IPTC 12637

Reservoir Characterization and CO2-EOR Injection Studies in Chicontepec

Turbidite Reservoirs, Mexico
Maghsood Abbaszadeh: Innovative Petrotech Solutions, Inc., Kenji Ohno: Japan Oil, Gas and Metals National
Corporation, Hirofumi Okano: Japan Oil, Gas and Metals National Corporation, Jorge Morales, Petroleos
Mexicana, Juan Manuel Riano Caraza, Petroleos Mexicana

Copyright 2008, International Petroleum Technology Conference

This paper was prepared for presentation at the International Petroleum Technology Conference held in Kuala Lumpur, Malaysia, 3–5 December 2008.

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 restricted 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
Chicontepec fields are tight clastic deposits comprised of stacked turbiditic events with extreme geological heterogeneity and
variable heavy oil content. Huge reserves of oil are left behind because of loss of system energy by primary depletion;
offering challenging opportunities for IOR/EOR processes. We first present a general methodology for integrated
geostatistical reservoir characterization, using the technique of multivariate multiGaussian formalism for the integration of
data sources and scales. The resulting geostatistical models are calibrated to available production data, and used as predictive
tools for CO2-EOR and alternative solvent injection studies in simulated sector model pilots.
In addition to CO2, the paper discusses injection of associated gas (NGL), N2, CH4, flue gas and their mixtures with
CO2. Studies are performed to reduce MMP of Chicontepec oil with enriched CO2 and hydrocarbon mixtures, and a
procedure is devised for enhanced MMP predictions. Continuous water and CO2 injection studies determine limiting values
for injectivity and incremental recovery factors. WAG studies for mobility control and improved sweep determine optimum
WAG ratio to be 1-3. MMP studies illustrate variability in fluid PVT and the required injection pressure ranges for
miscibility.
The CO2 utilization factor is less than 10 Mscf/bbl, indicating economic viability. EOR recovery factors in excess of
20% are possible for these tight rocks; depending on oil characteristics, MMP of solvents and geology of specific pilot sites.
This high recovery efficiency is remarkable as contrasted with the 2-5% primary recovery factor established by 20-30 years
of production history. Key parameters are identified that impact oil recovery in Chicontepec reservoirs, and it is
demonstrated that flue gas (and hence N2) and associated gas injections mainly act as pressure maintenance agents because
they are immiscible displacements.

Introduction
Chicontepec basin, 25 km in width (E-W) and 123 km in length (N-S), has been formed by a complex system of submarine
fan and turbidite sediment deposits in a deep-water canyon of the paleo-Gulf of Mexico. This field was discovered in 1931,
and commercial production commenced in 1952. The Chicontepec reservoirs consist of Upper Paleocene-Lower Eocene
alternating stack of sandstone and shale bodies. These bodies do not represent a continuous laminar extension throughout the
entire field, and wide variations in clay-shale content and secondary porosity alterations by diagenesis are recognized.
Extreme heterogeneity of rock fabric and petrophtycial properties as well as fluid variations occur both laterally and
vertically across the fields. The API gravity of oil decreases progressively from north to south, ranging from about 45 oAPI
in north to near 18 oAPI in the south fields.
Huge reserves of oil are left in the fields because of quick loss of system energy by primary depletion in this tight sands,
making them ideal candidates for alternative IOR/EOR innovative processes. Carbon dioxide, CO2, or combinations of it
with other fluids as miscible solvents is emphasized. This is justified due to the very large volumes of original oil-in-place,
the very low primary recovery, relatively-impoved recovery by waterflood and the desire to achieve high oil recovery
efficiencies over the life of the field.
In the past several years, research has been conducted to develop practical technology for integrated multidisciplinary
reservoir characterization, reservoir management and field developments by fluid injection1-5. Data sources are first
integrated into static reservoir characterization models through the technology of geostatistics. Simulation models are then
2 IPTC 12637

constructed by calibrating static reservoir characterizations with various dynamic data for reliable assessment of alternative
field development scenarios through well placement and EOR. South Chicontepec fields are selected for benchmarking the
developed technology as a prototype site to be expanded later to other fields in the Chicontepec basin. Fig. 1 shows the
location of the Chicontepec field and the selected study area for this paper.

Project Study Area

(8km x 20km)
Thickness 400-500 m
> 250 wells

3D seismic
Agua Fria

Coapechaca Satellie Image taken from NASA

Tajin
Chicontepec Fields
(25km x 123 km)

Fig. 1: Chicontepec field location of Gulf of Mexico and the selected study area model

This paper first presents a brief description of integrated pixel-based geostatistical methodology for reservoir property
distributions of Vsh, facies, effective porosity and water saturation. Next, the paper presents an evaluation of various EOR
solvent injection scenarios in three simulated sector pilot models in south Chicontepec fields for field development purposes.
MMP of reservoir fluid with mixtures of CO2, N2 and hydrocarbons of varying composition are evaluated through slim-tube
simulations with tuned PR-EOS. Emphasis is placed in identifying CO2 solvent mixtures with reduced MMP to attain
miscibility at practical injections pressures in Chicontepec fields. Comparisons in recovery performance of waterflood, CO2,
CO2-WAG, flue gas, associated produced gas and CO2+HC mixture solvents are provided. These studies illustrate the
potential of CO2-EOR technology as a visionary optimized oil recovery plan, and open possibilities for future enriched CO2
injection for improved miscibility.

