Download as pdf or txt
Download as pdf or txt
You are on page 1of 23

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/288935156

Reservoir connectivity: Definitions, examples and strategies

Article · January 2007

CITATIONS READS
6 4,241

7 authors, including:

John W. Snedden Peter Vrolijk


University of Texas at Austin New Mexico Institute of Mining and Technology
107 PUBLICATIONS   1,787 CITATIONS    91 PUBLICATIONS   3,644 CITATIONS   

SEE PROFILE SEE PROFILE

Liza Sumpter Michael Louis Sweet


Indiana University-Purdue University Indianapolis University of Texas at Austin
7 PUBLICATIONS   44 CITATIONS    37 PUBLICATIONS   860 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Gulf of Mexico Depositional Systems View project

Tuscaloosa Deepwater View project

All content following this page was uploaded by John W. Snedden on 05 July 2017.

The user has requested enhancement of the downloaded file.


IPTC 11375

Reservoir Connectivity: Definitions, Examples, and Strategies


John W. Snedden, Peter J. Vrolijk, Larry T. Sumpter, Mike L. Sweet, Kevin R. Barnes, Elijah White and Mike E. Farrell,
ExxonMobil Upstream Research Company

Copyright 2007, International Petroleum Technology Conference


field compartments. Analysis of dynamic connectivity is
This paper was prepared for presentation at the International Petroleum Technology essential to estimating ultimate recovery from a field.
Conference held in Dubai, U.A.E., 4–6 December 2007.

This paper was selected for presentation by an IPTC Programme Committee following review
of information contained in an abstract submitted by the author(s). Contents of the paper, as An example of RCA application in the Gulf of Mexico is
presented, have not been reviewed by the International Petroleum Technology Conference
and are subject to correction by the author(s). The material, as presented, does not
provided to illustrate how RCA can generate testable fluid
necessarily reflect any position of the International Petroleum Technology Conference, its connectivity scenarios and explain troubling production
officers, or members. Papers presented at IPTC are subject to publication review by Sponsor
Society Committees of IPTC. Electronic reproduction, distribution, or storage of any part of this anomalies.
paper for commercial purposes without the written consent of the International Petroleum
Technology Conference is prohibited. Permission to reproduce in print is restricted to an
abstract of not more than 300 words; illustrations may not be copied. The abstract must Introduction
contain conspicuous acknowledgment of where and by whom the paper was presented. Write
Librarian, IPTC, P.O. Box 833836, Richardson, TX 75083-3836, U.S.A., fax 01-972-952-9435.
Reservoir connectivity and its inverse, reservoir
Abstract compartmentalization, is a growing area of oil industry
research and business application, as large offshore
Reservoir connectivity, and its inverse, compartmentalization, discoveries go through development and established
is a critical area of petroleum industry research and business producing fields progress through their maturity cycle
application. However, significant differences in how it is (Smalley and Hale, 1996; Elshahawi et al. 2005). The size
defined, measured, and modeled exist among companies. For and scale of the associated monetary investments requires
some, connectivity is defined relative to an entity such as a three key foundations be established: 1) strict definitions of
well or set of perforations in a reservoir. Others prefer connectivity; 2) explicit workflows for analyzing connectivity;
reservoir connectivity indexes, using a set of often and 3) suitable analogs to be used as guides for investigating
subjectively defined criteria to gauge how problematic a field new cases. This paper provides each of these in order below.
will be to develop or exploit.
Previous Definitions of Connectivity
We have developed a technology called "Reservoir
Connectivity Analysis" (RCA) to investigate field Surveying the general terrain of reservoir connectivity reveals
compartments and associated connections (Vrolijk et al. significant differences among companies in how it is defined,
2005). A compartment is precisely defined as a trap which measured, modeled, and acted upon.
has no internal boundaries which would allow fluids to reach
equilibrium at more than one elevation. Compartment Schlumberger (2004) compared two unnamed North Sea fields
boundaries include sealing faults, channel margins, shale- (Fig. 1). Without specifying the connectivity measure, it was
draped clinoforms, paleokarst fractures and other diagenetic noted that the “highly connected” submarine fan reservoir and
boundaries. These can separate hydrocarbons and aquifers “poorly connected” deltaic reservoir fields differed
within a field or discovery. Connections between substantially in terms of estimated oil in place, ultimate oil
compartments include fault juxtaposition windows, erosional recovery (reserves), and cumulative oil production.
scours between channels, and capillary leakage. Compartment Particularly striking is the drop in oil reserves (ultimate
boundaries include spill and breakover points, defined on recovery) in the “poorly connected” deltaic reservoir field, just
topseal and baseseal maps. a short time after production start-up. By contrast, the highly
connected submarine fan reservoir had showed a steady climb
We also find that it is important to separately define and in oil reserves through time.
investigate "static" and "dynamic" connectivity. Static
connectivity describes the native state of a field, prior to For some petroleum and service companies, connectivity is
production start-up. Evaluation of static connectivity is the defined relative to an entity such as a well or set of
basis for proper assessment of original hydrocarbons in place perforations in a reservoir (Fig. 2). For example, a tabular
and prediction of fluid contacts in unpenetrated compartments. reservoir penetrated by a single well yields 100% reservoir
Dynamic connectivity describes movement of fluids once volume connected to the well, while more discrete,
production has begun. Initiation of production actually channelized reservoirs show lower connectivity (Ainsworth,
perturbs the original fluid distributions as pressure and 2005). However, drilling an additional well decreases the
saturation changes proceed in a non-systematic fashion across overall connectivity measure in most cases, as the more rapid
2 IPTC 11375

