Vcl5 . ∆t5

SGR = average Vcl past which a point has slipped

SGR = Sum [Vcli . ∆ti]

Throw

A

A’ B

B’

One SGR value per point

3

Siliclastic host series

FAULT SEAL THRESHOLD: EXAMPLE OF LOCAL CALIBRATION

Local calibration

• Cases where gas-bearing reservoirs forming the trap are juxtaposed against water-bearing reservoirs.

Threshold of maximum column heights

“Local” calibration based on a specific zone, using > 30 cases gas columns trapped against faults

FAULT SEALED COLUMNS IN AN OVERPRESSURE CONTEXT

5

• Cases where gas-bearing reservoirs forming the trap are juxtaposed against water-bearing reservoirs.

OVP cases Hydrostatic cases

Overpressured contexts: fault seal behaviour follows different rules

CAPILLARY SEAL CONCEPT – APPLICATION TO TOPSEAL

r

seal

reservoir Phc

Pe Pe

Pw

Hc Hc FWL

θ

Hc(max)

Fluid pressure

Dep

th

6

Phc Pe Pw

Hc Hc(max)

Phc

FWL

Hc Phc

Pe Pw

Hc Hc(max)

FWL

Hc Hc(max)

Phc

Pe Pw

Hc

FWL

Hc

Phc = (ρw- ρhc).g.H Pe : Capillary entry pressure

Pe = 2 *IFT*cos θ / r

Phc : HC-induced buoyancy effect

>

Cap rock Capillary leakage whenever: Phc > Pe

CAPILLARY SEAL CONCEPT – APPLICATION TO FAULTS

Phc(top) Pe(top) Pw

Fault entry pressure profile Topseal entry

pressure profile Phyllosilicate-rich fault zone of complex internal structure

Hc Fault zone water

saturated?

7

Hc FWL

Topseal

Fault seal

Phc(top)

Pe(top) Pw

Fault entry pressure profile Topseal entry

pressure profile Phyllosilicate-rich fault zone of complex internal structure

Fault zone water saturated?

Hc

FWL

Topseal

Fault seal

Hc

Phc(top)

Pe(top) Pw

Hc

FWL

Topseal

Fault seal

Fault entry pressure profile Topseal entry

pressure profile

Pe(FLP)

Phyllosilicate-rich fault zone of complex internal structure

Hc(max)

Hc(FLP)

Fault zone water saturated?

At Fault Leak Point (FLP): Pe(FLP) = (ρw – ρhc).g.Hc(FLP) Can not explain aquifer ∆Ppotential !

2) NEW APPROACHES : PERMEABILITY AND OVERPRESSURE

from Grauls et al. (2002)

Hc flushed through the fault Hc column “sealed” by aquifer pressure

Top Seal

Pw 2 Pw 1

∆Hc

F1 Channel OWC 1

OWC 2

Pw 2 Pw 1

∆ppot (aq) ∆ppot(aq) = ∆ρ. g. ∆Hc

Pressure equilibrium across the fault ρoil

ρaq

∆ρ = ρaq - ρoil

Pressure

Case study of pseudo-static Hc response to maintaint across-fault pressure equilibrium

HYDRODYNAMIC DRIVE THROUGH OVP

FAULT COMPARTMENTS : CENTRAL

NORTH SEA CASE 5

5 6

4

3 2

1

HYDRAULIC FRACTURE REGIME

NORMAL HYDROSTATIC PRESSURE REGIME Pressure (bars)

400 500 600 700

Dep

th (m

/sl)

3400

3600

3800

4000

S3

WEST EAST

From M. Sacleux & D. Grauls

‘DYNAMIC’ CONCEPTS: PERMEABILITY OF A FAULT SEAL

Presentation title - Place and Country - Date Month Day Year 10

1

100

10 000

1 000 000

100 000 000

1 10 100

1 mD

0.1 mD

10 µD

1 µD

0.1 µD

10 nD

1 nD

Clay-richfault gouge

Quartzo-feldsp.fault gouge

V(FWL)

θf θs d

A V(fault)

Time (years) Assuming steady-state flow across a 1m thick fault zone

• How long for across-fault leakage to cause a 100m rise in FWL?

Pressure potential difference (bars)

Permeability is a seal property relevant to “exploration” as well as production timescales

Fault permeability algorithms

x

y

z Fault zone

3) RECENT CASE APPLICATIONS

Well 1 Well 2 ∆OVP = 70bar

Q

Q A

1. North Sea case: aquifer pressure prediction

∆OVP1 = Xbar, L1 = xkm

Well-1 Well-2

Well-4 ∆OVP3 = Ybar, L3 = ykm

Prospect

2. CIS case: dynamic charge rates

APPLICATION 1: PROSPECT PRESSURE SCENARIOS

Range of Prospect A

U. Jurassic unconformity

Base chalk

Pressure potential (m)

Dep

th

Crest depth uncertainty

Well 2 aquifer

Well 1 aquifer

Approximate 1D calculation:

3m 1000m 100m

∆P1 ∆P2

∆PTOT = 70bar

Q Q Q Q

x

Prospect A

∆P1 = 0.02 bars ∆P2 = 69.98 bars

Results

Well 1

Q = k * A ∆P

µ ∆x

Well 2 ∆OVP = 70bar

Q

Q A

APPLICATION 2: DYNAMIC FILLING ESTIMATIONS

∆OVP1 = 14bar, L1 = 10km

Well 1 Well 2

NE

Flow restriction across low-permeability fault

What fault permeability required to support a given ∆OVPf ?