Reservoir Characterization
Geology
The Paleogene Chicontepec Basin is located between the Sierra Mandre Oriental and the Golden Lane Platform,
extending northwest-southeast along the Gulf Coast in Mexico. The main reservoir rock of the basin is submarine fan
turbidite sandstone derived from the Sierra Madre Oriental to the west of the basin. Previous sedimentological studies6
demonstrate that there were multiple sediment supply systems forming submarine fans along the western margin of the
Chicontepec Basin, and generating complex turbidite deposits comprised of multiple stacks of turbidite sequence events.
Extreme heterogeneity of rock fabric and petrophysical and fluid properties occurs in 3D across the fields. In addition to this,
irregular diagenesis prevails in the entire field, making reservoir characterization difficult and oil production non-uniform
across the fields. The Agua Fria (AF), Coapechaca (COA) and Tajin (TAJ) fields, as shown in Fig. 1, are situated in the
southern part of the Chicontepec basin within the area of interest in this study. According to the previous geological studies,
these fields are located in the depositional area of one of the submarine fans along the basin.
Geological and sedimentological analyses were conducted based on the concept of sequence stratigraphy for stratigraphic
division, facies modeling and digenesis pattern identification3-4. The south reservoirs are divided into 13 sequences and 26
subsequences, as seen in Fig. 2. Sequences AF100 and AF85 are absent in the Tajin field due to erosion, and Sequences
TAJ100, TAJ85 and TAJ60 are absent in the Agua Fria field, likely because of the by-passing of sediments in these fields
during pertinent geological deposition periods. Generally, three subsequences within each stratigraphic sequence deposit,
represented by stratigraphic transgressive surface TS and microcondenesed surface MCS, have been identified. Based on the
mapped sequence surfaces, the reservoir framework was constructed using seismic horizon and well marker information in a
coherent way (see Fig. 2 to the right). This framework was used in building property distributions for the reservoir units by
the geostatistical procedures.
Log patterns of gamma ray, resistivity and core facies recognized four major turbidite facies: sands (SA) as high density
turbidites, normal alternating sands (NA) of equal sand and shale beds, mud alternating sands (MA) of low density turbidites
and mud (M). Lower fan of sandy radial fan or levee of a channel levee system is considered as a sedimentary environment
for facies MA, and facies M is mainly shale deposited in a slope or basin floor environment. Secondary porosity is either
enhanced porosity by dissolution/leaching or reduced porosity by cementation. Primary porosity is assumed to be linked
directly to facies: facies SA has the highest porosity, whereas facies M has the lowest porosity. Diagenesis intensity is
measured as the degree of departure between secondary and primary porosities.
IPTC 12637 3

AF Field TAJ Field AF10

AF100

TAJ100
AF10
AF30

AF85
TAJ60

AF100
TAJ100

Tajin Canal

Fig. 2: Sequence stratigraphy of the reservoir intervals of south Chicontepec fields

Geophysics
Major horizons corresponding to sequence boundaries based on geological/sedimentological analysis were picked from
seismic data in time-domain and converted to depth through an iterative approach that used the velocity cube derived from
seismic migration velocity, VSP/check shot, sonic logs and geologic well marker picks5. Faults/fractures were identified
from seismic, coherency cubes and micro-seismic data at selected wells. All faults are normal faults with approximately
vertical dip and orientations from northeast to southwest.
Nine seismic attributes at the sequence interval scale (60-100 m) and acoustic inversion (AI) at the subsequence scale (20-
30 m) were extracted for integration with well petrophysics. The AI cube had finer resolution than the seismic attributes.
Thus a total of ten seismic attributes at different scales were used for integrated reservoir models2-4.

Geostatistical Modeling
Pixel-based geostatistical reservoir characterization was used to generate multiple equally probable static reservoir
models that honored hard data at wells and all available soft data from geology and seismic. The hard and soft data are at
different scales with different accuracy or resolution. The approach of mutlivariate multiGaussian (MSMG) integrated
sources of data from different scales and generated fine-scale property distributions within reservoir sequence deposits. This
method combines all secondary data of seismic attributes and geological diagenesis intensity maps into a “single” secondary
data – here called super secondary data through MSMG1-2, 7-9. The basic equation for super secondary data, yss, in normal-
scored Gaussian space is given by
n
y ss ( u j ) = ∑ λi yi ( u j ) (1)
i =1
[ ρ ij ] λi = ρ i0 , i = 1,..,n and j = 1,..,n (2)