pressure decline imparted by two wells (all other factors being because they have failed to trap hydrocarbons, as per
equal) reduces sweep during production. In our opinion, this definition, but for practical reasons may be included in a
scale-sensitive approach tends to mix time spans, e.g., connectivity evaluation (e.g. associated aquifers, etc.)
geologic versus production time scales. It is also very
sensitive to the well spacing, completion technique, and Compartment boundaries are defined by structural features
pressure maintenance. It is our view that connectivity like faults, stratigraphic features like channel-margins, and the
measures should relate instead to the inherent geology of a limits of topseal and baseseal (Fig. 4). For example, a
reservoir and field: the structural and stratigraphic framework, compartment can be completely bounded by faults or partially
fluid type and distribution. bounded by faults with a reservoir pinchout (enclosed
reservoir compartment). Faults or reservoir pinchouts may
Other companies prefer reservoir connectivity indexes, using a only bound three sides of a compartment (open reservoir
set of often subjectively defined criteria to gauge how compartment) In some cases, oil may become separated across
problematic a field will be to develop or exploit (Harrison, a fault and a separate gas-oil contact (GOC) and/or water –oil
2004). Equally important is the confidence values associated contact (WOC) may develop, thus defining two different
with each estimate used in the index, and these vary compartments (Vrolijk et al. 2005).
considerably among different field and data (seismic) types.
Two other important features also represent compartment
Proposed definitions of connectivity and boundaries (Fig. 5). Spillpoints separate buoyant fluids on
connectivity elements structural highs and thus can allow division of compartments
containing different hydrocarbon/water contacts, for example
One approach to connectivity is to eschew subjective (Fig. 5a).
numerical values and let the buoyant fluids tell us what is
connected. First, one needs to differentiate between static or A less well-understood compartment boundary is the
geologic time scale fluid connectivity and dynamic or breakover point (Fig. 5b). A breakover point separates denser
production time scale fluid connectivity. It is well known fluids (e.g. oil below gas, water below oil) allowing
from our field studies that over geologic timescales, buoyancy development of separate contacts and thus compartments. It is
and capillary forces work to allow lighter fluids like oil and broadly defined as the deepest elevation at which buoyancy
gas to find their way into virtually any water-wet, porous forces can trap a relatively heavy fluid below a lighter one.
sandstone regardless of quality (e.g., Sweet and Sumpter, in Compartment boundaries are associated with baseseal and are
press). A schematic example shows how fluids migrate into thus best identified from top baseseal maps. Oil and water
the crest of a trap and fill downward, eventually crossing a pressures no longer communicate through the crest of the
fault plane where a relatively poor quality (low porosity and structure, assuming that there is lateral bounding to prevent
permeability) delta front reservoir facies is juxtaposed across the heavier fluid from communicating around the flank of the
the fault from high quality inner stream mouth bar reservoir structure.
facies (Fig. 3a). This continues until the spillpoint is reached.
Lateral bounding to set up breakover (a type of
However, our field studies also have demonstrated that compartment boundary) is probably more common that most
understanding reservoir type and quality is critical to operators assume. Laterally bounding of fluids has a variety
understanding dynamic connectivity as well (Pfeiffer et al. of causes (Fig. 6). It can occur due to primary faulting as
2000; Sweet and Sumpter, in press). For example, the 4D observed in Hibernia Field of Canada (Vrolijk et al. 2005).
seismic response (dark blue polygon in Fig. 3b), resulting Diagenetic changes (e.g. downdip loss of porosity, presence of
from either pressure or water saturation changes due to fluid impermeable karst towers, etc.) contribute to bounding of
movement, often occurs at the best facies juxtaposition across fluids. Channelization in both alluvial and deepwater realms
a fault plane, here in the case of a deepwater channel axis on can lead to lateral bounding if the channel margins and
axis. That is a critical difference between the two but both are overbank are largely impermeable (Sweet and Sumpter, in
important: understanding static connectivity is critical to press). Even shale-draped delta clinoforms are known to trap
appropriate in-place hydrocarbon estimation and dynamic and bound fluids.
connectivity for accurate determination of ultimate recovery.
Connections
Compartments and Boundaries
There are also a number of connections between
Like connectivity, compartments and their connections are compartments to consider while investigating reservoir
defined by the presence of two or three fluids (Fig. 4). A connectivity. An obvious connection between compartments
compartment is a trap which contains no internal barriers that is due to the juxtaposition of sand-on-sand at a fault plane.
would allow the contact between two fluids (e.g. oil and Assuming limited occurrence of shale gouge or cementation,
water) to reach equilibrium at more than one elevation sand-on-sand contacts tend to leak over geologic timeframes
(Vrolijk et al. 2005). Within a static compartment, the contact (James et al. 2004). Over production timescales, the
between two fluids will settle at a single elevation over transmissibility of the fault zone controls the rate and volume
geologic time (Sweet and Sumpter, in press). Water-filled of fluid leakage (Lescoffit and Townsend, 2005).
reservoirs are generally not considered “compartments”
IPTC 11375 3

