IPA13-E-189

PROCEEDINGS, INDONESIAN PETROLEUM ASSOCIATION

Thirty-Seventh Annual Convention & Exhibition, May 2013

INVESTIGATION OF THE IMPACT OF CONDENSATE BLOCKAGE ON WELL

PRODUCTIVITY IN A TIGHT GAS CONDENSATE RESERVOIR – CASE STUDY OF TAWES
FIELD

Agung Ertanto*
Lukman Hidayat*
Imaduddin Kartoatmodjo*
Ludy Andria*

ABSTRACT This paper focuses on the use of numerical

simulation to investigate the condensate blockage
The Tawes Gas field in Block B, South Natuna Sea phenomena in the reservoir and also to examine if
was discovered in 1988 by the Tawes-1 exploration hydraulic fracturing could still be beneficial to
well and appraised by the Tawes-2 well in 1999. increase the productivity of a gas reservoir with
The primary reservoir is the Gabus-3B sandstone. condensate blockage.
Economic development of the field is challenging
due to its poor quality reservoir and fluid properties. INTRODUCTION

In low permeability reservoirs, rich gas produced In a gas condensate reservoir a condensate bank will
below the dew point has a negative effect on be built around the wellbore when the flowing
permeability in the near wellbore area. For tight gas bottom hole pressure falls below dew point
reservoirs such as the Gabus 3B, high drawdown pressure. The condensate bank can reduce the well
pressure is needed to flow gas, consequently productivity index. This effect is most severe in a
condensate saturation banking will occur near the low permeability gas reservoir when high
wellbore as the pressure drops below dew point drawdown is needed to flow gas from reservoir into
pressure. The buildup of condensate saturation is the wellbore. Condensate blockage is important if
called condensate blockage. This effect causes a the pressure drop from the reservoir to the wellbore
significant drop in gas relative permeability, which is a significant percentage of the total pressure drop
reduces the productivity of the well. from reservoir to delivery point (e.g. a surface
separator) from the time a well goes on decline.
To investigate and to model the condensate Reservoirs with low-to-moderate permeability (<10
blockage phenomena in a numerical reservoir – 50 md) are often “problem” wells where
simulation, a radial single well model was first condensate blockage must be handled properly.
used. For a sector or full-field model with a coarser Wells with high kh products (>1,000-10,000 md-ft)
Cartesian grid, Generalized Pseudo-Pressure (GPP) are typically not affected by reservoir pressure drop
was found to be the best method to capture the because the well’s deliverability is constrained
phenomena without the need to use fine grid almost entirely by the tubing. In this case,
simulation. condensate blockage is a non-issue (Whitson et al.
1999).
Hydraulic fracturing was evaluated as a method to To accurately forecast gas condensate well
enhance productivity. In an earlier initial productivity, a fine-grid simulation is required to
assessment, a study using a mechanistic model was model the formation of the condensate banking. In
performed and concluded that hydraulic fracturing a single well simulation a very fine-grid may be
could increase the well deliverability and the total used but, a full-field model using fine-grid
recovery of the field. However, the potential effect simulation would require long run times.
of condensate blockage was not taken into account
in that study. According to Fevang and Whitson (1995), when a
well is being depleted, three regions form around
* ConocoPhillips Indonesia Inc. Ltd the well bore:

© ,3$WK$QQXDO&RQYHQWLRQ3URFHHGLQJV
 Region 1: An inner near-wellbore region where TAWES FIELD OVERVIEW
both gas and oil/condensate flow
simultaneously at different velocities. The Tawes Field is located in the West Natuna
Basin of the South China Sea, Indonesia. It lies in
 Region 2: A region of condensate buildup water depths of approximately 300 feet and is
where only gas is flowing. located approximately 10 km Northeast of the
Belanak Field and 12 km southwest of the Kerisi-
 Region 3: A region containing single-phase Hiu Field (figure 1). The field was discovered in
(original) reservoir gas. 1988 by the Tawes-1 exploration well which
encountered gas-bearing sands within the Upper
Fevang and Whitson show that the loss in well Gabus Massive and Gabus Zone-3B in the East fault
deliverability in gas condensate wells is block. The hydrocarbon accumulation is trapped in
underpredicted in full-field simulation models when a southwest-northeast trending 4-way dip closed
using the traditional method to calculate rate. They anticline cut by two northwest-southeast trending
proposed a new method, called the Generalized normal faults and sealed by a thick Barat shale top
Pseudo Pressure (GPP) method based on seal. The two major normal faults divide the
observations of the three flow regions for gas structure into 3 fault blocks (figure 1). Gross sand
condensate systems. thickness varies from 50 to 520 ft. Two DST’s were
conducted at Tawes-1: DST-1 over the Gabus Zone-
This method has been implemented in some 3B and DST-2 over the Upper Gabus Massive. Only
commercial reservoir simulators. Barker (2005) DST-1 was succesful and flowed 0.785 mmscfd at a
discussed his experience with simulation of combined rate.
condensate banking effects in various gas
condensate reservoirs using the GPP method. He Following acquistion of 3D seismic acquisition in
found that the GPP method is not always valid. In 1997 the Tawes-2 well was drilled in 1999 3 km
this paper we evaluate the application of the GPP west of Tawes-1 to appraise the Tawes field west
method for modeling condensate blockage in a fault block. Tawes-2 encountered gas-bearing sands
coarse model using data from the Tawes field (West within the Upper Gabus Massive and Gabus Zone-
Natuna basin, Block B, Indonesia). 3B. Gross sand thickness varies from 50 to 800 ft.
Two DST’s were conducted: DST-1 over the Lower
In tight sandstone reservoirs, hydraulic fracturing is Gabus sands and DST-2 over the Gabus Zone-3B.
frequently required to improve productivity. In a Only DST-2 was successful and flowed gas at a rate
gas condensate reservoir, fracture stimulation of 2.7 mmscfd.
reduces pressure drawdown and thus leads to less
liquid dropout in the reservoir. In a fractured well In 2012, Tawes-3 was drilled to appraise the central
simulation, the size and aspect ratio of the grid cells block. Gas bearing sands were again encountered in
have a strong influence on our ability to correctly Gabus Zone-3B. Unlike the two offset wells,
predict the flow behavior for different well types, Tawes-3 was perforated underbalanced to minimize
fluid compositions and recovery processes. Several the formation damage. A DST was conducted and
methods are available for modeling hydraulic flowed gas at rate of 4.6 mmscfd.
fracturing using reservoir simulation including:
negative skin, equivalent effective well radius, The hydrocarbon reservoirs are relatively deeply
transmissibility multiplier and Local Grid buried (approximately -9,500ft TVDSS) which has
Refinement (LGR). Negative skin can provide an negatively impacted reservoir quality. The reservoir
approximation of flow due to the increased fracture exhibits tight sand properties, the average porosity
permeability within the coarse grid, but cannot is 12%, average permeability 0.1 - 2 mD, and
model condensate banking effects. Furthermore, average NTG 47%. The reservoir is normally
using a large negative skin results in errors in the pressured at around 4400 psia with high formation
mathematical calculation that can abort the temperatures (approximately 320 deg F). The
simulation run due to convergence issues. The observed condensate yield is approximately 40 - 70
coarse grid model was not suitable as it is unable to bbls/mmscf. Up to 40 ppm H2S and 10% CO2 were
correctly capture the condensate blockage process measured during the well test.
in the reservoir. LGR around explicit fractures was
therefore considered the most suitable approach to Overall the Gabus Formation is interpreted to have
examine the effect of hydraulic fracturing on been deposited as a sand prone interval in a
production enhancement at the Tawes field. distributary channel system to more distal
lacustrinal fluvio-deltaic system. The sediment second model a straight line relative permeability
transport direction is interpreted to be from the relationship was incorporated as representative of a
southwest. The retrogradational stacking pattern in miscible curve. The results from the two models
the Gabus Formation is probably due to regional were then compared. If the end of the plateau
subsidence and a rise in the lake base level which periods are similar, condensate blockage will have
reached its peak during deposition of lacustrine only a minor impact on the deliverability. If a larger
shales of the extensive Barat Lake, the main sealing difference is seen, a more detailed evaluation is
facies in this area. required.

