SPE-177445-MS

Production Performance of Gas Condensate Reservoirs: Compositional

Numerical Model - A Case Study of Hai Thach - Moc Tinh Fields
T.V. Tran, A. T. Ngo, H. M. Hoang, and N. H. Tran, Bien Dong POC

Copyright 2015, Society of Petroleum Engineers

This paper was prepared for presentation at the Abu Dhabi International Petroleum Exhibition and Conference held in Abu Dhabi, UAE, 9 –12 November 2015.

This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s). Contents
of the paper have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect
any position of the Society of Petroleum Engineers, its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written
consent of the Society of Petroleum Engineers 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 SPE copyright.

Abstract
The approach focuses on the use of analytical and 3D dynamic compositional numerical models to
physically describe the condensate banking phenomenon and to better understand the real deliverability
in terms of gas and condensate production. The model is also built to compare the benefits of hydraulic
fracturing over the horizontal well in Bien Dong POC development campaign.
In this study, an analysis of pressure transient test data is conducted to investigate the condensate bank
radius. Construction of a radial model is used as a benchmark model to compare it to an equivalent
Cartesian grid model in term of productivity. The paper is focused on the development of a workflow for
the analysis and selection of grid size and methods used in the 3D compositional model. Then the model
is used for horizontal and hydraulic fracturing designs.
When the flowing bottomhole pressure falls below dew point, the condensate dropouts in the near
wellbore reduce permeability to gas through relative permeability effects; hence the well productivity is
reduced significantly. The prediction of gas condensate well production will strongly depend on conden-
sate banking evaluation and modeling. It is important that a model is built to describe the condensate
banking phenomenon accurately to avoid over-estimating well performance. It is shown that an integrated
model which incorporates Generalized Pseudo Pressure (GPP) and Local Grid Refinement (LGRs) can
reproduce almost exactly with the benchmark model. The presented procedure and analysis can serve as
a guideline for full field gas condensate reservoir modeling. The study also suggests that un-fractured
horizontal well is not attractive compared to hydraulic fractured vertical well. Fractured half-length
optimization design should be based on economic criterion.
The study addresses the complexity of gas condensate behavior around the near wellbore region. The
workflow of building an accurate 3D compositional model allows Bien Dong POC to have a reliable
resource in reservoir development and management plans.
The applicability of this work is confirmed by actual field case study in offshore Vietnam. It contributes
to the knowledge of performance of gas condensate reservoirs and supports representative modeling of
detailed model and recovery processes.
2 SPE-177445-MS

Introduction
The prediction of gas condensate well production will strongly depend on oil banking evaluation and
modeling (El-Banbi, 2010). The liquid dropping out from the gas occurs when bottome-hole pressure falls
below the dewpoint pressure. In deep and tight gas formations, condensate banking can occur shortly after
production start-up as pressure drops below the dew-point pressure. Condensate blockage near the well
bore is a result of the decreased gas mobility around a producing well below dewpoint pressure.
Condensate dropout near the wellbore has two key negative results: (1) decrease in gas and condensate
production due to near-well blockage, (2) the produced gas contains fewer valuable heavy components.
Productivity reduction in production wells due to retrograde condensation has been greatly studied over
decates (Musket 1949). The availability of accurate full field dynamic model is essential for production
optimization and reservoir development. If the condensate banking phenomenon is not well understood
and accurately modeled, the company’s field development and operational decisions may not support the
maximum economic value of the reservoir.
To characterize gas condensate flow, a three-zone model (illustrated in Fig. 1) was first put forward by
Fevang and Whiston (1996). This three-zone model, as divided below, has been commonly adopted in the
oil industry to understand the near flow behavior in gas condensate wells (Fevang and Whitson, 1996;
Singh and Whitson, 2008).
● Region 1 – near wellbore region: both gas and condensate flow simultaneously at different
velocities. Condensate saturation is at its critical value. In this region, gas relative permeability
decreases significantly.
● Region 2 – quite near the well: condensate accumulation occurs in this region. Only gas flows in
this region because oil saturation is below its critical value. The radius of condensate accumulation
changes with time.
● Region 3 – Far away from well: Single phase exists. Gas preserves its original composition when
pressure is above dew-point pressure. When reservoir pressure falls below dewpoint pressure,
condensate forms throughout the reservoir, therefore region 3 no longer exists.