Gas and oil can also leak out of the top of a reservoir Comparison to Available Fluid Data
through the seal rock, forming another potential connection to
an overlying reservoir compartment (Fig. 5c). The ideas To ensure internal consistency, candidate compartments and
formulated by Sales (1997) to explain leakage and spill in connections are checked against available fluid data. Rather
traps can be extended down in scale to the compartment level than establishing gas, oil, and water gradients by plotting all
(Fig. 7). The base assumption is that the filling of traps (and (and sometimes erroneous) pressure data, available
thus compartments) depends on the seal capacity relative to compositional data (from PVT reports, etc.) is used to
closure size. Large traps (and compartments) tend to leak gas independently calculate gradients (Vrolijk et al. 2005).
and oil until the buoyancy forces and capillary forces are Petrophysical data is evaluated and compilations of the
balanced (Class 3). In this case, there often is more oil than observed contact information (including GDT’s, OUT’s,
gas in a trap or compartment as a given seal rock has less ODT’s, WUT’s, etc.) are reviewed. Inconsistencies like
capacity to trap gas than oil. Traps (and compartments) with different fluid types present across an open fault juxtaposition
less closure size for the same seal rock capacity are prone to window are rectified (i.e. by adjusting the fault window
leak gas but trap oil (Class 2). If oil is added to a depths).
compartment due to continued migration or leakage from an
underlying compartment, it will spill out of a filled The RCA workflow can be still carried out in the absence of
compartment. Smaller closures (or improved seal rock pre-production pressure data, as is the case for many older
capacity for the same closure) will trap oil or gas and spill fields. In fact, it is quite useful to construct schematic
these out of a compartment if additional oil or gas is added diagrams of pressure versus depth to help construct geologic
over geologic time (Class 1). These traps tend to be gas and connectivity models. For example, observations of
prone, as gas is often generated last during source rock different WOC’s in a channel crossing an anticline (Fig. 9)
maturation and gas displaces oil (Gussow, 1954). suggest one of two possibilities: 1) hydrodynamic flow
causing a tilted WOC; and 2) a breakover point is set up by oil
The Sales (1997) model and the RCA workflow itself fill down to the top of the baseseal and lateral bounding
assume access to adequate hydrocarbon charge and a system (Vrolijk et al. 2005). Constructing the schematic pressure
in equilibrium (steady-state). These assumptions have versus depth plots for the two scenarios reveals two key
worked in the vast majority of fields examined to date (Vrolijk differences which can be tested. The hydrodynamic case
et al. 2005; Sweet and Sumpter, in press). requires both a thicker channel, communication of water
pressures across the crest of the structure, and a regional or
Other connections include channel-on-channel contacts semi-regional pressure gradient from right to left. The
where the appropriate sandy lithofacies is present in the two breakover scenario indicates two separate water lines but does
channels. Channel scour is an effective means of connecting not require regional water gradients or variations in sand body
two channels and is analogous in many ways to sand-on-sand thickness (Vrolijk et al. 2005).
juxtaposition across faults (Fig. 8). For example, muddy
drapes at channel-bases are thought to restrict fluid While well-documented cases of hydrodynamically tilted fluid
communication in the deepwater channel systems of Gulf of contacts do exist, it is our experience that many times the
Mexico deepwater fields (Barton et al. 2004). It is analogous cause is breakover (separation of denser fluids across a
to shale gouge in fault zones and fault relays, as both restrict structure and lateral bounding). Perched water (water trapped
fluid movement (which can be good or bad for well sweep in a structural low) is also more common than previously
efficiency, depending on the fluid breaking through). Our thought and often misinterpretations of field-wide fluid
experience is that relatively few sand-on-sand contacts (faults contacts and free water levels can result.
or channels) tend to seal over geologic time scales. However,
during production, fluid leakage is more difficult to predict Building a Connectivity Model
due to complex processes controlling the formation of shale
gouge in faults or shale drapes in channels (Fig. 8). The goal of any reservoir connectivity analysis is to construct
a model showing how a field is plumbed together. Often
Connectivity in channelized systems has previously been construction of a connectivity diagram is a useful means of
explored through simple experimental or empirical tabulating observations of fluid contacts and keeping track of
relationships between the number of sand-prone channels and compartments and connections (Fig. 10). Compartments are
overbank mud (expressed as net-to-gross or NTG). 2D represented by rectangles and connections by arrows of
modeling work has established a threshold value of 66% NTG various types. Of course, building such a model requires some
above which communication is expected (Allen, 1978), while simplification. Compartments here are shown here as
recent 3D models show a lower threshold value of 30% NTG rectangles, in this case defined by stratigraphic level (N1, N2,
(King et al. 1990; Larue and Hovadik, 2006). These etc.) and structural block (West Central Fault Block, etc). The
investigations fail to appreciate the complexity of connections are illustrated by single-headed arrows, assigned
compartments and connections described above. to the type of connection (cap leak, channel cut window, and
spill). Reservoirs that potentially have separate fluid contacts
are tabulated in this chart as compartments.
4 IPTC 11375

