IPA14-E-065

PROCEEDINGS, INDONESIAN PETROLEUM ASSOCIATION

Thirty-Eighth Annual Convention & Exhibition, May 2014

AN INTEGRATED EVALUATION MODEL OF KAJI-SEMOGA FIELD WATER INJECTION

SYSTEM, INCLUDING SUBSURFACE AND SURFACE FACILITIES

Kamal Hamzah*
Adam Fatchur Rohman*
Yoseph Itok Arseto*
Virzi Firmansyah*

ABSTRACT Total Kaji-Semoga field liquid lifting capacity has

increased from 179,000 to 184,500 BLPD.
A water injection (WI) system plays a significant Elimination of oil deferment of 236 BOPD (caused
role for an oil field that is experiencing liquid lifting by shutting in high water cut wells due to water
with high water cut and production via handling constraint) has been achieved.
waterflooding method. Kaji-Semoga field is an oil
field with an average water cut of 93% and liquid INTRODUCTION
lifting of 180,000 BLPD. In order to achieve the
targeted oil production, it is a priority to increase Kaji-Semoga Field has 3 major oil producing layers
the liquid lifting rate. Artificial lift conversions, that consist of Baturaja limestone formation, Telisa
additional wells through infill drilling and field sandstone formation and Talang Akar sandstone
station upgrading have been implemented to formation. Baturaja limestone formation is the
optimize oil production. Furthermore, the water biggest contributor to oil production from Kaji-
injection system handles produced water to be Semoga field. Following the primary depletion of
injected as part of a field strategy to improved reservoir pressure, pressure maintenance has been
reservoir seep. In addition, optimized lifting implemented in the Baturaja formation at Kaji-
capacity eliminates oil deferment caused by water Semoga Field since 1999 by converting some
handling constraints. peripheral wells to water injection. Then, in 2003, a
full field waterflood was implemented to optimize
An integrated evaluation model of Kaji-Semoga oil recovery from Baturaja Formation.
water injection system was created to optimize
water injection rate and to de-bottleneck the water In a mature field such as Kaji-Semoga Field, with
flow system. A surface facilities model was an average water cut of about 93% of liquid
developed and simulated using integrated surface- production and liquid lifting reaching 179,000
subsurface software to evaluate the surface facilities BLPD, the main challenge is how to increase liquid
constraint. Production surveillance, analysis and lifting from an individual well to combat the oil
forecasting software was utilized to record the hall- production decline. An increase in liquid lifting
plot performance history of water injection wells brings more formation water produced from the
(WIW) to identify subsurface constraints. Based on reservoir and raises the challenge of managing the
the simulation results, some activities consisting of produced water to be reinjected into the reservoir.
well stimulations, water injection pump Increased produced water limits the ability to
optimization and the installation of additional optimize the waterflood program and also to
pipeline network were executed. The contribution maintain zero discharge to the environment.
of these activities to the field water injection system
was then evaluated. Figure 1 describes the liquid lifting trends in Kaji-
As a result, manifold system pressure has decreased Semoga Field since 2001. It shows that to maintain
from 1100 psig to 1050 psig, whilst average water oil production along with the water cut increment,
injection rate has increased from 168,200 BWPD to the liquid lifting had to be increased, which resulted
172,800 BWPD and monthly VRR (Voidage in an increase in water production. This provoked
Replacement Ratio) has increased from 0.911 to more efforts to manage the water production, water
1.048 for Kaji Field. injection rate and the capacity of water injection
system (field station, network and water injection
* Medco E&P Indonesia
wells) to prevent problems that might constrains 3. Water injection network simulation must be
liquid lifting. Figure 2, shows the increased number updated regularly to evaluate the capacity of
of water injection wells since 2003, which meant water injection network and water injection
there was an additional network and surface wells to accommodate the future plan of
equipment to be managed. An integrated model was development related to liquid lifting and liquid
then created to manage all the water injection injection. This activity takes some dedicated
system parameters. effort and time due to the relatively complex
water injection system in Kaji-Semoga field.
The integrated network modeling is a tool that
incorporates all water injection system components Network Overview
(reservoir, well by well injection, surface facilities
and pipeline networking) into a single model for The water injection system model covers the water
easier observation and evaluation. This model injection network in Kaji-Semoga field, as shown in
consists of all water injection well modeling, all Figures 3 and 4. It starts from the piping network
surface facilities modeling and all pipeline network inside the field stations and flow-line network from
modeling (well to well networks, station to station the stations to the clusters and wells. Three stations
networks and field to field networks). are involved in the water injection system in Kaji-
Semoga, i.e. Kaji station, Semoga station and Kaji
There are some concerns that should be understood Satellites station. The water injection network
in order to have good water injection management: interconnects each station. All flow-lines from the
station manifold to cluster manifold are covered in
1. Understand the maximum capacity of the this model. The model also covers 56 water
surface facilities (water injection pump and injection wells in Kaji Field located in some clusters
piping systems). that are supplied by Kaji station and Kaji Satellites
station, and 34 wells that are supplied by Semoga
2. Determine the optimum wellhead pressure and station.
water injection rate of individual wells in order
to maximize surface facilities’ capacity. METHODOLOGY
The methodology to build this integrated network
3. Plan well how to increase and/or decrease water
model started with gathering all necessary data,
injection rate in case of any problems on
including reservoir data, fluid data, wellbore data,
surface facilities.
surface facilities data and pipeline network data.
After all data have been compiled, each well model
4. Predict some scenarios on how to increase
was created, and then the surface facilities model
water injection distribution in the future (with
was built. The next step was to tie-in the well model
or without work-over).
to the surface facilities model to create an integrated
network model. History matching was conducted to
In Kaji-Semoga Field water injection operations
match the model result (production rate) with the
there are some daily challenges that require good
actual production rate. Once the history match was
management, for example:
validated, an integrated network model was ready to
be used for analysis, evaluation and enhancement
1. When one or more water injection pumps is
recommendations.
down (because of unplanned problems or
planned preventive maintenance of power plant
Gathering Data
generator set) water injection rate will decrease
and lead to rapid increase in the produced water
1. Water Injection Well data.
tank level at field stations. In this case, the
The well data needed to build the well model
emergency response plan to shut-in high water
consist of:
cut wells is applied to maintain the produced
