IPTC 14398

Validating of Reservoir Connectivity and Compartmentalization Through

the Use of the CO2 Compositional Gradient, Mass Transportation
Simulation, and Asphaltene Analysis
Saifon Daungkaew, Oliver C. Mullins, Zailily Johan, Eric Lehne, Julian Zuo, and Suresh Sinnappu,
Thomas Pfeiffer, Schlumberger, Tan Giok Lin, Tunku Indra B Tunku A Muthalib, Teh Yat Hong,
Hazwani Bt Rameli, PETRONAS Carigali Sdn. Bhd., John T Ludwig, Chulalongkorn University

Copyright 2011, International Petroleum Technology Conference

This paper was prepared for presentation at the International Petroleum Technology Conference held in Bangkok, Thailand, 7–9 February 2012.

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
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 necessarily
reflect any position of the International Petroleum Technology Conference, its 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 paper for commercial purposes without the written consent of the International Petroleum Technology
Conference is prohibited. Permission to reproduce in print is res tricted to an abstract of not more than 300 words; illustrations may not be copied. The abstract must 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 +1-972-952-9435

Abstract

More than 40% of the world's conventional gas reserves are in reservoirs that contain significant amounts of H2S and CO2.
The presence of these gases results in a number of challenges for the Field Development Plan (FDP). For a field with multiple
fault blocks with unknown fault transmissivity, a key challenge is to understand the field connectivity and
compartmentalization which impacts the ability to drain the reserves.

This paper presents a comprehensive study to understand reservoir connectivity in a gas and oil field located in Southeast Asia.
This particular field has variation of the CO2 content even in the same zone ranging from less than 10% to more than 80%wt.
A key for production strategy and facility design is to be able to accurately quantify the CO2 in each reservoir. Initially, the
CO2 study aimed to quantify the CO2 content for each reservoir using an advanced Downhole Fluid Analyzer (DFA), and then
to use the DFA measurement as well as the available PVT data from nearby wells to understand reservoir connectivity through
the use of a compositional gradient concept. Reservoir fluids that deviated from the compositional gradient were considered to
not be in equilibrium with each other, and subsequently indicate possible compartmentalization. In addition, the use of a PVT
thermodynamic calculation with a non-isothermal solution resulted in a possible identification of the CO2 charging location.
Although the geochemistry study was conducted to understand the source of CO2 and hydrocarbon isotopes, it did not provide
a conclusive result of the reservoir connectivity. As expected, the hydrocarbon and non hydrocarbon gases had different
charge sources. Since the geochemistry study was inconclusive, a mass transportation simulation was performed to understand
the reservoir connectivity and this information has had a great impact on understanding the production mechanism of this field.

In addition to the discussion of the horizontal connectivity between wells in the reservoir, a preliminary study of the asphaltene
composition and concentration in one of the wells was provided as part of this paper to determine the extent of vertical
connectivity between the thin beds. Specifically, the effect of the asphaltene content on the reservoir fluid’s optical properties
versus depth was analyzed and is presented in this paper.

A systematic process is provided to understand the reservoir connectivity by using integrated reservoir data including pressure,
DFA, PVT fluid properties, geochemistry, asphaltene science, as well as the geological and geophysical interpretations of the
reservoir. This paper offers an efficient way for reservoir characterization for proper field management for an important
hydrocarbon discovery in Southeast Asia.

Introduction

Carbon dioxide and Its Origin

The occurrence of CO2 in gas fields can be of primary or secondary origin, or a combination of both. For example, primary
origin(s) of CO2 are1,3,4,6-9,15-17,19;
2 IPTC 14398

• mantle and crustal emanations

• carbonate metamorphism and hydrolysis
• maturation of source rocks, particularly coals, under high thermal stress
• microbial activity

In addition, secondary physico-chemical partitioning processes, such as solubility-dispersive interactions may play a role in
increasing the concentrations of CO2 in reservoirs. In any event, it is very difficult to characterize migration pathways of the
CO2 in the reservoir as there can be multiple sources for CO2 having different migration pathways over geological times1,3,4,6-
9,15-17,19
.