ρ is correlation coefficient among all data sources at different scales. yi(uj) is normal score of variable i at location uj. yss
is used as a single soft data in conditional simulations of well data by the geostatistical collocated cokriging algorithm
n
y l ( ui ) = ∑ λα yα + µ y SS ( ui ) (3)
α =1
λα is for hard data and µ is for combined soft data. In order to impose the vertical trend on property distributions, the
vertical trend is binned into classes and normal score conditional petrophysical property is constructed for that class through
stepwise modeling10. All geostatistical realizations are conditioned to the binned values at their predominant depth classes. A
transformation table is constructed between a property and its normal score for each class, which is used for back
transformation to preserve the trend.
The geostatistical study model is represented by 125 x 400 grid cells laterally with dimensions of 50m x 50m. The
definition in the vertical direction is about 1 m for all sequences. The total number of grid cells in the vertical direction for
various sequences ranges from 20 to 160, most of them having 70 to 80 grid cells. The gridding in the areal direction is
4 IPTC 12637

conformed to the main direction of faults and fractures identified from seismic data analysis and information derived from
micro-seismic analysis at wells.
First, Vsh was modeled at the stratigraphic layer-scale in the order of 1 m. The vertical trend of Vsh from well data was
ss
calculated by averaging over 2-4 m layers. Second, all secondary data were integrated for super secondary data yVsh ( x , y ) at
the subsequence scale. Third, Sequential Gaussian Simulations (SGSim) collocated with the super secondary variable was
used to obtain realizations of yVsh distributions conditioned to transformed well data. The results were stepwise back
transformed for actual Vsh realizations at high resolution.
Second, multiple realizations of 3-D facies distributions were generated by first constructing “prior” facies models from
Vsh distributions and facies-Vsh calibrations. Sequential Indicator Simulations (SISim) was performed with prior facies
probability pk as locally varying mean (LVM) to generate facies realizations. Third, effective porosity was generated at the
layer scale for each facies given Vsh and other secondary data. SGSim collocated with the super secondary porosity data was
used to generate multiple realizations of effective porosity. Fourth, super secondary data for water saturation was built at the
layer scale by integrating Vsh and porosity realizations. SGSim collocated with the resulting super secondary data for Swi
generated water saturation distributions. Finally, permeability distributions were constructed from porosity and facies
distributions along with the core-based k-φ relationships for each facies.

Fluid Properties and EOS

Laboratory fluid analysis and PVT data from the reservoirs in this study show wide variations in bubble point pressure, pb,
saturation solution gas-oil-ratio, Rs, and composition as shown in Fig. 3. Such variations are attributed to heterogeneous
source rock and compartmentalization of reservoir. Fluid simulation studies with full compositional models indicate that the
effects of temperature and gravity on saturation properties of Rs and pb are negligible and that variations in oil composition is
the main source for the observed fluid behavior in the study area. Compositional analyses of reservoir fluid in Fig. 3 indicate
that the oil has large percentage of heavy components, nearly 10% of C30+ fractions. Also, oil compositional variations
maybe characterized by mole fraction of methane in oil, which is a hallmark indicator for solution gas-oil-ratio, Rs. Viscosity
measurements show oil viscosity of 5 cp at reservoir temperature of 85 0C (190 0F) and over 50 cp at standard condition.

40 5

AF-561
AF-584 Study Area
35 4

30 3

mol %
2
Agua Fria
25
TAJ-352
1
20
mol %

TAJ-367

0
TAJ-446
15

+
2

C 0
C 1
C12
C13
C 4
C 5
C16
C17
C 8
C29
C20
C 1
C 2
C23
C24
C 5
C26
C27
C 8
C 29
2
C1
C2
3
C4
C5
6
7
C8

C19

30
CO

1
1

1
1

2
2

2
N

C
C

C
COA-535, 571 TAJ-679
10
TAJ-318 Tajin
5

+
2

C0
C1
C2
C3
C4
C5
C6
C7
C8
C9
C0
21

C2
C3
C4
C5
26

C7
C8
C 29
2
1
2
3
4
5
6
7
8
C9

30
O

1
1
1
1
1
1
1
1
1
1
2

2
2
2
2

2
2
N
C
C
C
C
C
C
C
C
C

C
C

Fig. 3 - PVT variations, location of fluid samples and a compositional analysis of oil in south Chicontepec fields.

Equation of State (EOS) and Minimum Miscibility Pressure (MMP)

Four reservoir fluid samples from various wells in TAJ, COA and AF fields with full compositional analysis were used to
assess MMP with CO2 and with CO2-N2 mixture at reservoir temperature of 190 °F. The solution gas-oil-ratio at the
saturation pressure of 150.4 bara (2,180 psia) is 72.9 m3/m3 (410 scf/bbl). The fluid samples were characterized by splitting
the C8+ fractions into 3 pseudo-components and by lumping C2-C4 and C5-C7; while keeping CO2, N2, and methane as
individual components. The cubic PR-EOS model was tuned by adjusting the critical temperature, critical pressure, acentric
factor, volume shift and binary interaction parameters of non-hydrocarbons to match all PVT data, swelling tests of oil by
CO2 and N2 and slim-tube experiments with CO2 and N2 at different injection pressures. All HC-HC binary interaction
parameters were set to zero.
The characterized fluids were used to estimate the MMP of CO2 and CO2 mixtures with reservoir oil using slim-tube
simulations at reservoir temperature. For each pressure, the oil recovery at 1.2 hydrocarbon pore volumes (HCPVs) of CO2
injected was plotted and used to estimate the MMP. The MMP for the main sample (AF-561) is about 4,100 psi
(288Kg/cm2). The slim-tube simulations of CO2 injection with other fluid samples show that MMP varies from 2,800-5,200
psi (197-366 Kg/cm2) in the study area. The higher MMPs in south Chicontepec are associated with high percentage of
heavy fractions in oil samples.