water tank level.
a. The well type, such as production, injection
or pipeline, and flow type, such as annular
2. Problems that occur in the pipeline network,
or tubular.
such as leaks due to corrosion. This requires re-
distribution of water flow which involves
b. Reservoir data: injectivity index, IPR
opening and closing several valves along the
model, reservoir pressure, layer
water injection lines.
 temperature, parting pressure, completion collected and validated. The typical sequential steps
type, perforation depth and injection test to create a well model are shown below:
data.
1. Well identity was input into the generic nodal,
c. Reference depth that indicate well position and the flow type “tubular + annular” was
and location, consisting of: field location selected. The well type was set to injection
(offshore/onshore), KB (Kelly Bushing), instead of production.
BH (Braden Head) and GL (Ground Level)
elevations that are measured from MSL 2. Reservoir data was input into the reservoir
(Mean Sea Level). control tab. In the reservoir control tab, fluid
d. Wellbore data, consisting of: tubular and type was selected to water (because it is a water
casing configuration, well type (injection or injection well). Then, the reservoir model was
producer well), flow type (annular or constructed by editing layer pressure, layer
tubular), well trajectory and well flow temperature, parting pressure, mid-perforation
correlation. depth and the injectivity index (based on test
pressure and flow rate).
e. Water injection properties, consisting of
water specific gravity and water salinity. 3. For well orientation, vertical well
approximation was selected for directional
Those data mentioned above are the standard wells instead of horizontal well type. Next,
data needed to build a basic well model. For fluid parameters , the salinity and the water
more advance well modeling, layer parameters specific gravity, were input accordingly.
can be included.
4. Tubular configuration was constructed by
2. Surface Data.
inputting tubing size, tubing length and tubing
Major surface data that are needed for building
Outer Diameter (OD) / Inner Diameter (ID),
the model are:
while casing configuration was constructed by
a. Overall network data: Pipeline properties inputting the casing size, ID, with the casing
(length, outside and inside diameter) and tie length measured from the end of tubing up to
in location for each pipeline and also valves total depth on the tubular flow type. For annular
that are available in the network. flow type, casing length was measured from the
well head up to total depth. Well deviation data
b. Field data: Flow line and trunk line were input for deviated wells.
properties (length, outside and inside
diameter), tie in location for each flow line 5. Finally, the history matching process was
or trunk line and Cluster(s) pressure. conducted to compare the simulation model
result and the actual data field, and adjustments
c. Field station data: Well gathering
were applied if necessary.
distribution (high, medium or low pressure
system), separator(s) setting pressure,
The main output of this well model was
compressor(s) setting suction and discharge
inflow/outflow performance curves and
pressure, equipment maximum capacity,
pressure/temperature drop curves. When matching
flare system and tie in location for each
field data, the well models were ready to be
piping.
connected to the network model.
d. Field cluster data: choke opening size and
current producing well head pressures Creating Surface Model and Integrated Model
(WHP).
Commercially available PIM (production integrated
The data above were obtained from various model) software is utilized in building surface
sources, such as Process Flow Diagram, equipment facilities and integrated network model.
pipeline and flow-line survey map, design The major elements of the surface models of Kaji-
capacity data and daily surveillance data. Semoga water injection system are:
Constructing Well Injection Model
1. Three field stations, mainly consisting of:
The well model was built using generic nodal a. Water Injection manifolds.
analysis software after all well data have been b. Water Injection Pumps.
 c. SDV (shutdown valve) and block valve based on sequential-linear-programming (SLP)
representation for simulation maneuvers. technique. This technique uses a fundamentally
different approach from other software that uses
2. Water Injection lines, including trunk line and only nodal-analysis technique, so this modeling
branch line to manifold. technique can handle large and complex networks
with loops and branches. The system can perform
3. Wells or Cluster. optimization and simulation by accounting for
a. 90 water injection wells. targets, objectives, and constraints anywhere in the
b. Well chokes. network as reported by Prihandono Aditama et al.
c. Block valves. (2010).
d. Water injection manifold at cluster.
History Matching
4. Only one fluid model for formation water
injected.
After all surface facilities equipment models have
Building the surface model and integrated model been built and the well models have been connected
started by modeling the surface facilities equipment into the PIM model, the next step is to conduct
in the field station to the water injection manifold. history matching between the model and the actual
To simplify the process, the station model was field condition to validate that the model is
limited from water injection pump up to the station representative of the actual condition. This phase is
manifold. The process flow diagram of stations an important part of this type of study and requires
were used as base data for building the station careful analysis of different aspects of behavior of
models, with some simplifications, such as water each object in the model, i.e. well performance,
injection pump, which was modeled as water pressure drop across pipes and chokes as reported
source, and water injection manifold, described as by Aldani Malau et al. (2011).
junction, valve and simple link. From the water
injection manifold, the link was connected by History matching of the integrated model was
pipeline (trunk line) to the clusters, which consisted conducted in this sequence:
of a manifold cluster and well head symbol.
1. Matching the pressure drop in the pipe line
The pipeline properties, such as: the length of from the manifold station to the manifold
pipeline, roughness, outside and inner diameter, the cluster.
material of fabrication, density and the elevation
angle of the pipeline were then input. Flow method, 2. Matching the pressure drop between manifold
such as: flow correlation, L factor, D factor and U clusters to the well head of water injection
factor were tuned later during history matching. A wells. Adjustments were applied so that the
flow correlation commonly used for this model was flowing well head pressures in the model were
Beggs and Brill. Beggs and Brill, Hagedorn and similar to actual pressures (in this study, it was
Brown, Gray are the set of flow correlations based set that the difference between actual value and
on the Fanning friction pressure loss equation. the model value must be less than 20 psig)
These correlations can be used for gas-liquid
multiphase flow, single-phase gas or single-phase 3. Matching the water injection well rate for each
liquid flow. Beggs and Brill is a multi-purpose individual well in the cluster.
correlation derived from laboratory data for vertical,
horizontal, inclined uphill and downhill flow of gas- 4. Matching in the trunk-line that supplied water
water mixtures. injection to more than 3 clusters.