Literature Review

There have been a number of papers written that address the importance of identifying reservoir connectivity and
compartmentalization, and the use of pressure and fluid data to identify these issues. Mullins et al. (2005) used an analogy to
describe reservoir architecture as being like a kitchen sponge (open-cell), with connected compartments that would allow fluid
to flow freely between compartments12. Conversely, on the other end of the desired reservoir spectrum, the architecture could
be described like bubble wrap (closed-cell) which could also contain a lot of reservoir fluid, but that fluid would be trapped in
individual compartments unable to flow. Furthermore, separate from the question of reservoir architecture. Canas et al.
(2008) wrote that a complication in analyzing reservoirs is that several reservoir processes can be responsible for fluid
compositional variations2. Mullins wrote that the key to identifying reservoir complexities was to focus on the composition of
the reservoir fluid at reservoir conditions. This breakthrough by Mullins brought about the development of wireline tools
utilizing Downhole Fluid Analysis (DFA) tool in recent years. The DFA is a concept that utilizes optical spectroscopy, gas
refractometer, florescence detector, and vibration rod to obtain PVT level quality measurements downhole in real time.
Focusing on the reservoir fluid composition, Daungkaew et al. (2008) wrote that an understanding of the distribution and
quantifying of CO2 content is important for production strategy and facility design. Furthermore, certain aspects of a reservoir
fluid’s CO2 content and distribution can also indicate reservoir compartmentalization5. Along with measuring several other
reservoir fluid properties in real time, the current generation of DFA concept tools can obtain highly accurate measurements of
CO2 content in the reservoir. McKinney et al. (2007) wrote that in evaluating a reservoir fluid, no technique should be used in
isolation and that integration of all the available data including understanding the geological setting, fluid sourcing, migration,
and trapping mechanism is critical10.

Challenges For This Project

In offshore Peninsular Malaysia, Oil Field X has a big structure covering an extensive area with an East-West elongated
anticline and North-South trending faults. This particular oil and gas field is in relatively shallow waters of about 60 meters in
the central portion of the Malay Basin. The target reservoirs are comprised of channel sandstones and shallow marine sands
with interbedded mudstones consisting of transgressive shales and marine shales. The reservoir fluid is known to have a large
variation of CO2 concentration, waxy, viscous, and low Gas-Oil-Ratio (GOR) based on the early appraisal well production
tests. Currently, the field is undergoing a drilling program for a series of appraisal wells to acquire additional reservoir data
and reservoir delineation. Apart from the data quality, cost and timing are the key consideration for getting early reservoir
fluid properties as the field is geared for the early first oil. The challenge is to obtain critical reservoir fluid properties such as
CO2, GOR, fluid density, and fluid compositions (i.e. C1 to C6+) efficiently. This information is required for reservoir
modeling and it can be used to confirm reservoir connectivity in terms of fluid communication between fault blocks as
previously interpreted in the structure map.

FB III
FB VII
A-5
A-3 AS-1 A-8
A-8ST
A-1
A-3ST
A4ST A-6
FB I A-4 FB VI
FB IV FB V
FBII
A-2 A-6

30 SPE116501

Figure 1: The Structure Map of Oil Field X.

IPTC 14398 3

There are numbers of challenges for this reservoir:

1. Reservoir connectivity and compartmentalization is the biggest challenge for proper Field Development Plan (FDP).
The seismic interpretation first suggested multiple fault blocks in this field. The connectivity between each fault
block in terms of pressure and fluids are crucial information for this appraisal campaign.
2. This particular field was first discovered as a gas field with variation of CO2 concentration. The oil rim was
encountered after first few wells were drilled. Reservoir fluids in both gas and oil reservoirs have been found to have
varying of GOR, and fluid composition including CO2. The fluid property variations have been observed even within
the same sand. The CO2 contents needed to be quantified for each reservoir depth in order to understand the
distribution of CO2 versus depth. This measurement may lead to a better understanding of the CO2 charging in this
field.
3. Thinly bedded zones, so fluid identification from conventional logs were not conclusive. Standard resolution logs
and pressure gradient data cannot be relied on for fluid identification. The pump-out needs to be conducted with the
Formation Tester (FT) tool in order to conclusively identify reservoir fluid zone by zone.
4. The wells were drilled with WBM. CO2 quantification has been challenging with the previous generation of the DFA
tool (i.e. Compositional Fluid Analyzer (CFA*) when water is in the flowline5.
5. The complex hydrocarbon and CO2 charging paths and sources for this particular field.
6. The connectivity between the gas cap and the oil reservoir considering the variation of CO2. The different
connectivity possibilities would result in different drive mechanisms for the FDP.

Scope of the Downhole Fluid Analysis Study

Due to several challenges in this particular field, the Downhole Fluid Analysis (DFA) Study was proposed together with the
FT tools to this operator to understand these challenges. First of all, the data acquisition process and results will be discussed.
An advanced DFA tool, i.e. InSitu Fluid Analyzer (IFA*) that can provide accurate downhole fluid composition including CO2
was deployed and the real time PVT parameters were then compared with the PVT laboratory analysis that became available
later and also with the sampling collected during the Drill Stem Test (DST). This paper does not intend to discuss operational
aspects in detail since that was presented previously in the SPE 1165015. First, only a fluid comparison between different
methods is discussed before the other studies are introduced one by one to investigate each challenge for this particular field.
Second, the Equation of State (EOS) was then used to understand the connectivity between fault blocks. Third, the
Geochemistry was proposed to understand the CO2 Isotope from this field. Finally, the use of mass transportation modeling
was then used to model the CO2 variation versus depth. The conclusion from this simulation was also used to suggest reservoir
connectivity between gas cap and oil leg in one particular reservoir.