One way to think of reservoir connectivity is to hypothetically Field Application, South Timbalier-26, Gulf of Mexico
add a drop of oil to any compartment in the model and follow
it as it displaces hydrocarbons in the linked compartments We developed a workflow for investigating reservoir
finally to a system exit point such a fieldwide spill or capillary connectivity called Reservoir Connectivity Analysis (RCA) as
leak out of a topseal. In the case of the hypothetical oil field described by Vroljk et al (2005). It has been applied in cases
depicted by Fig. 10, a drop of oil added to Compartment N1 in ranging from discovered/undeveloped resources, early field
the East Flank (EF) will take the following path to the system life and mature fields. In mature producing fields, RCA has
exit point: explained troubling production anomalies, generated new drill
1. Leak via fault juxtaposition to compartment N5 in well opportunities, and boosted field reserves above original
the East Central Fault Block (ECFB). estimates. A case study of the Miocene O sand reservoir of the
2. Leak via capillary leakage from N5 to N2 South Timbalier Block-26 (ST-26) in the Gulf of Mexico
compartment in ECFB. illustrates how RCA and our connectivity definitions are
3. Spill to N2 in the west central fault block (WCFB) applied to: a) development of scenarios for pre-production
4. Leak via fault juxtaposition to the N1 reservoir in connectivity; and b) testing these scenarios against both pre-
the West Graben (WG) production (static) and post-production (dynamic) data. The
5. Spill to the system exit point updip of the WG data used are from publicly available sources (Farrell and
Abreu, 2006).
This analysis also reveals compartments that may be
shadowed from charge or that may be by-passed oil The ST-26 field is developed on the southern flank of a large
opportunities. In this example we find three unconnected salt dome (Fig. 11a). Salt traps hydrocarbons to the north
compartments (N2 in EF, N2 in WG, and N6 in ECFB). Of while radial NW-SE-striking faults trap laterally and internally
these, only N2 in EF is shown to be sourced (closest to source compartmentalize the field. In the “D” Fault block, the
kitchen). Miocene O sand traps oil in Miocene age fluvial and deltaic
reservoirs. Reservoirs are isolated vertically because of thick
Other potential compartments depend on hydrocarbon overlying and underlying shales. Initially, it was thought that
sourcing. N3 and N4 in the WCFB could be self-sourced a common oil-water contact (WOC) could be used for the
(transported organic matter), which explains their connection entire fault block. Subsequent development drilling revealed
to the system. N6 in ECFB requires charge from below or up significant variations in the WOC from well to well (Fig. 11b).
the fault plane.
The O sand had generally similar reservoir types, mainly
In most field cases, there is some uncertainty around fluvially-dominated deltaic lobes but it was apparent from
connections because of seismic resolution and/or well control. comparison of log motifs in the ST-26 #1 and #2 wells (drilled
The connectivity diagram can easily be modified to 2800 feet apart) that variable downcutting by overlying
accommodate different fault throws and create difference distributary channels had occurred. As illustrated in Fig. 11b,
connectivity scenarios. For example, changing the fault throw ST26#2 has thin distributary channel and #1 has a relatively
on the east flank could allow N6 in the ECFB to be juxtaposed thick one. It was not immediately understood how this could
against the N2 reservoir. Allowing for more erosion between impact the pre-production fluid contact elevations over just a
the N5 and N6 channels in the ECFB could link the two short distance.
reservoirs. The scenarios created are non-unique but testable
with additional drilling or examination of new data. Evidence of separate fluid contacts between wells indicates
that one of the two fluids; either oil or water is not in
The connectivity diagram constructed in Fig. 10 represents a communication between the two wells. Obviously, both
pre-production state and is particularly useful for estimating cannot be in communication otherwise the same fluid contact
original fluid contacts and original hydrocarbons in place. elevation would be present in both wells. The type of barrier
However, it is more difficult to estimate how the various was also problematic as faults were not recognized in wells or
connections will operate over production time scales. For seismic data. An additional concern surrounded the question
example, the formation of a shale gouge along the fault of whether this “barrier” would make a difference during
bounding the West Graben may limit recovery from wells production (dynamic connectivity).
drilled on either side of the fault. The presence of a shale base
drape between N3 and N4 channels in the WCFB may inhibit Using this case as an example of how to approach connectivity
communication there. Internal permeability barriers not seen problems, the first step is to propose some scenarios of how
in this gross reconstruction may also prove to be an issue the oil or water may have become separated over geologic
during production. However, our experience is that without time scales. As deltaic reservoirs are inherently architecturally
first understanding the pre-production distribution of fluids, complex, a series of stratigraphic models were generated (Fig.
illuminated through this work, it is even more problematic to 12). These represent “static connectivity” scenarios as defined
evaluate or predict production time scale field behavior. earlier.

