spe 163367-ms-p modelling of regional aquifer.....burgan field minagish reservoir, kuwait

15
SPE 163367 Modeling of Regional Aquifer System Allows Decision on Early Pressure Support to be Made for the Burgan Minagish Reservoir, Kuwait F.A. Al-Faresi, Kuwait Oil Company; J.T. Wang, W. Clark, M.L. Belobraydic, M. Yaser, F.O. Iwere, E. Gomez, O.M. Gurpinar, Schlumberger; K. Datta, A. Mudavakkat, L. Hayat, G.A. Al-Sahlan, R. Husain, A. Prakash, S.J. Crittenden, D.J. Bond, Kuwait Oil Company Copyright 2012, Society of Petroleum Engineers This paper was prepared for presentation at the SPE Kuwait International Petroleum Conference and Exhibition held in Kuwait City, Kuwait, 10–12 December 2012. 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 Burgan Minagish reservoir in the Greater Burgan Field is one of several reservoirs producing from the Minagish formation in Kuwait and the Divided Zone. The reservoir has been produced intermittently since the 1960s under natural depletion. A powered water-flood is currently being planned. The pressure performance of the reservoir has proved hard to explain without invoking communication with other reservoirs. Such communication could be either with other reservoirs through the regional aquifer of through faults to other reservoirs in the Greater Burgan field. Recent pressures are close to the bubble point. A coarse simulation model of the nearby fields and the regional aquifer was constructed based on data from the fields and regional geological understanding. This model could be history matched to allow all regional pressure data to be broadly matched, a result which supports the view that communication is through the regional aquifer. Using this model to predict future pressure performance suggested that injecting at rates that exceeded voidage replacement by about 50 Mbd could keep reservoir pressure above bubble point. It was recognized that the process of history matching performance was non-unique. This is a particular concern in the context of this study because the model inputs that were varied in the history matching process included aquifer data that was very poorly constrained. To address this problem multiple history matched models were created using an assisted history matching tool. Using prediction results from the range of models has increased our confidence that a modest degree of over-injection can help maintain reservoir pressure. This paper demonstrates the utility of computer assisted history match tools in allowing an assessment of uncertainty in a case where non-uniqueness was a particular problem. It also emphasizes the importance of understanding aquifer communication when relatively closely spaced fields are being developed. Introduction The Minagish Reservoir in the Burgan Field was discovered in the 1960’s. It has been producing intermittently since then. Production was shut down for a period (2005 - 2009) due to concerns about the H 2 S content of the produced fluids. Since 2005 there has been limited production to a gathering center in West Kuwait. It is planned to increase production significantly and to produce this to a newly built “sour service” gathering center. The reservoir pressure is determined to be currently at or slightly below bubble point of the fluid. It has been recognized that water injection would be needed maintain pressure and to support the increased off take. The pressure behavior of the reservoir over recent years was studied and was hard to explain. In particular there was pressure decline during periods with little or no production. Possible explanations for this included communication with nearby fields and cross-flow through the wells due to mechanical problems. The pressure has recently reached or dropped below the bubble point of the reservoir fluid. This gives the added urgency in the implementation of a pressure maintenance scheme. Early work focused on evaluating the scope for the pressure behavior being explained by cross flow. This involved determining the cross-flow volumes that would be required and reviewing the well data for indications of cross flow. After careful analyses, it was determined that cross-flow at the wells was not sufficient to account for the observed pressure trends. This prompted the need the work described here and the scope for aquifer communication was then evaluated. This was done in stages. Firstly, the information on the Minagish regional aquifer was reviewed. Special attention was given to the expected

Upload: stephen-crittenden

Post on 24-Jan-2017

17 views

Category:

Science


1 download

TRANSCRIPT

Page 1: Spe 163367-ms-p Modelling of regional aquifer.....Burgan Field Minagish Reservoir, Kuwait

SPE 163367

Modeling of Regional Aquifer System Allows Decision on Early Pressure Support to be Made for the Burgan Minagish Reservoir, Kuwait F.A. Al-Faresi, Kuwait Oil Company; J.T. Wang, W. Clark, M.L. Belobraydic, M. Yaser, F.O. Iwere, E. Gomez, O.M. Gurpinar, Schlumberger; K. Datta, A. Mudavakkat, L. Hayat, G.A. Al-Sahlan, R. Husain, A. Prakash, S.J. Crittenden, D.J. Bond, Kuwait Oil Company

Copyright 2012, Society of Petroleum Engineers This paper was prepared for presentation at the SPE Kuwait International Petroleum Conference and Exhibition held in Kuwait City, Kuwait, 10–12 December 2012. 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 Burgan Minagish reservoir in the Greater Burgan Field is one of several reservoirs producing from the Minagish formation in Kuwait and the Divided Zone. The reservoir has been produced intermittently since the 1960s under natural depletion. A powered water-flood is currently being planned. The pressure performance of the reservoir has proved hard to explain without invoking communication with other reservoirs. Such communication could be either with other reservoirs through the regional aquifer of through faults to other reservoirs in the Greater Burgan field. Recent pressures are close to the bubble point.