Data Acquisitions and DFA Results

With the current technology of FT, one method of quantifying CO2 in real time is by using near-infrared (NIR) optical
spectrometers mounted along the FT flowline. However, the main challenge for in-situ CO2 is the effect of water spectrum
overlapping the CO2 region and saturating the signal11. As a result, a new DFA tool which aims to measure CO2 needed to
have certain improvements to address these issues.

C o m p o s itio n fro m th e I F A to o l

C o m p o sit io n fr o m th e C F A to o l
Optical density

W a t e r fra c ti o n

Wavelength, nm

Figure 2: One of the main challenges here is to
quantify CO2 while there is presence of water in the
system. In spectrometer, the peak of CO2 optical
density can be hidden if there is some water the
system since water peak has similar wavelength as
the CO2 peak.
Figure 3: Fluid scanning in one of the gas sampling stations: fluid
composition measurement obtained from IFA is in the first row and
composition measured from CFA is in the second row. The water fraction is
shown in the third row. This figure shows clearly that the CFA (previous
generation of the DFA tool) can measure CO2 when the water fraction is
minimal, whereas the IFA can measure CO2 even when the water fraction is
high [Ref 5].
4 IPTC 14398

In order to obtain CO2 data from the reservoir, we employ the advanced DFA tool, i.e. IFA, coupled with a formation tester
tool. The idea is to characterize the CO2 vertically and laterally in the reservoir. This is best done by performing DFA stations
along with a few sampling stations. With this IFA tool, the CO2 can be quantified in real time, and is shown in Tables 1 to 3.

Table 1: Comparison of CO2 measurement from the DFA, FT samples, and with the samples taken during the production test facility.

MDT Sand Depth Fluid MDT DFA MDT Lab DST Onsite DST Lab
Unit (m MD) Type CFA IFA Total MDT Fluid Flash Gas
File no. Sample Separator Gas Separator Gas
% wt % wt % wt % mole % mole % mole % mole
1 T1 xx44.0 Gas 65.00 64.40 70.40 52.08
2 S1 xx14.5 Gas 73.00 69.00 72.59 53.22
3 S1 xx66.5 Gas 63.00 60.00 61.80 41.47
4 S2 xx97.5 Oil N/A 7.50 6.20 21.70 47.7 47 49
5 S3 xx21.0 Oil N/A 1.80 3.30 11.90 29.8
6 S3 xx21.0 Oil N/A 2.00 2.90 10.80 28.8 24.8 29

This particular well A-4 that is located in Fault Block (FB) II was logged in 2007 with the IFA field test version while the PVT
lab results were unavailable until mid 2009. Lab results of the Modular Formation Dynamic Testers (MDT*) show an
excellent agreement between DFA and DST analyses. However, DFA measurements (from both Compositional Fluid
Analysis (CFA*) and InSitu Fluid Analyzer (IFA*)) offer the downhole CO2 content, whereas the DST provides CO2
agreement to the flash gas from the downhole hydrocarbon. The DFA measurement, on the other hand, provides accurate
downhole CO2 measurement. The limitation of the previous DFA tool, i.e. CFA, is when the CO2 content is less than 40%wt.
In addition the CFA cannot detect the CO2 if there is some water in the flowline, as shown in Figures 2 and 3. However, with
more precise wavelength at CO2 peaks, the IFA can provide accurate CO2 measurement as low as 2%wt. In addition, two
DFA stations with the same depth confirm an excellent repeatability of the CO2 measurement even in the very low CO2. Not
only does the IFA provide the in-situ CO2 measurement, this tool also provides fluid composition, i.e. C1, C2, C3-C5, and C6+,
GOR and also density. Tables 2 and 3 show a comparison of fluid composition between the in-situ and lab measurements.
The GOR measurement for the IFA covers low GOR oil <200 scf/bbl up to gas condensate ranges. As a result, for the dry gas
case with high CO2, the GOR from Table 3 is approximate in comparison to the lab measurement. For black oil cases the
GOR also matched the lab quite well.

Table 2: Comparison of fluid composition obtained from IFA and lab measurements.

CO2, wt% C1,wt% C2,wt% C3-C5,wt% C6+,wt%

Sandface Depth, m MD Fluid In-situ LAB In-situ Lab In-situ Lab In-situ Lab In-situ Lab
T1 x244 Gas 64.0 70.4 29.7 20.7 1.0 2.4 0.0 3.0 4.9 2.9
S1 x414.5 Gas 63.0 72.6 22.1 20.9 0.6 2.2 0.0 2.0 4.5 1.8
S2 x466.5 Gas 60.0 61.8 29.1 28.5 0.7 2.8 0.0 3.2 7.3 3.0
S3 x497 Oil 7.5 6.2 2.2 1.9 0.5 0.4 0.9 1.1 88.8 90.3
S4 x521 Oil 1.8 2.9 3.2 2.0 0.4 0.4 1.1 1.3 93.5 93.7
S4 x521 Oil 2.0 3.3 3.1 2.3 0.4 0.4 1.4 1.1 93.1 92.8

Table 3: In-Situ measurement of density from DFA was used to identify reservoir fluid and confirm pressure gradient.