Enriched CO2 MMP Enhancement

Impurities in CO2 can change the MMP of the injectant fluid with reservoir crude oil. Methane and particularly nitrogen
increase the CO2 MMP, whereas H2S decreases it11-14. We consider the effect of CO2 impurity on MMP with the AF-561
crude oil sample. Flue gas, a mixture of approximately 86% N2 and 14% CO2, is included amongst the CO2 impurity cases.
IPTC 12637 5

Mixtures of hydrocarbons and CO2 are also studied. High resolution slim-tube simulations were performed for all of the
injected solvents to quantify the MMP under various solvent injection scenarios. Fig. 4 shows the slim-tube simulation
results, indicating that MMP increases rapidly with increasing N2 mole fraction in the injected fluid, but MMP decreases in
the presence of hydrocarbons heavier than C2 in the injection stream. Considering MMP at RF of 90%, the MMP of flue gas
is over 9,000 psi (634 Kg/cm2), and the recovery factor at 4,000 psi (281 Kg/cm2) for the flue gas injection is only 40%.
Also shown in the figure on the right-hand side are the relation between mole percent of each impurity component in the
CO2 stream and their corresponding reduced MMP value from the initial MMP of 4100 psi for 100% CO2. Furthermore,
addition of the heavier hydrocarbons to CO2 has more profound effect on the miscibility pressure compared to lighter
components, such as C2. The relations of Fig. 4 also provide means to predict MMP for mixtures of HC components and
CO2, such as CO2+LPG (liquid petroleum gas). Addition of NGL (natural gas) or produced associated gas to CO2 may also
reduce MMP, although NGL or produced gas contains large quantities of methane that tends to increase MMP.

100.00
4400
100% CO2 slim-tube simulations
90.00 4200 C2

N2 4000
H2S
80.00 % as
eg
C3
15
2, Flu Air 3800 iC4
70.00 CO nC4
% 3600
85 nC5
60.00
N2 3400 iC5
RF (%OOIP)

%
100 nC6

MMP
50.00 3200 C2

MMP, psi
3000
40.00
2800
30.00
2600
AF561 & CO2
AF561 & 0.15N2-0.75CO2 H2S
20.00 2400
AF561 & 0.8N2-0.2CO2 C6
AF561 & 0.86N2-0.14CO2 2200
10.00 iC4 C3
4100 AF561 & N2
2000
0.00 0 10 20 30 40 50 60 70 80 90 100
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 Mole Percent
M ole pe rce nt of Impurities
of im puritie s
Pressure, psia

Fig. 4: MMP for CO2 and CO2 mixtures. MMP can be reduced by CO2+hydrocarbon mixture solvents

CO2-EOR Simulation Studies

Three inverted 7-spot pilot models are considered in this study, two in the lower sand sequences of Tajin and Agua Fria and
one in the upper sand sequences of Agua Fria. These pilots are selected in different parts of the model study in order to
assess the impact of reservoir variability on various CO2-EOR and other solvent injection processes. Some general
information on the characteristics of these pilot sector models is given in Table 1.

Table 1: General characteristics of pilot models

Pilot location Average air Total model Average pilot Hydrocarbon Water
perm in sector thickness, m pore volume, pore volume, saturation
model area, md MM Rm3 MM Rm3
Lower Tajin 10 100-150 6.39 4.48 0.30
Lower Agua Fria 13 220-260 12.00 7.92 0.34
Upper Agua Fria 16 110-180 4.78 3.96 0.17

Tajin Pilot, Lower Sands

Simulation Model Construction. A 34x30x140 fine-scale sector model volume in Sequence TAJ100 was extracted from
one of the geostatistical reservoir characterizations (Fig. 5). The grid dimensions are: ∆x=∆y=50 m (164 ft) and ∆z=1-2 m
(3.3-5.6 ft). There are 15 wells in the sector model, 7 of which are arranged in a nearly inverted 7-spot pattern with the
central injector Well TAJ-317. Permeability distribution is according to the facies-based k-φ correlations for the lower
sequences. Three-phase relative permeability and capillary pressure curves as functions of permeability are from correlations
of SCAL data with Corey-type equations. All wells are hydraulically stimulated in this low permeability formation.
The fine-scale geocelluar model was upscaled in the vertical direction, where the 140 fine layers were upscaled to 16
coarse layers in the simulation model. The critical gas saturation was set to zero to account for larger size simulation
gridblocks compared to core-scale SCAL measurements. The central injector well, has a finite conductivity fracture with
half-length of 175 m, represented by a high-resolution amalgamated local grid refinement (LGR) scheme and validated
against the analytical solutions for flux distribution along the fracture length15. The producers were modeled with negative
skin factors to represent their stimulation status.
The history matching of the primary production data was performed with both black oil and compositional simulations in
order to calibrate the static reservoir characterization to dynamic data. Some adjustments to gas relative permeability,
6 IPTC 12637

solution gas-oil-ratio at bubble point pressure and the mole fraction of methane in oil were necessary to reach a history match
to gas production data in Fig. 5. These adjustments tended to increase the simulated gas flow.