The wellhead model in the PIM software The objective was that the pressure in each cluster
accommodates 3 type of wells i.e. Gas lift (G), ESP must be similar with the actual condition.
(E) and Natural flow (N). Water injection well can
be represented with natural flow model (N). The Model Application for Water Injection
performance well data must be imported from the Optimization
single well model that was already modeled by
using the generic nodal software. The integrated model (surface and well models) can
be used for optimization and simulation in order to
If there are no errors, the whole model can then be assess any changes in the network of water
run. The system network-optimizer technology is injection. The problem that frequently arises in
Kaji-Semoga Field is maintaining produced water the recommendation to install additional
tank level by shutting in some high water cut wells. pipeline to prevent high pressure drop or
However, the integrated model of water injection bottleneck in the pipeline system. Future
system can be very useful for the situation: development of the field related to the
increment of water injection system rate must
1. Determining the cause of low (or decreased) be considered as well.
water injection rate.
Restriction in the well can be caused by
Low water injection rate can be caused by formation damage, which may be caused by
several problems, i.e.: debris and small amounts of oil that
inadvertently have been injected along with
a. By upstream water injection system water. Scale growth in front of perforation can
problems, such as low efficiency of water also be one of the root causes of formation
injection pumps (WIP). This is indicated by damage. For preliminary evaluation, well that
low pressure discharge of the header or experienced formation damage would have
individual pump. The troubleshooting can lower injection rate compared to the adjacent
be focused at the pumps and the simulation wells in the same WHP. By this model, it will
can be localized at the field station. It can be easier to compare a well to others. For
be indicated by low tuning factor value in further investigation, hall-plot history of the
the pump model (flow, head and power). well will clearly told us the cause of low water
injection rate. Then, to overcome formation
b. By downstream water injection system damage on the well, stimulation of the wells by
problem. This is indicated by high acidizing can be conducted. The design of such
discharge pressure in the header or stimulation must consider these parameters:
individual water injection pump, caused by
restriction that occurred in the pipeline a. The target of injection rate required by the
network or at any well with low (or reservoir engineer in order to optimize the
decreased) injectivity index. pressure maintenance and waterflood
project.
For pipeline network analysis, after conducting
history matching at each node (e.g. each b. Wellhead pressure that can be generated by
manifold station, each manifold cluster, flow- the network at the well location. This can
line connection, and wellhead), it can be be simulated by the integrated network with
observed which network part has experienced the increment of flow rate (after acidizing)
pipe bottleneck by determining the velocity of in the pipeline that is directed to the well
the liquid (in this case in water single phase) location.
and the pipeline correction factor (L, D and U
factor). D factor of the pipeline that less than c. The influence on the rate increment on the
0.9 also indicates constriction ID because of whole water injection system. The
scale or other impurities growth that contributes stimulation priority will be based on the
to high pressure drop in the pipeline. It affected highest rate increment.
to low injecting WHP and yielded low water
injected rate. It’s experienced by wells in the 2. Redistribution of water injection network.
cluster L that was caused by flowline to cluster
L has D factor value about 0.78. Manifold Integrated simulation can accommodate and
Pressure at Cluster L is only 421 psig (as shown forecast the results of redistribution of the water
in table 1). Beside it, based on API RP 14E injection in the system to:
(1991) recommended practice concerning
Design and Installation of offshore production a. Mitigate problems such as pipeline leak due
platform piping systems, as shown in Figure 5, to corrosion or sabotage.
the recommended velocity of liquid flow
through pipeline is between 3ft/second until 15 b. Optimize performance to get a higher
ft/second. Figure 6 shows an example of liquid injection rate and prevent bottlenecking.
velocity in the network that exceeds 15 ft/s and
yielded high pressure drop (8 psig per 10 ft). c. Re-route water injection flow due to surface
This simulation result provides justification for network sabotage by local inhabitants.
 Redistribution can be conducted by setting the e. Flow-line from manifold station until
open/closed valve that connects between the cluster R.
flow-lines. It also can be done by opening or
closing the valves that interconnect between The expected manifold pressure increment at
field stations. In Kaji-Semoga field, valve each cluster is shown in Table 1. The increment
connection between stations and fields can also water injection rate is expected about 20,000
be adjusted to optimize water injection of the BWPD.
whole system.
2. Additional water injection wells to reduce
RESULTS bottlenecking in the trunk-line. These wells are
located far from the field station. Three wells
In Kaji-Semoga field, the integrated model of water are already being proposed to be converted to
injection system is a very useful tool for evaluation, water injection wells.
optimization and troubleshooting in order to
increase water injection rate. Some 3. Installation of additional water injection pumps
recommendations for the water injection system to accommodate the future plan of water
optimization in Kaji-Semoga can be divided into injection rate due to increment of liquid lifting
short term plan and long term pan. The short term and water cut to combat oil production decline.
plan consists of:
After the short-term recommendations were applied,
1. Stimulation of wells that experiences formation for Kaji field station, as shown in Figure 8,
damage (can be observed by Hall-plot chart). In manifold system pressure has decreased from 1020
Kaji-Semoga, hydrochloric acid (HCl) acidizing psig to 980 psig (on average, based on trend line).
for stimulation is regularly conducted by using Average water injection rate has increased from
60 GPF (gallons per feet) acid. This is expected 65,000 BWPD to 70,000 BWPD and monthly VRR
to obtain a higher injectivity index to meet the for Kaji Field has increased from 0.911(August
required target of water injection. Figure 7 2012) to 1.048 (Jun 2013). Semoga Field monthly
shows the example of a Hall-plot trend in one VRR has increased from 0.818(August 2012) to
well in Kaji-Semoga field before and after 0.835 (Jun 2013) as shown in Figure 9. For the
stimulation. whole system at Kaji-Semoga field, liquid lifting
capacity has increased from 179,000 to 184,500
2. Redistribution of water injection flow to get the BLPD. The final objective to increase oil
maximum water injection rate of the system, by production by eliminating oil deferment 236 BOPD
adjusting the valve interconnection between due to water handling constraint has been achieved.
Kaji and Semoga field stations.
The long term plan consists of: The long term plan for upgrading the water
injection system in Kaji-Semoga field has not been
1. Installation of a new flow-line to re-route the applied yet. A proposal to the Indonesian
water injection flow in order to de-bottleneck Government has been issued and approved.
parts of the network (for some flow-lines that Currently, the upgrading is still under construction.
have liquid velocity exceed the maximum limit After the upgrade project has been completed, the
15 ft/s). Based on the model evaluation, it is capacity of the water injection system in Kaji-
recommended that additional flow-lines are Semoga field is expected to reach 220,000 BWPD.
required in these locations (as also described in
Figures 3 and 4):
CONCLUSIONS
a. Flow-line from jumper cluster AF until
1. Oilfield Water Injection Systems can be
jumper cluster Y.
modeled and simulated in Production Integrated
Model software as a tool to conduct system
b. Flow-line from manifold station until
optimization and troubleshooting.
jumper cluster L.