METHODOLOGY A comparison of the Tawes models using rock and

miscible relative permeability curves is shown in
To study the effect of condensate blockage in a figure 3. The plateau period of the first model is 10
reservoir, different methods have been proposed by months whereas in the second model it is 6.2 years.
various authors. In this study, the Fevang and It clearly shows that condensate blockage occurs
Whitson method was used (Fevang and Whitson and will impact the productivity of the Tawes field.
1995).
A fine grid radial model can accurately describe In order to investigate condensate accumulation
pressure and saturations in the reservoir around the around the wellbore, condensate banking
wellbore and can therefore be used to model well development is plotted against radial distance from
deliverability changes due to liquid banking in the the wellbore (figure 4). It is clearly seen that
near wellbore area. All proposed methods to model condensate accumulation increases with time. As
the condensate accumulation are compared to the the dew point is close to the initial pressure, the
fine grid radial model. The radial grid was built condensate bank forms immediately after the
using constant ratio of 1.5. commencement of production. DST results in
Tawes-2 showed that a very high pressure
Fluid properties were taken from Tawes-2 PVT drawdown (3500 psi) is needed to flow gas at a rate
analysis results. The initial reservoir pressure is 200 of 2.7 mmscfd. This high drawdown results in a
psi above the dew point pressure. Condensate yield flowing bottom hole pressure which is much lower
is 40-70 stb/mmscf, hence the fluid is classified as a than the dew point pressure.
lean gas condensate. The components were lumped
into 9 pseudo components. In late 2012 the Tawes-3 appraisal well was drilled
with the objective of assessing commerciality of the
Initial Assessment. tight reservoir including data gathering to assess the
impact of fracture stimulation. The data acquisition
The first question was would the condensate program was intensive, and included pressure
blockage effect have an impact on the well points, bottom hole sampling, whole core, side wall
deliverability? As a best practice, an initial core, saturation log, vertical seismic profile, image
assessment should be performed before a more log, DST, PLT and Pin Point Injection tests. At the
detailed resource intensive evaluation is conducted. time of writing data evaluation and laboratory
For this purpose, two fine grid radial models were analysis were in progress. However, the preliminary
built to assess the importance of condensate results of pressure transient analysis are shown in
blockage with respect to production performance. figure 5. An analytical approach was conducted.
The fine grid models contain 20 x 1 x 1 grid cells, Three radial stabilizations were observed (figure 5).
with an inner grid cell size of 0.13 ft and an outer The first stabilization shows a zone in the vicinity
radius of 432 ft. A single numerical layer with a of the well with high capillary number which
thickness of 80 ft was used. The porosity and increases the gas relative permeability. The second
permeability were set at 12% and 1 md, stabilization shows a zone near to the well with
respectively. No Special Core Analysis (SCAL) increased condensate saturation and lower gas
data were available for this study; hence a relative mobility. The third stabilization shows a zone away
permeability curve from the analogue North Belut from the well with the initial liquid condensation
field was used. A vertical lift curve was also saturation. It indicates that condensate blockage is
required for this step. A grid schematic is shown in likely near the wellbore. A radial composite
figure 2. The maximum gas production was reservoir model with three regions best matches the
constrained to 2.5 mmscfd to provide a plateau results. However, the software only permitted use of
period for comparison purposes. In the first model, a dual radial composite model. Consequently, it is
the rock relative permeability was used; while in the difficult to match the third stabilization.
Sector Model. considered sufficient for the future full field model
provided GPP is included.
A coarse grid cell size of 100 m x 100 m is
normally standard for planning purposes during Tawes field consists of stacked independent sands
early phase development, and works well for dry separated by shale stringers. In order to evaluate the
gas reservoirs. However, for a gas condensate GPP method in a multi layered reservoir, the
reservoir special well treatment is required to get a models were divided into 8 equal layers of 10 ft
more correct production forecast. thicknesses in the z-direction. The permeability in
the z-direction varied from 0.5 to 2 md. The
A coarse grid model containing a grid of 20 x 20 x 1 petrophysical properties for each layer were held
was employed for comparison with a benchmark constant. Figure 13 shows the simulated gas
model. The fine grid radial model from the initial performance of fine grid and coarse grid models.
assessment was used as the benchmark. Both radial Figure 14 and 15 show the BHP and cumulative gas
and Cartesian models had the same reservoir of the two models. The results are similar, which
volume and rock and fluid properties. A maximum indicates that the GPP method with multi layered
gas rate of 3 mmscfd and a minimum bottom hole reservoirs works well for forecasting the gas
pressure of 1000 psia were used as constraints. performance in a coarse model.