In scenario 1, it was speculated that amalgamation of the


upper delta reservoirs allowed a fluid connection in the oil leg
but separation of water in the more distal delta facies (Fig.
IPTC 11375 5

12a). In general, this scenario assumes a limited aquifer OWOC. This delta lobe could provide considerable aquifer
volume. volume and strength to support pressures.

In Scenario 2, the connection of fluids was also presumed to Plotting GOR’s on the seismic facies map shows a clear
be in the oil leg but through an incised distributary or fluvial relationship with the mapped distributary channel trend (Fig.
channel sand (Fig. 12b). This model also incorporates a 15). GOR’s appear increase steadily away from the channel,
connection to a larger volume of aquifer below the OWOC, as suggesting that the wells near the channel are receiving aquifer
the channel sand extends downdip (during progradation or pressure support, assuming similar rates of production and
lowstand shifting). associated drawdown.

In Scenario 3, a similar channel is present, but located below The observed trends support connectivity scenario 2 and
the delta lobes (Fig. 12c). The fluid connection is through the argues against scenario 1 (Fig. 12). Scenario 1 presumes a
aquifer. This model also presumes access to a large downdip connection in the oil leg through toplap and amalgamation of
aquifer. delta lobes but a poor connection to an aquifer and thus poor
pressure support. By contrast, scenario 2: the connection is
The models described above can also be depicted in the also in the oil leg but through a downcutting distributary
connectivity diagram discussed earlier (Fig. 12). The different channel. This would result in a good connection to a downdip
scenarios for pre-production fluid connectivity are established aquifer and thus good pressure support.
by varying both compartments and connections.
To summarize, a series of connectivity scenarios were
The three connectivity scenarios form hypotheses which can established to explain differences in OWOC’s over short
be tested by evaluation of pressure data or other relevant distances where faulting plays a minor or no role. Use of early
measures of communication between wells and reservoirs. production pressure data and dynamic GOR information
Unfortunately, this publicly available dataset did not contain allowed testing of these different hypotheses. By eliminating
pre-production pressure data, which is usually the case for scenarios which are inconsistent with the observations, one is
older fields like this one (production began in the 1970’s). able to establish the most likely connectivity model for a
mature producing field. Mitigation strategies and infill
However, early production pressure data was available and drilling opportunities were not explored here, due to the fact
assuming limited initial drawdown, could be used as a proxy that the field was operated by another company. Yet the
for original pressures (Fig. 13). The early production pressure insights and workflow were significant and supported
data from the oil-reservoirs indicated that most wells had applications elsewhere.
roughly similar pressures during the first few months of
production, prior to significant drawdown. Wells D4, D11, Strategies for Application of Reservoir Connectivity
D15, and D17 all had pressures within 200 psi of each other Analysis
for the first 14 months of production. This is reasonably
similar, considering the quality of measurements from 1970’s We would suggest disciplined evaluation of reservoir
vintage strain gauge data. Thus, it is concluded that fluid connectivity should be carried out where appropriate. Clues
communication must be occurring through the oil column. or signs that this is needed include:
This deduction indicates that connectivity scenario #3
(communication through the aquifer) is not supportable (Fig. • Fields with multiple fluid contacts, even where
12). However, scenarios 1 and 2, which both allow compartments are not completely enclosed.
communication in the oil leg, are consistent with the data • Maps that show fault-defined compartments and a
above (Fig. 12). potential for multiple fluid contacts;
• Stratigraphic complexity (e.g. lateral reservoir
Differentiating between connectivity scenarios 1 and 2 bounding);
requires examination of more “dynamic” data like gas-oil- • DHI’s or AVO that don’t quite conform to
ratio’s (GOR’s). In the ST-26 Field, the O sand reservoir structure;
contains a near-critical fluid and significant production-related • MDT pressure data suggests limited vertical
pressure declines are usually accompanied by increased GOR. connectivity;
This significantly impacts oil production and overall oil • “Perched” or trapped water
recoveries. Co-mingling of reservoirs is limited in the field, • “Tilted” hydrocarbon/water contacts;
so changes in GOR reflect mainly pressure drawdown and • WUT’s above field/sub-regional FWL; multiple
degree of aquifer support in the O sand reservoir. contacts, unpenetrated compartments;
• Variations in HWC’s that cannot be explained by
Detailed seismic facies mapping (Abreu et al 2002) provided depth measurement error;
the rough geometry of the delta lobe reservoirs in Fault Block • A lack of wells penetrating a true HWC (e.g. many
D and the areal extent of the distributary channel cutting ODT’s)
through the delta lobes (Fig. 14). In fact, one could interpret
the channel connecting with a large delta lobe below the
6 IPTC 11375