c. Flow-line from manifold station until 2. The water injection model includes
cluster D. performance data that cover subsurface
parameters (well and reservoir performance
d. Flow-line from manifold station until data) and surface parameters (pipe line and
jumper cluster U. water injection pump performance data).
3. Based on the water injection model, analysis The authors are also grateful to colleagues in P.T.
and evaluation can be performed, which have Medco Area Engineering Department Office that
resulted in optimization recommendations, were very kind and always supported completion of
including well stimulations and upgrading all of the jobs to finish this manuscript.
surface facilities equipment.
4. The results of water injection system REFERENCES
optimization based on water injection
integration model are: Aditama, P. et al., 2010, Construction and
implementation of integrated production Network
 Increase in the water injection rate yielding Model in a complex onshore operation: Proceedings
an increase in VRR value. of SPE OGIC, p. 2-3.
 The minimizing of oil deferment due to
maintain produced water tank levels. Malau, A. et al., 2011, Gas Field Production
 An increase in oil production rate. Optimization Using Integrated Network Modeling –
Case Study of SSE Gas Management: Proceedings
ACKNOWLEDGEMENTS of Indonesian Petroleum association, 35th Annual
Convention & Exhibition, p. 3-6.
The authors would like to acknowledge the support
of Rimau Asset management of PT. Medco E&P
Indonesia, particularly Petroleum Engineering Recommended Practice for Design and Installation
Rimau manager who has encouraged all of his of offshore production platform piping system, API
subordinates to gain new knowledge as deeply as Recommended Practice 14E (RP 14E), Fifth
possible. Edition. (1991).
 TABLE 1