i
j
Gas production rate

Fig. 5: Tajin sector model with the injection pilot configuration and history match to gas production rate

EOR Performance Predictions: In all prediction studies, the wells were considered to operate at constant bottomhole
pressure. Constant pressure injections include continuous solvent and solvent-WAG injections with different WAG ratios.
The solvent injections include CO2, CO2 and hydrocarbon mixture, CO2 and LNG mixture, LNG or associated gas, flue gas
and methane. Water injection is also simulated to provide a basis for comparison. The injection pressures considered are
3000, 3500, 4000, 4500 and 5000 psi to cover injection cases below MMP and near or above MMP, where the CO2 MMP is
about 4,000 psi for the AF-561 fluid sample. The WAG cases use slug size of 2% HCPV within the hexagon pilot area,
because generally the CO2 slug size is not an important factor in CO2-EOR recovery performance predictions16. The WAG
ratios considered are 0.5, 1 and 3.
Fig. 6 shows cumulative oil recovery and oil production rate for continuous CO2 and water injection at various injection
pressures. At low injection pressures, the recovery mechanism is vastly different from the process mechanism at higher
injection pressure. The recovery factors for continuous CO2 injection range from 7.7% to 19.5% when injection pressure
varies from 3,000 to 5,000 psi. The recovery factors for water injection range from 7.7% to 11.3% for the same injection
pressure values.

Solid = CO2 5000 psi Solid = CO2

Dash = water 5000 psi Dash = water

4500
4500
4000
Cumulative Oil Production

3500
Oil Production Rate

4000

3500

Pinj=3000 psi

Pinj=3000 psi

Injection Period Injection Period

Fig. 6: Cumulative oil production from pilot TAJ-317 for continuous CO2 and water injection.

One important feature of CO2 injection, especially at higher injection pressure, is formation of oil bank13,14, where
residual oil saturation Sor is lowered towards zero upon contact with CO2 at reservoir pressures higher than the MMP.
Waterflooding does not produce an oil bank. At low injection pressure of 3,000 psi, waterflood and CO2 injetions show
similar performance, indicating a much poorer CO2-miscibility process mechanism, with the CO2 injection approaching
immiscible gas injection case.
Fig. 7 shows cumulative oil production for CO2-WAG injections at the injection pressure of 4,000 psi. Also included are
flue gas WAG injections and waterflood cases. The hybrid injection refers to cases where the reservoir is initially
waterflooded and then CO2 or CO2-WAG is implemented. This is either to pressurize the system by a cheaper fluid before
CO2 or to subject the reservoir to a secondary recovery process before implementing EOR injection schemes. Fig. 7 shows
the results of water injection for 5 years followed by CO2 and CO2-WAG injections. Because a GOR limit is imposed in
IPTC 12637 7

most field operations, the recovery of various processes should be calculated at a certain GOR limit, such as 6000 sm3/sm3
(30,000 scf/bbl).
The circles in Fig. 7 show cumulative oil produced from the combined pilot wells at the economic limit of GOR=6000
sm3/sm3. It is obvious that continuous CO2 injection cases cannot be tolerated in practice. Also injections with small WAG
ratio may not be attractive either. The case of WAG-ratio of 3 seems to be optimum for this particular pilot study. WAG
controls mobility and prevents rapid breakthrough of gas either by channeling or by gravity over-ride in many field cases17,
as is also the case in this study. Hybrid runs require lower CO2 consumption than WAG injections without the initial
waterflood phase; however, this may not be that significant. Therefore, if operationally and economically feasible, it would
be beneficial to start WAG injections from the onset before implementing a waterflood program.
Although the results indicate that continuous CO2 injection may yield higher or comparable oil recovery to WAG
injections, it also yields higher GOR that makes the process impractical and uneconomical under most conditions. WAG
injections result in a similar recovery factor but at considerably lower produced GOR and improved volumetric sweep. In
general, CO2-WAG is superior to both continuous CO2 and water injections. Flue gas WAG injections are inferior to
waterfloods because flue gas is practically an immiscible gas injection process.

P_inj=4000 psi WAGR

d O2)
G 0.5 bri
-WA Hy ous C
CO2 １ nu
３
P_inj=4000 psi
ont i
(C
0.5 1 3
Cumulative Oil Production

AG

Cumulative Oil Production

,W
brid
Hy
flood
Water
2
s CO
ou
3 nu
nti
１ Co 2O
us H
tinuo
0.5 Con

Flu e Gas-WAG

Injection Period Injection Period

Fig. 7: CO2-WAG, flue-gas-WAG (left) and hybrid WAG injections (right), TAJ-317 pilot.