Density (g/cc) GOR, scf/bbl

Sandface Depth, m MD Fluid Gradient In-situ Lab* In-Situ Lab
T1 x244 Gas 0.11 0.14 N/A 79,825 184,097
S1 x414.5 Gas 0.19 0.19 N/A 78,630 297,407
S2 x466.5 Gas 0.16 N/A N/A 52,087 175,795
S3 x497 Oil 0.76 0.78 0.82 443 367.6
S4 x521 Oil 0.78 0.77 0.82 334 277.5
S4 x521 Oil 0.97 0.80 0.81 337 304.4

The data above is from the logging campaign in 2007. In the second campaign in 2009, two more wells and one sidetrack
were logged, i.e. A-6 in FBVI and A-8 main hole in FBIII and A-8 ST in FBIV. With confidence in the DFA measurement,
this operator requested use of the in-situ fluid data in real time for all hydrocarbon zones. No PVT quality sample was
required to send to the lab for reservoir fluid property measurement. Only two PVT samples were required for a flow
assurance study. In addition, the in-situ CO2 from the IFA tool allows the geoscientist team to make decision on the facility
selection material without conducting the DST for this campaign. The result from the second campaign is shown in Figure 4.
From this figure, it shows clearly that there is a significant compositional gradient in this particular reservoir. Even though the
oil column is only 55 meters, the GOR varies from 240 scf/bbl at the bottom to 477 scf/bbl at the top of oil column. Another
interesting point is that CO2 varies significantly from 62%mol at the top of gas column to 26.7%mol in solution gas at the
GOC to the 1.5%mol in solution gas at the bottom of the oil column (at the OWC).
IPTC 14398 5

Figure 4: With enough reservoir fluid information, reservoir fluid for this particular reservoir shows more compositional
grading than what client expected, i.e. change GOR, Pb with depth even in the oil with 55 m thick.

Using Equation of State (EOS) to Understand Reservoir Connectivity and Compartmentalization

In this second step of this DFA study, the main objective is to use pressure and fluid data to prove pressure and fluid
communications between each fault block. The EOS was generated from the available PVT data from the nearby well, and
then compared with the IFA data acquired in the current well in 2009. This methodology helps to identify a possible
communication between multiple faults. In other words, it helps to prove whether these faults are sealing or non sealing from
both pressure and fluid properties. The most important aspect of this EOS work is that when the EOS was generated from the
previous PVT data, a compositional grading due to temperature effect needs to taken into account in the EOS equation. This
was done using the PVTPro* Software. The actual workflow is shown in Figure 5. Figure 6 shows the cartoon of data
availability and the location of each well in different fault blocks.

EoS PVT Pro Field X

Avalaible Field Data
Tuned EoS model WFT, PVT ncompositions
A5S8 Block III
PVT
A3S3 IFA 2009
PVT Lab Composition
~ match
Campaign
Lumping Compositions
DFA logs on A 4
FB I
Block I
Compositional Grading DFA on 2nd Campaign, 2009 A4S4 IFA Field Test 2007
A8 and A8 ST1
FB II
Block II Block IV

Compositions & Pf Trends PDPlot Multiwell

Reservoir Connectivity

FDP Planning

Figure 5: The workflow of this study which takes into account the compositional Figure 6: Cartoon for the well location
grading due to temperature when generating the EOS. The PVTPro software was simplified from Figure 1.
used to generate the EOS for this study.
6 IPTC 14398

The well A3 (in Fault Block, i.e. FB I) has the complete PVT data. Well A4 (in FB II) has the DFA data obtained from the
IFA in 2007 and the preliminary PVT report. Well A8 is the new well that was planned to log in the 2009 campaign. The
workflow is to use EOS data derived from well A3 (in FB I) to predict fluid composition including CO2 contents to wells A4
(in FB II) and A8 (in FB III), and then compare to IFA data real time. This study aim to prove/confirm reservoir
compartmentalization between fault blocks I, II and III.

Compostion Grading Prediction on Block III Pressure MDT Pressure

Predictive Pressure
-4785
2090 2095 2100 2105 2110 2115 2120 2125 2130 2135 2140 2145 2150
-4790

-4795

-4800
D e p th F T T V D S S

-4805 MDT Oil Gradient = 0.277 PSI/FT

Predicted Oil
`
-4810 Gradient =
0.307 PSI/FT
-4815

-4820

-4825

-4830

-4835
Pressure, Psia

Figure 7: Understanding Pressure Connectivity in the X field for Block I and Block III. The predicted and measured pressures
match.