Of course, economic criteria may also play a role in deciding Farrell, M. and V. Abreu, 2006, Reservoir Connectivity in
when to investigate connectivity (e.g, significant anticipated Fluvial/Deltaic Depositional Environments: South Timbalier
costs, need to arrest rapid pressure decline, etc.) 26 Field Study: AAPG Bulletin, v. 90 (2006). AAPG
International Meeting Program Abstracts (Digital).
Conclusions
Gussow, W. C., 1954, Differential Entrapment of Oil and Gas:
We have found that reservoir connectivity analysis is essential A Fundamental Principle: AAPG Bull., vol. 38, Issue 5, p.
to proper reserve estimation, realizing reserve adds from
816 - 853,
established fields, appropriate development, efficient
depletion. ExxonMobil developed Reservoir Connectivity
Analysis (Vrolijk et al. 2005) which involves: Harrison, C. J., 2004, Method for computing complexity,
– Strict definitions of compartments and connections; confidence and technical maturity indices for reservoir
– Integration of stratigraphic, structural, and fluid analyses; evaluations: United States Patent Application Publication, US
– Evaluation of pre-production fluid distribution as a 20040158406A1, 10 pages.
necessary foundation for understanding later production
behavior; Larue, D. K. and J. Hovadik, 2006, Connectivity of
– Application of fundamental geoscience and engineering channelized reservoirs; a modelling approach: Petroleum
practices. Geoscience. 12; 4, Pages 291-308.
In application work like this, success often comes from a
novel integration of diverse datasets, as we have done here. King, P. R., 1990, The connectivity and conductivity of
overlapping sand bodies. In, Buller Anthony T. et al.,eds,
North Sea oil and gas reservoirs; II, Proceedings of the North
Acknowledgments Sea oil and gas reservoirs conference. [Book, Conference
Document] Pages 353-362.
The authors would like to thank ExxonMobil Upstream
Research for permission to publish this paper. The review by James, W. R. et al., 2004, Fault-seal analysis using a
Ryan Ruppert is appreciated. stochastic multifault approach: AAPG Bull., vol. Volume 88,
Issue 7, p. Pages 885 – 904.
References Cited
Lescoffit, G., and C. Townsend, 2005, Quantifying the impact
Abreu, V., et al. 2002, Reservoir Characterization of the South of fault modeling parameters on production forecasting for
Timbalier 26 Field: The Importance of Shelf Margin Deltas as clastic reservoirs, in P. Boult and J. Kaldi, eds., Evaluating
Reservoirs in the Gulf of Mexico; AAPG BulletinVolume 86, fault and cap rock seals: AAPG Hedberg Conference Series 2,
Issue 13. (abstract) . p. 137 149.

Ainsworth, R. B., 2005, Sequence stratigraphic-based analysis Pfieffer, D. S., 2000, Mensa, Mississippi Canyon Block 731,
of depositional architecture-a case study from a marginal Gulf of Mexico-An Integrated Field Study: in Weimer, P., et
marine depositional setting: Petroleum Geoscience, v. 11, p. al. Deepwater Reservoirs of the World, p. 756-778.
257-276.

Allen, J. R L., 1978, Studies in fluviatile sedimentation; an Sales, J. K., 1997, Seal Strength vs. Trap Closure--A
exploratory quantitative model for the architecture of Fundamental Control on the Distribution of Oil and Gas:
avulsion-controlled alluvial sites: Sedimentary Geology, vol. AAPG Memoir 67: Seals, Traps, and the Petroleum System, p.
21; 2, Pages 129-147. 57 – 83.