Flow simulations with various CO2+HC mixtures show that MMP has the first degree effect on oil recovery of miscible
displacement gas injection process and that the type of CO2 impurity has relatively unimportant second order effect.
Therefore, the recovery factor corresponding to any miscibility enhancement scenario can be quantified if only one
simulation at a given MMP is conducted irrespective of the composition of the injected impure CO2 stream. It is found that
the recovery factor is doubled from 9% to nearly 20% when MMP is reduced from the original value of 4100 psi to 2600 psi
and the injection pressure is set at 3500 psi.

injection BHP=3,500 psi

(below parting pressure)

CO2 Mixture
Recovery Factor, %OOIP

(MMP=2600)
CO2 utilization Factor
vs. MMP
(Mcf CO2 / bbl Oil)

100% CO2
(MMP=4100)

Injection Time, days

Fig. 8: Impact of MMP and CO2 impurity on recovery factor and CO2 utilization factor, TAJ-317 pilot.

The study shows that the flood performance, or gas utilization factor (Mcf of solvent injected per barrel of oil produced),
at injection pressure of 4000 psi is minimum for MMP of 2,900 psi, indicating an optimum MMP with corresponding
miscible injectant mixtures for increased oil recovery factor of 23%. Gas utilization factor is a direct measure of project
economics in connection with field results, such as total oil production and total amount of solvent injected. This measure
shows that as MMP increases flood performance is deteriorated because of higher amount of solvent that needs to be injected
per barrel of oil to be produced.
In overall, however, the simulations clearly demonstrate that the increase in oil recovery by reduced MMP is significant,
making the idea of enhanced miscibility by enriched CO2 mixtures attractive in Chicontepec and worthwhile for further
8 IPTC 12637

investigation. Proper design of CO2 flood will require considerations of economics to determine optimum values for the
three interrelated parameters of MMP, gas utilization factor and oil recovery factor.
Fig. 9 shows simulated oil recovery for various solvent injections. The fluids injected are: CO2+HC mixture with
MMP=2600 psi, CO2 with MMP of 4100, NGL or associated gas of composition (N2=7.6, CO2=0.3, C1=74.1, C2-C4=16.4,
C5-C6=1.3%), mixture of 20% NGL and 80% CO2, methane, flue gas and water. All injections are at injection pressure of
4000 psi to assess the impact of solvent type and process mechanism on oil recovery. The results clearly show that the
recovery factors by methane and flue gas are less than waterflood because these fluids are by in large immiscible
displacement processes. NGL due to its content of C2+ components benefits somewhat from partial miscibility and oil
vaporization mechanism to yield higher recovery factor compared to waterflood. CO2 and CO2+HC mixture yield
considerably larger oil recovery. In fact, a recovery factor of 25% after 10 years of CO2+HC mixture is substantial, and
offers an attractive EOR process for developing the Chicontepec oil fields. For field implementation purposes, however, the
established technology of WAG injection with enhanced MMP of CO2+HC mixture solvent should be considered for
improved volumetric sweep and economical solvent utilization factors.

0.3

Cumulative Oil Production from Pilot, sM3

0.3 CO2 Injection pressure = 4000 psi
0.25 02NGL08CO2 CO2 Mixture Injection pressure = 4000 psi
(MMP=2600) L
0.2 CH4
NG
10 years of simulation G 1:
RF, %OIP

0.25 Flu
WA
0.15

2 L
NGL
CO NG
G3:
0.1

: WA
0.05 WAT 2
C O 1 00 ) G1
0 C22600-4000 P =4 WA 2
%OOIP

(M M O
L
C

as

0.2
4
L

2O
2

G
H

:C
H
O

G
N

H4
2+

C
C

H
2+

e
O

Flu

G3
O

3 :C
C

H2O
C

WA G1,
%OOIP

O2 WA
%C
(pilot),

80
0.15 L+ L
RF-Pilot,

NG NG L
20 % NG
Recovery Factor

r
W a te

CH4
0.1

CH4

0.05

Flue Gas

0
0 1000 2000 3000 10 yr 4000 5000 6000 7000 8000
Time, Days
Time, days
Injection Time, Date

Fig. 9: Performance prediction of varuous fluid injection processes, TAJ-317 pilot.

Agua Fria Pilots, Lower and Upper Sands

Two inverted 7-spot simulation pilots in the Agua Fria field, one in the upper sands and another in the lower sands are
considered. The lower sand pilot has large pore volume compared to the upper sands (see Table 1). The injectors are
hydraulically fractured and are modeld with the scheme of LGR. The formations in these two pilots are different from the
Tajin pilot study area presented before.
Continuous CO2, water and flue gas as well as their WAG injections were studied in these two pilots. The WAG gas
slug size is 2% of HCPV within the hexagon pilot volume and the WAG ratio is 1. The producers are at constant BHP of 125
bars (1,813 psi), and the injectors are at constant bottomhole pressure but at different levels. The simulation models were
first calibrated against primary production data from these pilot wells using compositional simulations with the same EOS as
the one used in the Tajin pilot. Slight adjustment to methane mole fraction in oil was necessary to match gas production data.