Figure 7 shows the prediction of formation pressure and pressure gradient and the actual measurement from the FT in the new
well (in 2009). The fluid data from FBI was used to generate pressure gradient in FBIII, and the result suggests pressure
communication between FBI and III. In addition, when the EOS generated from PVT data in the FBI (Well A-3) was used to
match the IFA data from FBII (Well A-4), the composition for each component matched quite well including CO2 content, as
shown in Figure 8. However, the EOS generated from FBI (well A-3) does not match well to the new well A-8 located in
FBIII. This suggests a possibility of CO2 charging between nearby FBIII. This is also supported by Figure 10 where all wells
on the right-hand side of the map are gas and have high CO2.

Comparison Between PVT Composition, IFA data and Compositional Grading Prediction in FBII CO2 Charging Possibility: Comparing CO2 between Block I, Block II and Block III
100
100 100
90
90 90
80
80 80
70
70 70
Biggest Impact on FDP in handling CO2
60 60 60
w t%

w t%

50 50 50

40 40 40

30 30 30
Good Match indicates same fluid in FBI and FBII
20 20 20

10 10 Possibility of CO2 Charging in Block III

10

0 0
0
CO2 C1 C2 C3-C5 C6+
CO2 C1 C2 C3-C5 C6+

Figure 8: Understanding Fluid Communication: Good match Figure 9: Possibility of the CO2 charging between blocks. FBI
indicates same fluid between FBI and FBII. agrees very well to FBII. However, FBIII has significantly high
CO2 compared to other fault blocks.
IPTC 14398 7

FB III
FB VII
A-5
A-3 AS-1 A-8
A-8ST
A-1
A-3ST
A4ST A-6
FB I A-4 FB VI
FB IV FB V
FBII
A-2 A-6

30 SPE116501

Figure 10: The fluid composition in one single reservoir. All wells in the right-hand side of this field have significantly high CO2
compared to the left side.

Geochemistry for Isotope Analysis

All gas samples show similar composition of hydrocarbon (H.C.) gases. These hydrocarbons are wet thermogenic gases.
Hydrocarbons were generated from lacustrine or marine shales at higher thermal stress. The gases are likely mixtures of
different charges having slight different thermal maturities.

Ar and N2/O2 content of gases show linear correlation indicating a common origin for non-hydrocarbon gases. Constant He
content suggests that those gases are non-magmatic in origin. The geochemical study shows that CO2 from thermal
breakdown of carbonates (heavy isotopic signature) and CO2 from coals (light isotopic signature) generated under high
thermal stress have mixed in different proportions, which resulted in variable CO2 abundance and isotopic characteristics in
the field. The interrelationship between CO2 content and δ13C of CO2 of samples suggests that in addition processes, such as
driven by carbonate precipitation, were active to alter the two in tandem.

The isotope analysis suggests different origins for hydrocarbon and non-hydrocarbon gases. As a result, this study suggests
different migration paths for the hydrocarbon versus CO2 for this particular field.

Mass Transportation

The geochemistry study suggests different origins for H.C. and non H.C. gas including CO2. This implies two different
charges into one reservoir. However, another challenge for this FDP is whether the gas cap connects to the oil rim in a single
reservoir (as shown in 4). The oil has a low GOR in the range of 477 scf/STB to 243 scf/STB. A large gradient in CO2,
shown in Figure 4, is revealed by DFA and confirmed by laboratory work. A relatively high concentration of CO2, above
60%mol, is found in the gas cap at different locations in the reservoir. The CO2 concentration in the oil column is
monotonically declining down to concentration of only 2%mol at the bottom. GOR data and pressure gradients are also shown
in this figure. The gradient in the western direction of the gas cap is monotonic with a falling CO2 concentration with depth.
It stretches across two faults. A plausible explanation for this gradient would be a charge of CO2 into the gas cap that diffuses
down in the oil column. Using this monotonic CO2 gradient as a benchmark, a one-dimensional diffusion model is set up to
back this hypothesis.
8 IPTC 14398

A number of variables can be adjusted to make a match between the real gradient and the simulated gradient. Most of those
variables are unknown. The simulation will give at best an approximation of the zeroth order. A strategy can be derived from
this approximation to study higher order effects and essentially improve the understanding of the real reservoir architecture.
The model’s dimensions are an educated guess on the lateral diffusion length. The height of the oil column is known to be
55m TVD. Lateral distances between the wells increase the diffusion length. A column height of 500m is therefore not
unreasonable. The model is tilted 45 degrees to introduce gravity effects and reflect the angle of the geologic structure.
Porosity and permeability are assumed to be uniform throughout the model. A series of sensitivities have been evaluated to
see the effect of initial composition, diffusion length and the influence of barriers. Figure 11 shows visualization the model.