Barton, M. et al., 2004, Understanding hydrocarbon recovery Schlumberger, 2004, Managing Uncertainty in Oilfield
in deepwater reservoirs; modeling outcrop data in the third Reserves, Middle East Well Evaluation Review: Volume 12
dimension. Anonymous In: AAPG annual meeting. (Abstract) http://www.slb.com/content/services/resources/mewr/mewr12.
Annual Meeting Expanded Abstracts - American Association asp
of Petroleum Geologists, vol. 13; p. 11.
Smalley, P.C., Hale, N.A., 1996, Early Identification of
Elshahawi, H., M. Hashem, O. C. Mullins, and G. Fujisawa, Reservoir Compartmentalization by Combining a Range of
2005, The missing link Identification of reservoir Conventional and Novel Data Types: SPE Formation
compartmentalization through downhole fluid analysis: Proc. Evaluation, v.11, p. 163-169.
2005 Society of Petroleum Engineers Annual Technical
Conference and Exhibition, Dallas, Texas, October 9 12, Sweet, M. L., and L. T. Sumpter, in press, Genesis Field, Gulf
2005, SPE Paper 94709, p. 1 12. of Mexico: recognizing reservoir compartments on geologic
and production timescales in deep-water reservoirs: AAPG
Bull.
IPTC 11375 7

Vrolijk, P. et al., 2005, Reservoir Connectivity Analysis - Common Abbreviations Used:


Defining Reservoir Connections & Plumbing: SPE Middle
East Oil and Gas Show and Conference, Kingdom of Bahrain GOC: gas/oil contact
#93577-MS. WOC: water-oil contact
OWOC: original water/oil contact
HWC: hydrocarbon/water contact
DHI: direct hydrocarbon indicator
AVO: amplitude versus offset
MDT: measurement downhole tool
WUT: water-up-to
ODT: oil–down-to
GDT: gas-down-to
OUT: oil-up-to
PVT: pressure, volume, temperature
GOR: gas-oil ratio
RCA: Reservoir Connectivity Analysis
a. Highly Connected Reservoir

b. Poorly Connected Reservoir

Figure 1. Importance of Reservoir


Connectivity/Compartmentalization.
Modified from Schlumberger (2004)
100% of reservoir volume
connected to the well

100% of reservoir volume


a. Connectivity connected to the well

relative to one 70% of reservoir volume


connected to the well
well 80% of reservoir volume
connected to the well

50% of reservoir volume


connected to each well

b. Connectivity 50% of reservoir volume


connected to each well
relative to two 10% and 70% of reservoir

wells
volume connected to the wells,
respectively
20% and 60% of reservoir
volume connected to the
wells, respectively

Figure 2. Connectivity measure relative to a) one well; and b) several wells. Modified from Ainsworth (2005).
Figure 3. Static versus dynamic connectivity.
A compartment is a trap containing no -1000

identified barriers that would allow the -1100

contact between two fluids to reach Compartment1


-1200

equilibrium at more than one elevation. -1400


-1300

-1500

Compartment boundaries are defined by: a. Compartment2

-structural features
-stratigraphic features -11000000
-
-limits of topseal or baseseal
--11220000 C1
C1 -1600

Spillpoint and Breakover points --11440000


C2
C1
also define compartment boundaries --11660000

--11880000
00
--220000

D
Figure 4. Definition of a reservoir compartment. a) Enclosed Map represents
reservoir compartment; b) Open reservoir compartment. Inset
shows schematic pressure versus depth plot for this example.
top of reservoir
b.
P
P

c. Capillary leak

D
al
s e
p
a. Spill to
Point
aquifer
overpressuring

b. Breakover point:
Highest point of base reservoir*
Oil and water pressures no
longer communicate through
crest of structure
Figure 5. Connections: a) spill, b)
breakover and c) capillary leak. *Assumes lateral Assumes efficient
Inset shows schematic pressure boundaries (e.g. channel source rocks and
versus depth plot for this example.
or fault boundary) filled compartments
• Lateral bounding of reservoirs sets up breakover, as fluids cannot
communicate around structure
• Lateral bounding can occur due to
– Structural causes: primary faulting
– Diagenetic causes (porosity loss, karst towers)
– Depositional causes:
• Channelization
–Alluvial
–Deepwater
Faults
– Delta Clinoforms

Dip view

Channel reservoir

Breakover point Shale seal


Strike SE Channel margins
View
Figure 6. Geologic factors causing lateral bounding.
Clinoforms
Crest of Trap

? Figure 7. Classification of traps by tendency to capillary leak or


spill. Filling of traps with oil and gas depends on seal capacity
relative to closure size. This is not a seal classification scheme, as
Max Gas size of closure relative to seal capacity controls leak or spill. In
Column RCA, this concept is extended to reservoir compartments, as cap
leak allows oil and gas to move upward until the capillary forces
and buoyancy forces are balanced. Modified from Sales (1997).
Max (Oil-dominated)
Column

CLASS 1 CLASS 2 CLASS 3

Pressure Pressure Pressure


Reserve Seal Strength (Oil)
Seal Strength Reserve Seal Strength (Oil) Top of Trap Seal Strength Top of Trap Seal Strength Gas and Oil Leak Top of Trap

Class 1 Trap Class 2 Trap Class 3 Trap


Both Oil and Gas
Spill (Nothing Leaks) Equilibrium
Gas/Oil Contact
Gradient of Maximum
Spill Point Possible Gas Column

Depth
Depth

Depth

Maximum Possible Maximum Possible Maximum Possible


Gas Column Gas Column Gas Column

Spill Point
Maximum Possible (Oil- Maximum Possible (Oil- Maximum Possible (Oil-
dominated) Column dominated) Column dominated) Column Nothing
Spilling oil Oil/Water Contact Reserve Closure
Spills

and gas Spilling oil Leaking oil and gas Spill Point

Leaking gas
Static Dynamic
Fault and Channel Connectivity Connectivity Connectivity

Channel 2 May leak or seal,


depending on
Generally presence of shale
leaks drapes at base of
Channel 1 channel (and
Channel base scour appropriate sandy fill).