EXPECTED MANIFOLD CLUSTER PRESSURE CONDITION BEFORE AND AFTER

ADDITION OF PROPOSED SURFACE FLOW-LINES

Additional Manifold Cluster Pressure (psig)

No Remarks
Line Proposed Before After
1 To Cluster Y 647 838 At Cluster Y
2 To Cluster A 814 1021 At Cluster A
3 To Cluster L 421 822 At cluster L
At cluster R, anticipating
4 To Cluster R 972 988
additional WIW
5 To Cluster V 847 922 At Cluster V
Figure 1 - Kaji-Semoga Field Production History.

Figure 2 - Kaji-Semoga WIW Count.

 New Proposed flow-line
from jumper cluster AF
until jumper cluster Y

New Proposed flow-line

from manifold station
until jumper cluster L

New Proposed flow-line

from manifold station
until jumper cluster D

Figure 3 – Schematic Diagram of Kaji WI Network system model.

New Proposed flow-line

from manifold station
until jumper cluster R

New Proposed flow-line

from manifold station
until jumper cluster U

Figure 4 - Schematic Diagram of Semoga WI Network system model.

 Figure 5 – Flow velocity in liquid lines (Reference: API RP 14 E (1991)).

Figure 6 – Example of liquid velocity of pipes in the Kaji-Semoga field network.

 Formation Damage
Indication

Rate increased
after acid 60 gpf

Figure 7 – Example Hall plot of a WIW in Kaji-Semoga Field.

Figure 8 - Kaji field station WI Performance History.

 VRR Kaji Jun ‘13: 1.048
VRR Kaji Aug ‘12: 0.911

VRR Semoga Jun ‘13: 0.835

VRR Semoga Aug ‘12: 0.818

Figure 9 - VRR History in Kaji-Semoga Field.