A Reservoir Simulation of The Muara Laboh Field, Indonesia
A Reservoir Simulation of The Muara Laboh Field, Indonesia
A Reservoir Simulation of The Muara Laboh Field, Indonesia
ABSTRACT
This paper describes the application of reservoir simulation to evaluate the fracture permeability distributions as well as to forecast
future reservoir performance of the Muara Laboh Geothermal Field. Data from six full diameter exploration wells were used in the
evaluation. Dual porosity approach was implemented using PetraSim. Reservoir model was coupled with wellbore model to complete
forecast simulation.
1. INTRODUCTION
Muara Laboh Geothermal Field is located in South Solok Regency, West Sumatra, Indonesia. The prospect is situated along the segment
of the Great Sumatra Fault. Three principle surface manifestations, including Patah Sembilan (PS) fumaroles to the south, Idung
Mancung (IM) fumaroles in the middle, and Sapan Malulong (SM) springs to the north define the presence of geothermal potential in
the area. The later geological, geochemical, and geophysical surveys, as well as gradient-hole drilling delineate the potential extension
of the geothermal system. Figure 1 presents the location of the Muara Laboh field, and its surface features distribution. The blue line in
Figure 1 indicates boundary between concession area to the north and the national forest to the south.
In 2012-2013 PT Supreme Energy Muara Laboh (SEML) drilled six full diameter exploration wells. All wells were drilled directionally
from the well-pad located within the concession area. Two of them were completed within geothermal reservoir to the south, underneath
the national forest. The results of the exploration have been used to define the temperature distributions, resource size, permeable
structures, and hydrology of the system. At the initial state, it appears that hot geothermal fluid is upwelling from the deep part of the
resource to the south, close to the PS fumaroles. The fluid flows to northwest along the path almost parallel to NW-SE structures. When
fluid reaches just north of Pad-H, it turns to follow NE-SW permeable structures, passing ML-A1 well, to Idung Mancung (IM)
fumaroles before it then flows toward Sapan Malulong (SM) springs to the north along the NNW-SSE permeable structures. Throughout
its movement from the southern upwelling to the northern outflow, the reservoir fluid experiences a temperature decrease from 310oC to
200 oC
In the most part, the reservoir contains a single phase liquid. A small developed steam cap exists in the shallow part of the reservoir,
beneath IM fumaroles as confirmed by one of the exploration wells. The field conceptual model is illustrated in Figure 2.
1
Situmorang et al.
Figure 2: North-south cross section showing conceptual model of Muara Laboh field
During the course of exploration drilling program, the comprehensive well testing program including pressure-temperature (PT)
surveys, pressure-temperature-spinner (PTS) surveys, and interference test were conducted, providing fundamental information needed
for reservoir evaluation. The well testing information gathered has been used to verify the numerical model of the field previously
constructed. Through model calibration processes, the fracture permeability distribution of the model was reassessed. Once the reservoir
model has been well calibrated, it was then used to forecast reservoir response to exploitation.
Reservoir simulation is very time consuming and offers a non-unique solution with a high degree of uncertainty. With the intention to
providing a careful recommendation for field development plan, the Muara Laboh Numerical Model was built deterministically by
applying these following approaches:
1. All possible lateral or horizontal reservoir extensions delineated by Geophysics, Geology and Geochemistry information but
have not been proven by drilling were treated as non-reservoir. Most of the potential extension comprises the part of resources
beneath the national forest, far to the south, that presently cannot be accessed by drilling from the concession area.
2. Reservoir area, thickness, and permeability structures assigned in the model were constrained by the measured wells’ pressure
and temperature, and reservoir information collected during the interference test. The reservoir parameters for the model were
obtained through a process involving a lot of iterations of model calibration.
3. Conservative values were assigned to the parameters that are poorly constrained such as matrix permeability, porosity,
fracture spacing, and fracture porosity. The values assigned are based on the other similar geothermal fields.
PetraSim, a pre- and post-processor for TOUGH2, was used to develop the reservoir model using dual porosity approach. In addition to
PetraSim, separate functions and modules were created in Microsoft Excel Visual Basic for Applications (VBA) to accomplish the
reservoir simulation process. In the following paragraphs the workflow of Muara Laboh Reservoir Simulation is presented.
2
Situmorang et al.
Rock materials in the model are specified mainly upon different fracture permeability. In general, the rock distribution is illustrated in
Figure 5 and described in Table 1. These fracture permeability distribution were obtained from over a hundred times of trial and error
3
Situmorang et al.
until natural state and production history matching are achieved. The permeability values obtained through this iteration process is
comparable to the values reported in some other volcanic hosted geothermal fields such as Ogiri Geothermal Field (Itoi et al., 2010) ,
and on the other fields that the authors have worked before.