Figure 6 shows the gas rate comparison from fine Hydraulic Fracturing.
grid radial and coarse grid Cartesian models. The
coarse grid model overestimates the gas production In order to model the effect of hydraulic fracturing
because the bottom hole pressure (BHP) in the on production at Tawes field in the presence of
coarse grid is higher than the radial model (figure condensate blockage, a cartesian grid with LGR
7). Furthermore, the coarse grid cannot account for around the wellbore was used. The grids around the
the reduction in gas phase relative permeability near wellbore and fracture tips had to be fine enough to
the wellbore and therefore cannot model the capture the pressure and saturation distributions.
production loss correctly. Figure 8 shows the The grid chosen therefore consisted of 9 x 27 x 6
comparison of cumulative gas production. grid cells with very small grid cells at the center
(figure 16). The thickness and matrix permeability
The generalized pseudo pressure (GPP) method was of each layer are shown in table 1. The drainage
tested using rock and fluid data from the Tawes area was assumed to be 320 acre. The representative
field to see if it is accurate enough to be used in a fracture width in the model was defined based on
coarse model to forecast gas production and thereby the formula:
eliminate the need of a fine grid model. The
accuracy of the GPP method is somewhat kf w = kG wG.
dependent on the well grid size, generally requiring:
dx = dy = 50 – 100 m for lean gas condensates; and where, kf and w are fracture permeability and
dx = dy = 100 – 200 m for rich gas condensates fracture width, respectively; wG is grid block size
(Singh 2008). The number of grid cells in the coarse representing fracture width in the model.
model is a sensitivity parameter evaluated to find
the optimum grid size giving similar results to the A fracture permeability of 50,000 md and fracture
fine grid model. width of 0.2 in was used. The grid size in the
fracture region was 2.5 ft giving a corresponding
The coarse grid with GPP is compared to the fracture permeability of 332 md. A straight line
benchmark model (figure 9, 10 and 11). Both relative permeability relationship was used for
models have an initial gas in place of 3.5 BCF. The fracture regions. A single vertical well located in
results show that the gas performance is in very the centre of the reservoir was modeled. The well
close agreement between the two models. To was modeled to produce for 10 years with bottom
evaluate the effect of reservoir volume, gas in place hole pressure constant at 1000 psi. The fracture half
was increased to 11 BCF. The results are quite length was assumed to be 250 ft. For comparative
similar (figure 12), but the fit is not as good as for purposes, an unfractured vertical well was
the models using lower volumes. Figure 12 shows compared to a fractured vertical well. Figure 17 and
the comparison assessing sensitivity to different figure 18 show the comparison of gas rate and
grid sizes: 52 m, 75 m and 95 m. In the models with cumulative gas between the two wells. It is
GPP, grid size does not have a significant effect on observed that the fracture stimulation could improve
the results. Therefore, a grid size of 95 ft was the initial production about 4 times that of an
unfractured well. Gas recovery of the fractured well ACKNOWLEDGEMENTS
was 14 BCF, about double the recovery of the
unfractured well. The authors would like to thank ConocoPhillips
Indonesia Inc. Ltd, MIGAS and Block B Partners
(Chevron and Inpex) for permission to publish this
Fracture Half Length Sensitivity. paper.