Dash: Waterflood Pinj=4500 psi Dash: Waterflood CO2, 4500

Solid: CO2 and Flue gas Solid: CO2 and Flue gas
4000 4000 Flue gas, 4500 4500

CO2, 4000
3500
Co Flue gas, 4000 4000
Cumulative Oil Production

nt
in 3500
uo
us 3000
CO CO2, 3500
2
Cumulative Oil

4500 Flue gas, 3500 3500

4000
W 3000
at 3500 CO2, 3000
er
flo
od 3000
3000 Flue gas, 3000
4500
Flue
Gas
3000

Injection Period
Fig. 10: Cumulative oil production and average reservoir pressure in Agua Fria lower sands pilot for continuous CO2, water and flue
gas injections.

Fig. 10 shows cumulative oil production for continuous fluid injections. Flue gas injections are very inefficient and are
inferior to waterfloods. CO2 injection appears advantageous over waterflood, where CO2 can recover 1.5 to 2 times more oil
IPTC 12637 9

than waterflood and that flue gas injections will result in oil recovery less than half of waterflood. It is interesting to compare
average reservoir pressure of the pilot under these three fluid injection scenarios. As the figure shows, the average pressure
under the three fluid injections of CO2, flue gas and water are very similar. However, despite similar pressure maintenance
capability offered by these fluid injections, their recovery performances are vastly different. This finding illustrates that
increased pressure or supplied energy to the system is not the sole pararameter for increased recovery factor in Chicontepec.
It is the process mecahanism and miscibility characteristics of CO2-EOR with reservoir oil that yields substantial oil
recovery.
Fig. 11 compares the recovery factor vs. pore volumes of fluid injected into the three pilots considered under continuous
CO2 and WAG injection processes. The results indicate that CO2-WAG injection is preferred over continuous CO2 injection
because of improved volumetric sweep efficiency offered by the WAG processes. The poorer response of the pilot in the
Tajin field compared to those in the Agua Fria field is primarily attributed to lower permeability of the formation and higher
degree of reservoir heterogeneity in Tajin. These simulation studies clearly demonstrate the role of reservoir characteristics
on CO2-EOR processes.

0.3
ANTPLT-CO2-P4000
AF-Antares Pilot ANTPLT-WAG1-P4000
AG)
ANTPLT-WAG1-P4000 O2-W
wer (C
(Upper sands) 0.25 TAJPLT-CO2-P4000 AF_lo
AG)
TAJPLT-WAG1-P4000
2-W
AFPLT-CO2-P4000
pper (CO
u
AFPLT-WAG1-P4000 AF_
0.2 AG)

%OOIP Oil Recovered

O2-W
Tajin Canal TAJ (C
AF-846 Pilot
0.15
(Lower sands)
)
er (CO2
AF_low
0.1
TAJ-317 Pilot O2)
TAJ (C
Waterflood Pilot Area
(Lower sands) O2)
per (C
0.05 AF_up

0
0 0.5 1 1.5 2 2.5 3
PV CO2 Injected, frac.

Fig. 11: Performance of CO2 and CO2-WAG injection in three simulated pilots in South Chicontepec fields

Conclusions
1. The multidisciplinary approach of multivariate multiGaussian integrated sources of data with different scales and
accuracy into geostatistical reservoir characterizations of south Chicontepec fields. The resulting high-resolution
geostatistical realizations honor well data, 3-D trends and secondary geophysical and geological data.
2. The analysis of PVT samples and fluid characteristics indicate that there are large lateral and vertical variations in
reservoir oil characteristics in different parts of the system, and hence large uncertainty exists in determining fluid
properties within the sand bodies of reservoirs.
3. Representative and calibrated EOS model studies indicate that the MMP of CO2 with reservoir fluids is in the range
of 2800-5200 psi, depending on oil composition. Reduced MMP is achieved through mixtures of CO2 and
hydrocarbons of C2 and higher carbon numbers (enriched solvent).
4. Three pilot models in Agua Fria and Tajin field sands were constructed for various solvent injection simulation
studies. Better recovery factors are obtained with WAG injections that combine improved displacement efficiency
by gas and sweep conformance by mobility control of water for the highly heterogeneous Chicontepec formations.
5. Considerable improved recovery is obtained when the MMP of CO2 with reservoir oils is reduced by enriching the
CO2 injection stream with a mixture of light hydrocarbons. CO2 gas utilization factor is also lower with reduced
MMP solvent injectants, making the enriched CO2 even more attractive.
6. This study demonstrates that despite extreme heterogeneities that are dominant in Chicontepec, a sound engineering
approach for successful field development would require WAG injection of enriched CO2+HC mixture solvents for
lower MMP and enhanced miscibility at injection pressures below formation parting pressure. This is an attractive
strategy for implementing CO2-EOR injection plans in Chicontepec.