Ro c k C o lo r Kx, k y, k z ( m D ) T o p , b o t t o m ( m a s l)
R OC K4 5, 5, 20 -100, -1600
The results of the natural state calibration are summarized in Figure 6. Pressure and temperature in the model have been successfully
matched against the measurements. Figure 7 presents a north-south cross section of the model natural state temperature, showing a good
agreement with the conceptual model previously presented in Figure 2.
Figure 7: A north-south cross section of the model natural state temperature distribution
A plot of pressure changes over time can be generated from the collection of PT shut-in data. Pressure at feed point or major permeable
zone is extracted from each survey and plotted on one chart against time. Figure 8 shows how the pressure values at feed point of ML-
A1 have been extracted to generate a profile of pressure changes with time. It appears that the pressure history obtained from the static
PT surveys correlates with the well production history, showing a pressure buildup after the production stopped.
Figure 8: ML-A1 pressure changes record extracted from PT shut-in surveys collection
Furthermore, during the interference test, pressure changes were also recorded in the observation wells ML-B1 and ML-H1. The
pressure data from capillary tubing obtained during the interference test was combined with the pressure changes recorded from the
static PT surveys, which are presented in Figure 9.
5
Situmorang et al.
The calibration results are presented in Figure 10. A reasonable agreement between model and measured pressure changes was achieved
by assigning fracture permeability distribution described earlier in Figure 5. Three permeability barriers (ROCK 7) were assigned, one is
NEE-SWW direction and the other two are SE-NW, parallel one and another. The resulted fracture permeability and temperature
structures infer that the southern deep heat sources mainly move upward through permeability of “high temperature” zone (ROCK4) and
change direction through “high temperature and permeable” zone (ROCK3) before move laterally through “northern narrow permeable
outflow” zone (ROCK2) and finally ascend to the surface. Due to the present of SW-NE permeability barrier, only a small fraction of
the deep heat sources moving through ROCK5 & ROCK6.
4.2 Assumptions
The field is assumed to utilize dual flash technology, allowing further production of steam at lower pressure (LP) from separated brine
produced by high pressure (HP) production wells. At the same amount of energy extracted from two phase reservoir fluid, the dual-flash
technology produces higher power generation compared to single flash plant. Steam demand, separation pressure, operating wellhead
pressure (WHP), and operation availability assumptions of HP and LP systems are given in Table 1.
D e s c rip t io n Un it HP LP
Steam deliverability of a geothermal well depends on reservoir fluid enthalpy, productivity index (PI, a measure of reservoir
permeability), reservoir pressure, and casing geometry. In wellbore modeling, all those factors are processed to estimate well steam
deliverability. Presently, production capacity of future production wells in Muara Laboh is still uncertain as reflected by the wide range
7
Situmorang et al.
of production capacities of two wells completed inside the reservoir. The main uncertainties lies on the permeability (PI) and fluid
enthalpy wells would encounter in the targeted reservoir. To define a reasonable assumption on average initial well steam deliverability
of future wells, a Monte Carlo simulation, involves wellbore modeling, is carried out.
The future well is modeled to have encountered shallow and deep fractures (feedzones) at around 1000 and 1400mMD respectively. The
shallow feedzone is assumed two-phase with enthalpy ranging from 1043 to1559 kJ/kg, while the deep feedzone is assumed liquid with
1038 to1110 kJ/kg enthalpy range. The productivity index (PI) values of shallow and deep feedzones are in the range of 1-90 kg/s-bar
and 1-20 kg/s-bar respectively. Big hole casing configuration is used, with 13 3/8” production casing set at around 800mMD. A chance
of well penetrating each feedzone is assigned 80%. This is translated into a modeling computation in that whenever a random number
value is higher than 0.8, the corresponding feedzone will not be active (i.e. it will not be contributing to a well flow). If both feedzones
are not active, the estimated production capacity will be equal to zero. Fluid enthalpy, productivity index (PI), reservoir pressure, feed-
zone locations, and casing configuration used in the calculation are presented in Figure 12. These input values are decided base on the
drilling and well testing data collected during the exploration program.
Figure 12: Assumptions used in wellbore model for future well initial steam deliverability calculation
Well is set to flow at operating wellhead pressure 10.1 bar and separation pressure 9.1 bar. A total of 10000 computation trials were
done to generate the initial steam deliverability probabilistic curve shown in Figure 13.The modeling outcome shows initial steam
deliverability of 15.6, 30.3, and 44.1 kg/s at the probability of occurrence P90, P50, and P10 respectively. For this instance, the P90
value of 15.6 kg/s is assigned as future well average initial steam deliverability assumption in the forecast simulation.