Fracture length and fracture conductivity control the REFERENCES

effectiveness of hydraulic fractures. In order to
evaluate the effect of fracture half length on the Barker, J.W.: “Experience with Simulation of
well productivity, simulation runs with different Condensate Banking Effects in Various
fracture half lengths were performed. Fracture half Gas/Condensate Reservoirs” paper IPTC 10382
lengths used were 250 ft, 350 ft and 500 ft. The presented at the 2005 International Petroleum
simulations show that increasing fracture half length Technology Conference, Doha, Nov. 21 – 23
results in production improvements and increased
recovery (figure 19 and 20). Fevang, Ø., Whitson, C.F.: “Modeling gas
Condensate Well Deliverability”, paper SPE 30714
presented at the 1995 SPE Annual Technical
CONCLUSIONS Conference and Exhibition, Dallas, Oct. 22 – 25.
Singh, K., Whitson, C.F.: “Gas Condensate
1. A fine grid radial model shows that condensate Pseudopressure in Layered Reservoirs, paper SPE
blockage occurs in Tawes field and results in 117930 presented at the 2008 Abu Dhabi
reduced well deliverability. International Petroleum Exhibition and Conference,
Abu Dhabi Nov. 3 – 6
2. A GPP model is required for future evaluations
of the Tawes full field development scenarios in Whitson, C.F., Fevang, Ø., Sævareid, A.: “Gas
order to take into account the effects of Condensate Relative Permeability for Well
condensate banking. Calculations”, paper SPE 56476 presented at the
1999 SPE Annual Technical Conference and
3. Hydraulic fracturing can improve production Exhibition, Houston, Oct. 3 – 6.
performance at Tawes field
The information contained herein is for
4. Tawes full-field models should incorporate presentation and discussion purposes only.
condensate blockage and hydraulic fracturing ConocoPhillips does not make any representations
effects. or warranties regarding content accuracy.
 TABLE 1

THICKNESS AND MATRIX PERMEABILITY OF EACH LAYER FOR MODELING

HYDRAULIC FRACTURING

Layer Thickness ( ft) Permeability (md)

1 16 0.35
2 11 0.79
3 8 1.08
4 27 0.61
5 10 0.57
6 34 0.76  
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 Figure 1 - Tawes field location in Block-B PSC area.

Figure 2 - Schematic of fine grid radial model (top view).

 Figure 3 - Gas rate of radial model with rock and miscible curves.

Condensate Banking Development

0.35

1 month
0.3
3 month
7 month
Condensate Saturation

0.25 1 year
2 year
0.2 5 year
10 year
0.15

0.1

0.05

0
0.1 1 10 100 1000 10000

Distance from wellbore (ft)

Figure 4 - Condensate banking development in the near wellbore.

Figure 5 - Tawes-3 log-log derivative plot of main Pressure Build Up (PBU).

Figure 6 - Gas rate of radial and Cartesian models.

 Figure 7 - BHP of radial and Cartesian models.

Figure 8 - Cumulative gas of radial and Cartesian models.

Figure 9 - Gas rate of single layer radial and Cartesian – GPP models.

Figure 10 - BHP of single layer radial and Cartesian – GPP models.

 Figure 11 - Cumulative gas of single layer radial and Cartesian – GPP models.

Figure 12 - Gas rate of single layer radial, Cartesian coarse grid, Cartesian – GPP models with more
gas in place.
Figure 13 - Gas rate of multi layer radial and Cartesian – GPP models.

Figure 14 - BHP of multi layer radial and Cartesian – GPP models.

 Figure 15 - Cumulative gas of multi layer radial and Cartesian – GPP models .

Figure 16 - Schematic of Cartesian grid with LGR to model hydraulic fracturing (top view).

 
 Figure 17 - Gas rate of unfractured and fractured wells.

Figure 18 - Cumulative gas of unfractured and fractured wells.

 Figure 19 - Gas rate of fracture half length sensitivity.

Figure 20 - Cumulative gas of fracture half length sensitivity.