Figure 1—Condensate saturation profile around the wellbore (Fevang and Whitson, 1996)
SPE-177445-MS 3

The condensate banking may appreciably reduce gas well deliverability, though the severity depends
on reservoir and well parameters. Condensate blockage can be important if the pressure drop between the
reservoir and the wellbore is a significant percentage of the total pressure drop from reservoir to delivery
point (e.g. a surface separator). Its effect can reduce the gas relative permeability up to 95% (Al-Anazi
et al., 2005).
In terms of well performance, the near-well flow behavior, determined by the near-well relative
permeability functions, is the dominant factor, and Region 1 is the main source of deliverability loss in
a gas-condensate well. Gas relative permeability is reduced owing to condensate build-up. Region 2 has
somewhat reduced gas relative permeability, but this generally has a second-order effect on well
performance (Fevang and Whitson, 1996).
Methodology

In this study, the fluid properties and equation of state (EoS) are validated from experimental data. A
synthetic fine grid radial single well model is constructed with the compositional reservoir simulator to
describe the condensate blockage and its effect on well productivity. This is considered as a benchmark
model. Once the radial model is built, Cartesian model model (incorporating GPP and LGR) is calibrated
by adjusting grid sizes to match the results of the benchmark model. The calibrated Cartesian model is
then applied for full-field simulation. Fig. 2 illustrates the workflow of building an accurate condensate
banking model.

Figure 2—Workflow to build an accurate condensate banking model.

Fluid properties
In order to characterize the fluid properties, heavy components C7⫹ are reported as C7, C8, C9, C10, C11,
and C12⫹. The C12⫹ is splitted into two pseudo components FRC1 and FRC2. During the regression
4 SPE-177445-MS

process, OmgA, OmgB, and BIC of C1 and FRC1, FRC2 are found to be the most sensitive parameters.
Therefore, they are put in one regression in order to match the experimental data. The total number of
components is reduced from 18 to 6 (C1⫹, C3⫹, C7⫹, C9⫹, FRC1, FRC2) by grouping components
using Molar Weight Technique. This process produces better matching and reduces the simulation run
time. When the best match is found, the fluid properties are exported and put into the simulation for
history matching and production forecast. Fig. 3 shows the comparison between 18 and 6 component EoS
matching with experimental data.

Figure 3—18 components vs. 6 components: a) phase behavior; b) CCE Relative Volume; c) CVD 2 phase-z; d) CVD Liquid Drop-out.

Model Validation
Fine grid radial model was built using 30x1x1 grid cell (Fig. 4a). The length of the radial gridblock (N)
was distributed according to the following equation:
(1)

where re is the external radius (no flow boundary) usually taken as the mid interval between producing
wells, and rw is the wellbore radius. An equivalent Cartesian model (Fig. 4b) was built with different
grids: 43x43x1, 87x87x1, and 175x175x1. The same fluid and rock properties were used in all models.
The results showed that coarse Cartesian grid models over-predicted productions compared to the
benchmark radial model (Fig. 5).
SPE-177445-MS 5

Figure 4 —a) Single well radial model b) Single well Cartesian model.

Figure 5—Cartesian vs Radial model: coarse Cartesian model over-estimated production rate.

Local grid refinement (LGR) and generalized pseudo pressure (GPP) methods were applied to the
Cartesian model to describe the condensate banking phenomenon around the wellbore. The results showed
that LGR incorporated with GPP was a simple option providing reasonable results with Cartesian model
(Fig. 6). The selection of grid size and LGR algorithm was then used in the fullfield model for production
forecasting.
6 SPE-177445-MS

Figure 6 —Comparison Radial model vs Cart. Model with LGR & Cart. Model with LGR ⴙ GPP.

Field Applications
Well HT-1 is the first development well producing from MMF reservoir of Hai Thach field. It is one of the two
wells planned primarily to produce from this reservoir, which is interpreted as a gas - bearing turbidite sand
with total in-place reserve of 200 bcf of gas and 18 mmbbl of condensate. The production profile of this well
will be used as reference for optimizing the second well HT-2 drilling into the MMF reservoir.
The dynamic model of MMF is generated based on the static model framework (Fig. 7) which has
dimensions of 41⫻161⫻40. The total number of cells is 264,040, in which 206,334 cells are active. One
PVT set, one fluid region, and one saturation table are used in the MMF dynamic model.

Figure 7—Geological model of MMF.