Figure 11: Titled vertical one-dimensional model to investigate Figure 12: Base case simulation without barriers. The red
CO2 charging into an oil column. crosses are the data points as measured in the field. A good
match between simulation and data is found in the 250000 year
gradient line.

The initial composition of the gas and the oil is unknown. The gas cap is initialized with a gas of high CO2 content, similar to
gas found in the Malay Basin field. Faults manifest themselves in step changes in the gradient. The magnitude of the step is a
function of time and transmissivity of the fault. From the field measurements, the CO2 concentration at the bottom of the oil
column is 1.5%. Well A-4, in the middle of the oil column, has a CO2 concentration of 11.4%. Well A-8 shows a CO2
concentration of 27.6%, just below the GOC. On the map each of these wells is situated in different fault blocks.

A base case simulation model assumes all faults are 100% transmissive, i.e. no barriers are present. It is set up with a high
initial CO2 concentration of 60% in the gas cap and zero CO2 content in the oil column. The largest gradient after around
250,000 years lines up well with the data measured in the three wells (see Figure 12).

Two faults are introduced in the model between the wells. The transmissivity of the faults was varied over a range from 100%
to 0.01%. Two plots are presented here. In each plot one of the gradients lines up with the data points. This suggests that a
match like in the case without faults can be found in the presence of faults. Figure 13 shows the resulting gradients for a fault
transmissivity of 1%. Figure 14 presents the gradients that result from faults of 0.01% transmissivity. The model can line up
the gradient to the data points in either case – transmissive faults or sealing faults, yet under reasonable assumptions one
scenario is more likely, to occur than the other.

The CO2 in such high concentration is unlikely to come only from kerogen. Thus the hydrocarbon and the CO2 charge into the
reservoir come from two different locations. Both simulation cases (faulted and unfaulted) introduce the CO2 from the top.
Charging the CO2 exclusively from the top makes sense for the unfaulted case. If the entire reservoir is connected the light
CO2 may enter the reservoir on a high mobility charge plane. It accumulates at the top without percolating through the oil
column. This scenario would be in line with Stainforth’s charge model18. After its accumulation at the top, CO2 diffuses into
the oil column. As a result, a large continuous and monotonic gradient is formed.
IPTC 14398 9

Figure 13: Results of the simulation model with two faults of 1%
transmissivity. The gradient at 750,000 years matches the data
measured in the field.
Figure 14: Results of the simulation model with two faults of
0.01% transmissivity. The data suggests that a gradient
after 16 million years matches the data measured in the
field. Please notice the vertical profile between the faults. A
feature that is missed by having only one data point per
fault block.

The case of sealing faults seems contrived to match the data and raises a number of questions: How does the CO2 reach the gas
cap without getting trapped lower in the column? The reservoir would have needed to be charged with hydrocarbon fluid. If it
was faulted subsequently it would be likely that it would lose its charge in the process. Subsequently the CO2 would have
needed to find a way into the gas cap without getting trapped at the faults lower in the column. Compared to the connectivity-
scenario the case of sealing faults is less likely to occur. An increased number of DFA data points would help differentiate
between a continuous smooth CO2 gradient versus stair step discontinuous fluid gradient.

Asphaltene Analysis

While asphaltenes are present to some degree in almost all crudes, until recently very little was known about them besides
their contribution to the color of crude oil and the production problems for which they are responsible. The study of
asphaltenes took a leap forward recently with the introduction of the Yen-Mullins model which addresses the structure and
size of asphaltenes (Mullins et al, 2011). Asphaltenes are found nano-colloidally suspended in crude oil and while the
individual asphaltene molecules are not as large as previously thought the molecules tend to group together in nanoagreggates.
As asphaltene concentration increases, the nanoaggregates come together to form large cluster structures. Fluid density and
solubility act to drive asphaltenes from the top of the oil and gas column and they have been observed to grade continuously in
an oil and gas column (Zuo et al, 2010). Therefore, an asphaltene distribution will be observed with lower concentration of
asphaltene high in the oil column and increasing asphaltene concentration (and corresponding increases in observed
nanoaggregates and clusters) will be found lower in the oil column. The optical spectrometer used in DFA measures the
electronic absorption spectrum, or color, of the crude oil and is therefore perfectly suited to determine asphaltene content.

It follows that a reservoir fluid in equilibrium will have a continuous asphaltene gradient with the optical density (OD) or color
of the fluid increasing with depth. As part of the integrated approach to reservoir characterization for this paper, a plot of the
OD versus depth is provided in Figure 15 below.