May leak or seal,


Fault juxtaposition depending on
Generally
Leaks presence of shale
flow gouge in fault.

Figure 8. Schematic diagram showing how fault juxtaposition is analogous to connectivity at the
base of erosional channels. Our experience is that both generally leak fluids on geologic
timescales but can leak, seal, or baffle during production
0
10
A.

14 0
13 0
11 0

12 0
B.

Figure 9. Simple illustration showing difference between a hydrodynamically trapped oil accumulation (A) and alternative reservoir/seal geometric interpretation (B) that satisfies
same fluid pressure observations. Blue lines in (A) indicate potentiometric contours in aquifer. In both cases, map shows same distribution of well observations of oil and water and
same oil and aquifer pressures in wells. Note that reservoir isopach in hydrodynamic case (A) must be thicker at crest to allow continuous water stream to cross crest. An important
distinction between interpretations is hydrodynamic trapping requires aquifer flow, while breakover component of RCA permits but does not require flow of denser water phase.
Modified from Vrolijk et al. (2005)
A.

B.

Figure 10. A) hypothetical field compartments; B) representative compartment diagram


Salt
ST26 #2 ST26 #1
2800 ft apart

2900ms
D15

1 1

12 1
1 D17

3000 ms
1 D21
1 D28
1 1D11
D4

3300ms

South Timbalier 26

350
Time Structure (CI=100ms) OWOC
350
O Sand Fault Block D
0m
0m
s

s
A. 1 Mile
OWOC

Distributary channel

Figure 11. South Timbalier 26 Field, Gulf of Mexico. A. Time structure map, O sand
Delta lobe
reservoir, Fault Block D; B. Well log comparison showing variation in OWOC and thickness
of distributary channel facies in wells ST26 #1 and #2, located just 2800 feet apart. Modified
from Abreu et al. (2002).
B.
A.
• Scenario 1: Connection in oil leg
through delta lobe toplap and
amalgamation. Toplap/Amalg.

Aquifers*

B.
• Scenario 2: Connection in oil leg
through distributary channel
Erosional
Channel

Continues to aquifer

Downdip
Local Aquifers Aquifer

C. • Scenario 3: Connection in aquifer


through underlying fluvial channel
system.

Local aquifers
Figure 12. Three connectivity scenarios for ST-26 Fault Block D. with
Connectivity diagram format as discussed in text shown (insets). basal
*Normally not depicted but included here connection
for illustrative purposes
ST 26 Well Test Data
6800

Salt 6600

6400

6200

Pressure (psi)*
D15
1 1 6000
1
2
1
1
D17
D17

3000 m
s 5800
1
1 1 1 D21 D4
D28 D11
D4 5600
D11
D15
5400
s
3200 m D17

5200 D21
D28
5000

Jan-71
Jan-70

Apr-70

Oct-70
Oct-69

Jul-70
*1970’S VINTAGE STRAIN GAUGE MEASUREMENTS
• Figure 13. Early production oil pressure data, ST-26 Field, Gulf of Mexico:
A) Map of wells; B) Production oil pressures versus time. Modified from
Abreu et al. (2002).
Salt

D15
1 1

2 1 11
1 D17 C2

s
3000 m
1 D21
1D28 1 1D11
D4

South Timbalier 26
Seismic Facies Classification and Trace Shape
Seismic Facies Classification
350
0m
s Time Structure (CI=100ms)
O Sand Fault Block D
1 Mile

Figure 14. Delta lobes and distributary channels resolved using seismic trace classification technique. Interpreted distributary channels
shown by arrows (modified from Abreu et al 2002).
View publication stats

Salt

D15
D15
1
111
1

2 112
1 D17
D17
C2
C2

s
3000 m
1 D21D21
1D28
D28
1 1D11D11
C4
D4
D4
x
C18
x xC3 xC16
C13
x
x x C9
C7

South
South Timbalier
Timbalier 26
26
Seismic Facies Classification and Trace Shape
GOR
Seismic Facies Classification
Time Structure (CI=100ms)
0 - 1000 350
0m
s Time
GORStructure (CI=100ms)
of O Sand, Oct 1974
1000-2000
O Sand Fault Block D
2000+ 11 Mile
Mile

Figure 15. Comparison of GOR (gas-oil ratio) data to seismic facies maps of Fig. 14.
Modified from Farrell and Abreu (2006).

You might also like