SPE-177445-MS 7

The fullfield dynamic model is built based on the optimized grid block size and LGR obtained from
single well Cartisean model. The model is set at equilibrium with gas water contact at 12204.72 ft TVDss.
The initial reservoir pressure is 8461.5 at 11651.82 ft TVDss and the temperature is 305.6°F. The water
saturation of MMF is set at an average value of 0.25 for the whole reservoir. The average permeability
obtained from pressure transient analysis in HT-1 is estimated to be 0.57 mD. However, permeability
distribution throughout the reservoir is based on porosity-permeability relationship function obtained from
core analysis.
The production data of well HT-1 was used for history matching, which consisted of 3 main steps: (1)
initialization, (2) sensitivity analysis on parameter uncertainties, (3) optimization.
Sensitivity analysis on parameter uncertainties:
Historical production data is put in schedule to run the history matching process. The sensitivity of each
parameter is studied by varying only one parameter and keeping all other parameters the same as the base
case. It is found that the model is most sensitive to permeability and second most sensitive to skin, as
shown in Fig. 8. Number of uncertainties is run for permeability and skin to obtain the best match. Fig.
9 shows the uncertainty analysis and the best match result which is used as the ⬙matched model⬙ for future
forecast.

Figure 8 —Tornado Chart – Sensitivity for paremeters

Figure 9 —Sensitivity under uncertainties and optimization.

8 SPE-177445-MS

Simulations to evaluate the benefits of horizontal drilling and hydraulic fracturing

Number of studies were done to provide guidelines that demonstrate which reservoirs may benefit from
horizontal wells instead of fractured vertical wells (Seah et al, 2014). In this study, the fullfield simulation
is conducted to provide an extensive comparison of the performance of fractured vertical well and
unfractured horizontal well deployment for well HT-2. For simplicity, only well HT-2 is assumed to
produce the entire MMF reservoir.
A compositional numerical model is developed with the Peng-Ronbinson equation of state to model the
phase behavior of reservoir fluid. A sensitivity study is carried out on the grid size and LGR area to avoid
any numerical errors. After benchmarking with radial model, the full field reservoir simulations are
performed to compare the benefits of base cases obtained from hydraulic fractured vertical well,
horizontal well. Fig. 10 demonstrates the LGR techniques used to model the fracture around the
perforations.

Figure 10 —Refine the cell geometry around the fracture using LGR.

The most likely fracture half-length and horizontal length used in the studies are xf ⫽ 328 ft (100 m)
and L ⫽ 1312 ft (400 m), respectively (based on consultants’ experiences).
Three simulation cases are set up to forecast the ultimate recovery of HT-2 toward the end of the field’s
life.
● HT-2: most likely un-fractured vertical well (based on analog well HT-1)
● HT-2-HW: most likely horizontal well (based on consultant’s experience)
● HT-2-HF: optimal vertical hydraulic fractured well
Fig. 11 shows that the ultimate recovery is higher in fractured case than in horizontal case, this is also
somewhat in agreement with Brown and Economides (1992) studies for low permeability reservoir (less
than 5mD).
SPE-177445-MS 9

Figure 11—HT-2 gas and condensate production profiles

Table 1 shows the improvement in production compared to the most likely HT-2 case. The total
production recovered in verical fractured well is reported to be 54.80 BCF (117% increase) and 3.50 (25%
increase) mmbbl of gas and condensate, respectively.

Table 1—HT-2 Ultimate recovery comparison

HT-8P_actual HT-8P_HW_Base increase HT-8P_HF increase

Cummulative gas production [bcf] 25.26 34.06 35% 54.80 117%

Cummulative condensate production [mmbbl] 2.79 2.96 6% 3.50 25%

Detailed study on the hydraulic fracturing candidate ranking and the influence of control parameters
is presented in Appendix 1. The full numerical modeling is designed to simulate the production
performance of a horizontal well drilled in MMF30. The local grid refinement (LGR) method is used to
describe the effect of condensate blockage near well bore. Three different cases are investigated based on
the horizontal length (Lmin ⫽ 656 ft, Lbase ⫽ 1312 ft, and Lmax ⫽ 4810 ft). The results show that longer
horizontal lengh does not improve the productivity significantly, in fact, it will create higher risk of
condensate banking damge around the well bore (due to high drawdown and longer perforation interval).
Table 1 shows the total production forecasted for three different development scenerios. This suggests that
hydraulic fracturing is the most benifical deployment for the HT-2.
Conclusions
The condensate banking phenomenon is investigated and modeled accurately by incorporating GPP and
LGR techniques. The study suggests that:
1. Horizontal wells are likely to be damaged, severely in certain cases. Longer length is required to
compensate the productivity loss.
2. Un-fractured horizontal well is not attractive compared to hydraulic fractured vertical well.
3. Fracture half-length optimization design should be based on economic criterion, e.g. NPV.