Figure 15: Plot of Optical Density (Color) vs. Depth

10 IPTC 14398

The OD is observed to increase with increasing depth which would indicate that there are no sealing barriers in this area of the
reservoir. As DFA evolves it is expected that in the event that a break in asphaltene gradient is encountered during fluid
sampling operations, additional stations could be added to the sampling program at the top and bottom of the suspected sealing
barrier. Combined with evidence of pressure communication across the barrier, a continuous asphaltene gradient across a
barrier would imply connectivity. Further discussion of asphaltene optical properties and its application to determine vertical
connectivity across suspected sealing barriers will be provided in a future paper.

Conclusions

1. This paper presents a comprehensive study to address reservoir connectivity and compartmentalization in one of the gas
and oil fields located in the Southeast Asia. The integrated study of data acquisition, using EOS to prove pressure and
fluid communication between faults, the Isotope Analysis to understand the source of H.C. and non-H.C. gas, and the
mass transportation simulation to simulate the variation of CO2 within the same depth.

2. With current DFA technology, the fluid composition including CO2, GOR, and live density can be accurately quantified in
real time. As a result, DFA helps understanding of the reservoir complexity, thus enabling better decisions in
development strategy. In addition, quicker and more efficient way to characterize reservoir.

3. The EOS can be used to predict and compare with the actual DFA data to identify reservoir compartmentalization by
taking into account of the composition gradient effect.

4. The gases show similar hydrocarbon composition, which indicate wet thermogenic gas based on compositional and carbon
isotopic signature. Considerable differences between studied samples occur in the amount of non-hydrocarbon gases.
Varies parameters suggest that the CO2 originated from thermal breakdown of carbonate minerals in deep regimes and
was transported through faults to the reservoirs. Changing concentrations of CO2 among samples might be related to
mixing of CO2 from two different sources, or to fractionation processes during migration.

5. There is only one data point in each fault block available and therefore it is hard to make an assumption of the shape of the
gradient within the fault block. Knowledge of the shape of the gradient within the fault block would largely increase the
confidence of the interpretation. It would help to identify the gradient that applies and make it easier to conclude a firm
statement about compartmentalization. Additional DFA data points within each fault block would yield that knowledge.

6. The CO2 concentration exhibits the largest gradient in the data. Hence CO2 is the chemical analyte of interest when trying
to use gradients to constrain compartments. The GOR of black oil is less graded and thus not as suitable as an indicator to
constrain compartments in this case. It is suggested to investigate the asphaltene gradient in this field since the difference
in CO2 concentration might have impacted the stability of the asphaltene suspension.

7. Asphaltenes are typically found in all crude oil and grade continuously in an oil and gas column. Therefore, a continuous
asphaltene gradient shown as increasing optical density versus depth, along with pressure communication, will indicate
connectivity.

Acknowledgements

We would like to thanks PETRONAS Carigali Sdn. Bhd. and Schlumberger to allow us to present this work. We also would
like to thank the operation team that allows us to have such a great data to work on. In addition, we would like to express our
thanks to the DBR lab in Edmonton who conducted the geochemistry study for this research study.

References

1. Baric, G., Mesic, I., Jungwirth, M. and Spanic, D., 1991. Gas and gas condensate fields in the north-west of the Drava
Depression, Yugoslavia. In European Association Petroleum Geologists, ed. A. M. Spencer. Special Publication 1.
Oxford, University Press, Oxford, pp. 322-339.

2. Canas, J., Pop, J., Dubost, F., Elshahawi, H., 2008. Advanced Compositional Grading Analysis. SPE 115429,
presented at the SPE Annual Technical Conference and Exhibition, Denver, 21-24 September.

3. Clayton, C.J., 1991. Effect of maturity on carbon isotope ratios of oils and condensates. Organic Geochemistry 17,
887-899.
IPTC 14398 11

4. Cooper, B. A., Raven, M.J., Samuel, L., Hardjono, and Satoto, W., 1997. Origin and Geological Controls on
Subsurface CO2 Distribution with Examples from Western Indonesia. Paper presented at the International
Conference on Gas Habitats of SE Asia and Australasia.

5. Daungkaew, S., Harfoushian, J., Cheong, B. C., Sukahar, W., Tan, G. L., Sun, B., 2008. First Field Applications of
Downhole CO2 Measurement in Asia Pacific. SPE116501 presented at the SPE Asia Pacific Conference and
Exhibition, Perth, 20-22 October.

6. Fujiwara, M., Yamada, M., Sasaki, A., 1996. Hydrothermal Origin and E-log Identification of High CO2 Zones,
Platong Gas Field, Gulf of Thailand. Paper presented at the Second Well Logging Symposium of Japan, 26-27
September.

7. Imbus, S.W.; Katz, B.J.; Urwongse, T., 1998. Predicting CO2 occurrence on a regional scale: Southeast Asia
example. Organic Geochemistry 29, pp. 325 345.