P ro b a b ilit y In it ia l S t e a m R a t e ( k g / s )
P 10 44.1
P 50 30.3
P 90 15.6
Figure 13: Probabilistic curve of future well initial steam deliverability estimates
8
Situmorang et al.
Table 2 summarizes brine injection temperature, condensate injection flow rate, condensate injection temperature, and average initial
well steam deliverability assumptions. Location of production and injection wells in the simulation grids is given in Figure 14.
Production wells are put in the “high temperature” and “high temperature and permeable” zones. Brine injection wells are assigned in
the “northern narrow permeable outflow”, around 2 km from the closest production well.
D e s c rip t io n Un it Va lu e
o
B rine inje c tio n te m pe ra ture C 144
Figure 14: Location of future production and injection wells in the numerical model
4.3 Methodology
The methodology designed and applied allows more realistic forecast simulation.
1. All assigned production wells are constrained to operate at a constant 10.1bar WHP for HP and 5bar for LP throughout the life
of the project. This is done by coupling the numerical model with a wellbore model.
2. At the beginning of simulation, each individual well Productivity Index (PI) value is adjusted to meet the assumed steam
deliverability 15.6 kg/s. As the simulation is executed (i.e. reservoir exploitation begins), the steam deliverability of individual
well is updated every three months following the pressure and enthalpy changes caused by the exploitation.
3. Total steam available of active wells of HP & LP systems are evaluated to obtain production decline. If total steam available is
lower than steam demand, additional active well(s) (make up well) is put online.
4. Mass extraction imposed in the model is equal to 96% of the total HP & LP steam demand. If the total steam available is
higher than the demand, the imposed production rate of all production wells is proportionally reduced.
5. If at some point in the future an HP well is no longer able to sustain production at required high pressure, the well can be
switched to feed the LP system.
6. Brine injection flow rate is evaluated base on flowing enthalpy changes. Condensate injection flow rate is set constant.
Figure 15 presents the flowchart of simulation process. Forecast modules are created to allow communication between reservoir
simulator and wellbore simulator as well as to run forecast simulation more automated.
9
Situmorang et al.
We ll In it ia l S t e a m N u m o f P ro d u c t io n We ll N u m o f In je c t io n We ll
Cas e D e liv e ra b ilit y P a d-A P a d-H
P a d-D P a d-B
(kg / s ) ( E xc lu d e M L- A 1) ( E xc lu d e M L- H 1)
( B rin e ) (C o nde ns a te )
"E xt ra c t io n S t ra t e g y" "E xt ra c t io n S t ra t e g y"
4 4
1 15.6 3 1
"S ha llo w & De e p" "S ha llo w & De e p"
4 4
2 15.6 3 1
"S ha llo w" "S ha llo w"
3
"De e p"
4
3 15.6 1 3 1
"S ha llo w & De e p"
"S ha llo w & De e p"
10
Situmorang et al.
Figure 16: Case 1 total mass rate and enthalpy extracted for 30 years
Figure 18: Case 1 production wells requirement to sustain 30 years steam supply
11
Situmorang et al.
deep liquid reservoir, gives the lowest production decline. In this case, make-up wells drilling can be delayed by up to five (5) years
after start-up. It is interesting to note that whenever the portion of deep extraction increases, production decline decreases and therefore
make-up wells drilling can be delayed longer.
Figure 19: The results of forecast simulation in the form of filed wide production decline for all production cases
The analyses show that having more extraction from deep liquid reservoir will sustain steam supply longer before drilling make-up well
is required. Therefore, for this particular example, it is recommended that future production wells be targeted both to shallow and deep
reservoir (below 500masl).
5. CONCLUSION
From this reservoir simulation work the following conclusions can be drawn:
1. Reservoir simulation has been applied to assess fracture permeability distribution of Muara Laboh field using dual porosity
formulation.
2. The numerical model has been successfully calibrated to initial state and production history data.
3. A workflow has been designed and applied to enable more realistic forecast simulation
4. The applied workflow has allowed different development scenarios to be assessed thoroughly, enabling a careful
recommendation for field development plan to be made.
ACKNOWLEDGEMENTS
The authors wish to thank the management of SEML for the permission to publish this paper. Discussions with SEML Subsurface team
members are very much appreciated.
REFERENCES
Itoi, R., Kumamoto, Y., Tanaka, T., and Takayama, J.: History Matching Simulation of the Ogiri Geothermal Field, Japan, Proceedings,
World Geothermal Congress 2010, Bali, Indonesia (2010).
Supreme Energy Muara Laboh: Muara Laboh Geothermal Power Project-Stage 1 Development Feasibility Study, Unpublished Report,
(2014).
Supreme Energy Muara Laboh: Muara Laboh Development Drilling Strategy, Unpublished Report, (2016).
12