Acknowledgement
The authors would like to thank PetroVietnam (PVN) and Bien Dong POC (BDPOC) for permission to
publish this paper.
10 SPE-177445-MS

References
Al-Anazi, H.A., Solares, J.R., and Al-Faifi, M.G. 2005. The Impact of Condensate Blockage and
Completion Fluids on Gas Productivity in Gas-Condensate Reservoirs. Paper SPE 93210 presented
at the Asia Pacific Oil & Gas Conference and Exhibition, Jakarta, Indonesia, 5-7 April.
Brown, J.E. and Economidies, M.J. 1992. An Analysis of Hydraulically Fractured Horizontal Wells.
Paper SPE 24322 presented at the Rocky Mountain Regional Meeting, Wyoming, USA, 18-21
May.
El-Banbi, A.H. 2010. Optimizing Simulation Studies for Gas Condensate Field Development and
Management. Paper SPE 128448 presented at the North Africa Technical Conference and
Exhibition, Cairo, Egypt, 14-17 Febuary.
Giamminonni, D., Fanello, G., Kfoury, M, Colombo, I. and Bonzani, A. 2010. Condensate Banking
Phenomenon Evaluation in Heterogeneous Low Permeability Reservoirs. Paper SPE 131582
presented at the SPE EUROPEC/EAGE Annual Conference and Exhibition, Barcelona, Spain,
14-17 June.
Muskat, M. 1949. Condensate Reservoirs. In Physical Principles of Oil Production, Chap. 13, pages
738 –800. McGraw-Hill.
Fevang, Ø. and Whitson, C.H. 1996. Modeling Gas-Condensate Well Deliverability. SPE Res Eng 11
(4): 221–230.
Seah, Y.H., Gringarten, A.C., Giddins, M.A. and Burton, K. 2014. Optimizing Recovery in Gas
Condensate Reservoirs. Paper SPE 171519 presented at the SPE Asia Pacific Oil & Gas Confer-
ence and Exhibition, Adelaide, Australia, 14-16 October.
Singh, K. and Whitson, C.H. 2008. Gas Condensate Pseudopressure in Layered Reservoirs. Paper SPE
117930 presented at the SPE Abu Dhabi International Petroleum Exhibition and Conference, Abu
Dhabi, UAE, 3-6 November.
Whitson, C.H., Fevang, Ø. and Sævareid, A. 1999. Gas Condensate Relative Permeability for Well
Calculations. Paper SPE 56476 presented at the SPE Annual Technical Conference and Exhibition,
Houston, Texas, USA, 3-6 October.
SPE-177445-MS 11

Appendix 1
Hydraulic Fracturing Parameters

Hydraulic Fracturing Candidate Ranking

Post fracture performances are analyzed to identify candidate that is suitable for hydraulic fracturing job. Reservoir
permeability is varied from 0.5 mD – 5 mD – 50 mD in this sensitivity study. Fluid properties used in the model are taken from
HT-1 fluid sample analysis.
IPR model is built to evaluate post fracturing job based on skin reduction. The skin effect is expressed in equation A.1:
A.1

Where:
s is the skin effect interpreted as altered permeability zone
K is the original reservoir permeability
Ka is the permeability of the alter zone
rw is the well bore radius
ra is the alter zone radius
Since the model is built for gas flow, non-Darcy skin must be applied as in equation A.2:
A.2

Where s’ is the non-Darcy skin, D is the non-Darcy flow coefficient, and Qg is the gas flow rate.
Dimenssion less fracture conductivity is expressed in equation A.3:
A.3

Equation developed to represent the Cinco-Ley and Samaniego relationship curve is used to estimate rw’:
A.4

The IPR for gas well is generated using equation A.5:

A.5

Fig. A.1 shows the post fracture evaluation for three cases (K ⫽ 0.5 mD, 5 mD, and 50 mD). The results show that low
permeability resvoir (K ⬍ 10mD) is an excellent candidate for hydraulic fracturing.
12 SPE-177445-MS

Figure A.1—Post fracture evaluation results

Hydraulic Fracturing Parameters

A sensitivity analysis is conducted to investigate the effects of fracturing parameters on well performance. The selected
parameters used in the study are fracture permeability, fracture height, fracture width, and fracture length. Table A.1 represents
range variations for the above parameters.

Table A.1—Sensitivity Parameters

Fig. A.2 shows that fracture permeabilities do not impact on well productivity. Productivity improvement is not impacted
significantly by fracture half length (xf). This is well understood because even at min. case, xf ⫽ 50 m is already greater than
near wellbore damage radius, therefore fluid can bypass damage zone through fracture channels. This suggests that an optimum
fracture half length should exist to maximize the hydrocarbon recovery.
SPE-177445-MS 13

Figure A.2—Sensitivity analysis: a) Fracture half-length; b) Fracture width; c) Fracture conductivity.