8. Imbus, S. W., Wind, F. H., and Ephraim, D., 1999. Origin and Occurrence of CO2 in the Eastern and Andaman Sea,
Offshore Myanmar. Paper presented at an International Conference on Gas Habitats of SE Asia and Australasia.

9. Martini, A. M., Budai, J. M., Walter, L. M. and Schoell, M., 1997. Microbial generation of economic accumulations
of methane with shallow organic-rich shale. Nature 383, 155-158.

10. McKinney, D., Flannery, M., Elshahawi, H., Stankiewicz, A., Clarke, E., Breviere, J., Sharma, S., 2007. Advanced
Mud Gas Logging In Combination With Wireline Formation Testing and Geochemical Fingerprinting For An
Improved Understanding of Reservoir Architecture. SPE 109861, paper presented at the SPE Annual Technical
Conference and Exhibition. Anaheim, California, USA.

11. Müller, N., Elshahawi, H., Dong, C., Mullins, O.C., Flannery, M., Ardila, M., Weinheber, P., and McDade, E. C.,
2006. Quantification of carbon dioxide using downhole Wireline formation tester measurements. SPE100739, paper
presented at the 2006 SPE Annual Technical Conference and Exhibition held in San Antonio, Texas, U.S.A., 24–27
September.

12. Mullins, O.C., Elshahawi, H., Hashem, M.N., Fujisawa, G., 2005. Identification of Vertical Compartmentalization
and Compositional Variation by Downhole Fluid Analysis Towards a Continuous Downhole Fluid Log. SPWLA
685146, presented at the SPWLA 46th Annual Logging Symposium, New Orleans, 26-29 June.

13. Mullins, O.C., Andrews, A., Pomerantz, A., Dong, C., Zuo, J., Pfeiffer, T., Latifzai, S., Elshahawi, H., Barre, L.,
Larter, S., 2011. Impact of Asphaltene Nanoscience On Understanding Oilfield Reservoirs. Paper SPE-146649
presented at the SPE Annual Technical Conference and Exhibition held in Denver, Colorado, USA, 30 October–2
November.

14. Pfeiffer, T., Reza, Z., Schechter, D.S., McCain, W.D., Mullins, O.C. 2011. Determination of Fluid Composition
Equilibrium under Consideration of Asphaltenes – a Substantially Superior Way to Assess Reservoir Connectivity
than Formation Pressure Surveys. Paper SPE-145609 presented at the SPE Annual Technical Conference and
Exhibition held in Denver, Colorado, USA, 30 October–2 November.

15. Pisutha-Arnond, S., and Sirimongkolkitti, A., 2004. Carbon Isotropic Signature of CO2 in Arthit Gas Field, Northern
Malay Basin, the Gulf of Thailand. Petroleum Geochemistry and Exploration in the Afro-Asian Region, paper
presented for 6th AAAPG International Conference, Beijing, China, 12-14 October.

16. Prinzhofer, A., Mello, M.R., Takaki, T., 2000. Geochemical characterization of natural gas: a physical multivariable
approach and its applications in maturity and migration estimates. AAPG Bull. 84, 1152-1172.

17. Seewald, J. S., Seyfried, W. E, Jr. and Shanks, W. C. III, 1994. Variations in the chemical and stable isotope
composition of carbon and sulfur species during organic-rich sediment alteration: an experimental and theoretical
study of hydrothermal activity at Guaymas Basin, Gulf of California. Geochimica et Cosmochimica Acta 58, 5065-
5082.

18. Stainforth, J.G., 2004. New Insights into Reservoir Filling and Mixing Processes. In Understanding Petroleum
Reservoirs: Towards an Integrated Reservoir Engineering and Geochemical Approach, ed. J.M. Cubitt, W.A.
England, and S.R. Larter. SP 235: 115-132. London: Geological Society.
12 IPTC 14398

19. Suklis, J, Ames, A., and Michael, E., 2004. CO2 in South Sumatra-Observations and Prediction. Indonesian
Petroleum Association, paper presented at Deepwater and Frontier Exploration in Asia & Australasia Symposium,
December.

20. Xu, L., Cai, J., Guo, S, Yi, P., Xiao, D., Dai, Y. D., Yang, S. K., Khong, C. K., Fujisawa, G., Dong, C., and Mullins,
O. C., 2008. Real Time Carbon Dioxide Quantification Using Wireline Formation Tester to Optimize Completion and
Drill Stem Testing Decisions. Paper presented at the 14th Formation Evaluation Symposium of Japan, 29-30
September.

21. Zuo, J., Freed, D., Mullins, O.C., Zhang, D., Gisolf, A., 2010. Interpretation of DFA Color Gradients in Oil Columns
Using the Flory-Huggins Solubility Model. Paper SPE-130305 presented at the CPS/SPE International Oil and Gas
Conference and Exhibition held in Beijing, China, 8-10 June.