Wat

er h

ead

? OVP ~ 200b OVP ~ 190b

Well 4 (OVP 40 – 70b)

Q Q

Q

If: 10 m wide fault zone : kfault = 0.32 µD

1 m wide fault zone : kfault = 32 nD

Hydrodynamic gradient : 14m/kmGWC tilt : 22m/km (with TAF=1.53)

Azimuth: N10°W

BC i 4 P ti l f lt bilit K 0 005

Hydrodynamic gradient : 14m/kmGWC tilt : 22m/km (with TAF=1.53)

Azimuth: N10°W

BC i 4 P ti l f lt bilit K 0 005

2D water potential model (internal study 2011)

∆Pf 130 bars

Approximate 1D model:

A

10km 10m or 1m

∆P1 = 14bar ∆Pf

x k1 = 3mD

kf ?

If Q constant : Q = k1 * A * ∆P1 µ x1

kf * A * ∆Pf µ xf

=

IS THE FAULT PERMEABILITY ESTIMATE REASONABLE?

Fault permeability estimate is compatible with prediction from algorithm

10 m wide fault zone

1 m wide fault zone

0.32µD

32nD

nD range

mD range

FWL

PRESSURE SYNTHESIS

80 - 100 bars across-fault difference in gas legs

Max ~135 bars across-fault difference in water legs

Min ~110 bars across-fault difference in water legs

S crest against fault

Well 3 Well 2

N S

Main fault

Crests against fault

∆GWC

Static pressure potential

Nor

th c

ompt

Sout

h co

mptGas gradient

Potential (m)

Dep

th (m

ss)

Nor

th c

ompt

Sout

h co

mptGas gradient

Potential (m)

Dep

th (m

ss)

Sout

h co

mptGas gradient

Potential (m)

Dep

th (m

ss)

Gas leg in static equilibrium

∆OVP(fault)

N crest against fault

Down-dip spill to N

Down-dip spill to N

+100m uncertainty in spill depth

Pressure potential

(Grosjean et al., 2008)

Pre-drill view Post-drill view

16 -

DYNAMIC TRAPPING: FAULT COMMUNICATION BEHAVIOUR

Well 2

Appr

ox.

scal

e

xx00

xx00

xx00

xx00

xx00 m

Relay tip-controlled fault spill

Across-fault hydrodynamic leakage

Up-fault hydrodynamic filling

Well 3

Strong hydrodynamics: • No classic “fault sealing” methodology applicable • but pre-drill fault permeability prediction conforms to pressure compartmentalisation

Charge from basin Up-fault

charge

Strong hydrodynamic gradient

Seismic amplitude anomalies indicate possible filling down to

spill depth.

Quaternary fill-to-spill through fault is reasonable: across-fault -> 21TCF/Myr, up-fault 1 – 10 TCF/Myr

17 -

CONCLUSIONS

● Case studies of hydrodynamics allow us to advance concepts;

● Fault permeability is an important seal property in exploration as

well as production issues;

● Recently-developped fault permeability algorithms have been

implemented into new workflows where OVP / hydrodynamics is

important.

● Applications to pressure/seal evaluation include aquifer PPP and

dynamic charge/leak rates for prospect in-place volume risking

DISCLAIMER and COPYRIGHT RESERVATION

The TOTAL GROUP is defined as TOTAL S.A. and its affiliates and shall include the party making the presentation. Disclaimer This presentation may include forward-looking statements within the meaning of

the Private Securities Litigation Reform Act of 1995 with respect to the financial condition, results of operations, business, strategy and plans of TOTAL GROUP that are subject to risk factors and uncertainties caused by changes in, without limitation, technological development and innovation, supply sources, legal framework, market conditions, political or economic events.

The TOTAL GROUP does not assume any obligation to update publicly any forward-looking statement, whether as a result of new information, future events or otherwise. Further information on factors which could affect the company’s financial results is provided in documents filed by TOTAL GROUP with the French Autorité des Marchés Financiers and the US Securities and Exchange Commission.

Accordingly, no reliance may be placed on the accuracy or correctness of any such statements.

Copyright All rights are reserved and all material in this presentation may not be

reproduced without the express written permission of the TOTAL GROUP.