Acknowledgement
We thank the management of PEMEX and JOGMEC for permission to publish this paper. The work represents a portion of
the collaborative joint research project between PEMEX and JOGMEC for optimization of development scheme of
Chicontepec basin. Technical contributions of the JOGMEC and IPS team members are also acknowledged.
 10 IPTC 12637

Nomenclature
K = permeability
MMP = mimimum miscibility pressure
pb = bubble point pressure
Rs = solution gas oil ratio
Sw = water saturation
Sor = residual oil saturation
u = location being estimated
Vsh = shale volume
yss = super secondary data in normal score
yi = attribute value in normal score
WAG = water alternating gas
µ = weight factor for integration of yss
λ = weights in multivariate Gaussian model
ρ = correlation coefficient among variables
φ = effective porosity

References
1. Abbaszadeh, M., Ohno, K.: “Water, CO2-WAG and Flue Gas IOR/EOR Injection Pilot Studies in South Chicontepec Reservoirs,
Mexico”, paper presented at the annual meeting of CIPM, AIPM, AMGP, AMGE, and SPE in Veracruz, Mexico, June 28-31, 2007.
2. Takahashi, S., Abbaszadeh, M., Ohno, K., Salazar Sato, H. and Alcazar Cancino, O.: “Integrated Reservoir Modeling for Evaluating
Field Development Options in Agua Fria, Coapechaca and Tajin Fields of Chicontepec Basin”, SPE paper 103974 presented at the First
International Oil Conference and Exhibition in Mexico held in Cancun, Mexico, 31 August–2 September 2006.
3. Abbaszadeh, M., Takano, O., Yamamoto, H., Shimamoto, T., Yazawa, N., Sandria, M. F. and Guerrero Z. D., Garza, F. R.: “Integrated
Geostatistical Reservoir Characterization of Turbidite Sandstone Deposits in Chicontepec Basin, Gulf of Mexico,” SPE 84052
presented at SPE Annual Technical Conference and Exhibition, Denver, CO, 5-8 October 2003.
4. Abbaszadeh, M., Takano, O., Shimamoto, T., Yazawa, N., Sandria, M. F. and Guerrero Z. D.: “Geostatistical Modeling of Sandstone
Distributions in Chicontepec Turbidite Deposits-Tajin Field, Veracruz, Mexico,” presented at EXITEP 2001 meeting, Mexico City,
Mexico, 5-7 February 2001.
5. Yazawa, N., Yamamoto, H., Takano, O., Shimamoto, T., Abbaszadeh, M., Sandria, M. F. and Guerrero Z. D.: “Geostatistical Modeling
of Agua Fria, Coapechaca and Tajin Field – Chicontepec Basin”, presented at AIPM 2001 international meeting, Villahermosa,
Mexico, 14-17 June 2001.
6. Bitter, M. R. “Sedimentation and provenance of Chicontepec sandstones with implications for uplift of the Sierra Mandre Oriental and
Teziutlan Massif, east-central Mexico. In Mesozonic and Early Cenozoic Development of the Gulf of Mexico and Carribean Region,”
Gulf Coast Section SEPM, 13th Annual Research Conference, 157-172, (1993).
7. Deutsch, C.V.: Geostatistical Reservoir Modeling, Oxford University Press, New York, 2002
8. Johnson, R. and Wichern, D.: Applied Multivariate Statistical Analysis, 5th edition, Prentice Hall Inc., New Jersey, 2002.
9. Deutsch, C.V. and Zanon, S.D. :”Direct Prediction of reservoir Performance with Bayesian Updating,: Jour. Cand. Pet. Tech., Vol. 46,
No. 2, Feb. 2007,
10. Leuangthong, O., and Deutsch, C.V.: “Transformation of Residuals to Avoid Artifacts in Geostatistical Modelling with a Trend”,
Mathematical Geology, 36 (3), April 2004, pp. 287-305.
11. Ali Danesh: PVT and Phase Behaviour of Petroleum Reservoir Fluids, Developments in Petroleum Science series, Elsevier, 2001
12. Lake, L. W.: Enhanced Oil Recovery, Prentice Hall Englewood Cliffs, NJ, 1989.
13. Stalkup, F., "Miscible Flooding Fundamentals”, Soc. of Pet. Eng. Monograph Series, 1983.
14. Jarrell, P., Fox C., Stein M., “Practical aspects of CO2 flooding”, Soc. of Pet. Eng. Monograph Series, 2002.
15. Cinco-Ley, H., Samaniago, F. and Domínguez, A.: “Transient Pressure Behavior for a Well with Finite-Conductivity Vertical
Fracture,” Soc. Pet. Eng. Jour., pp. 253-264, August 1978
16. Christensen, J.R., Larsen, M., Nicolaisen: H., “Compositional Simulation of Water-Alternating-Gas Processes,” Paper SPE 62999,
presented at SPE Annual Technical Conference and Exhibition, Dallas, Texas, 1-4 October 2000.
17. Christensen, J.R., Stenby, E.H., Skauge, A.: “Review of WAG Field Experience,” SPE Reservoir Evaluation & Engineering, April
2001.