A coarse simulation model of the nearby fields and the regional aquifer was constructed based on data from the fields and regional geological understanding. This model could be history matched to allow all regional pressure data to be broadly matched, a result which supports the view that communication is through the regional aquifer. Using this model to predict future pressure performance suggested that injecting at rates that exceeded voidage replacement by about 50 Mbd could keep reservoir pressure above bubble point. It was recognized that the process of history matching performance was non-unique. This is a particular concern in the context of this study because the model inputs that were varied in the history matching process included aquifer data that was very poorly constrained. To address this problem multiple history matched models were created using an assisted history matching tool. Using prediction results from the range of models has increased our confidence that a modest degree of over-injection can help maintain reservoir pressure.

This paper demonstrates the utility of computer assisted history match tools in allowing an assessment of uncertainty in a case where non-uniqueness was a particular problem. It also emphasizes the importance of understanding aquifer communication when relatively closely spaced fields are being developed.

Introduction The Minagish Reservoir in the Burgan Field was discovered in the 1960’s. It has been producing intermittently since then. Production was shut down for a period (2005 - 2009) due to concerns about the H2S content of the produced fluids. Since 2005 there has been limited production to a gathering center in West Kuwait. It is planned to increase production significantly and to produce this to a newly built “sour service” gathering center. The reservoir pressure is determined to be currently at or slightly below bubble point of the fluid. It has been recognized that water injection would be needed maintain pressure and to support the increased off take.

The pressure behavior of the reservoir over recent years was studied and was hard to explain. In particular there was pressure decline during periods with little or no production. Possible explanations for this included communication with nearby fields and cross-flow through the wells due to mechanical problems. The pressure has recently reached or dropped below the bubble point of the reservoir fluid. This gives the added urgency in the implementation of a pressure maintenance scheme.

Early work focused on evaluating the scope for the pressure behavior being explained by cross flow. This involved determining the cross-flow volumes that would be required and reviewing the well data for indications of cross flow. After careful analyses, it was determined that cross-flow at the wells was not sufficient to account for the observed pressure trends. This prompted the need the work described here and the scope for aquifer communication was then evaluated. This was done in stages. Firstly, the information on the Minagish regional aquifer was reviewed. Special attention was given to the expected

Page 2: Spe 163367-ms-p Modelling of regional aquifer.....Burgan Field Minagish Reservoir, Kuwait

2 SPE 163367

reservoir properties between Burgan Minagish and nearby fields and on the scope for there being fault barriers between the fields. Based on this work, communication between the fields could not be ruled out. Secondly a model of the aquifer system was constructed to demonstrate that it was possible, very broadly, to simultaneously account for the pressure behaviors of the Burgan Minagish reservoir and nearby reservoirs with a model with “reasonable” aquifer properties. This involved calibrating the simulation model to the pressure data (history matching). It was recognized that the aquifer properties, and also the properties of possible tar mats, were not well defined by this process. Thirdly a series of matches to the data were produced. These were used to investigate the range of future pressure performance that would be expected and to help investigate the range of water injection rates that may be needed to support the planned production.

Inception of the study The Burgan Minagish reservoir pressure data were examined in 2010 for historical trends. The reservoir pressure in the last decade showed a significant decline, even when production from the reservoir was shut in for several prolonged periods of time. The pressure is approaching the bubble point and any further significant pressure decline was considered undesirable. Figure-1 gives a plot of pressure data (derived mainly from SIBHP surveys) and production.

A team was established within KOC to investigate the reasons for such an abnormal decline. This comprised members of the concerned production assets and staff from KOC’s exploration department. Two possible causes were investigated. Firstly cross flow to other reservoirs, and secondly, a reduction of aquifer support. Both reasons had to be investigated thoroughly to provide an effective diagnosis, in order for correct actions to take place.

Investigation of possible cross-flow As noted above, the pressure decline appeared to be inconsistent with the reported production. Additional productions, such as those caused by the flow behind pipe to other reservoirs, could help explain the anomaly. This was investigated first by estimating the level of such additional production that would be needed and then by assessing whether such cross-flow could be taking place in the wells.

The first task was approached by building a simple material balance model. The model was incorporated with average rock and fluid properties, a “deterministic” STOIIP and best estimates of blowout volumes (from the period immediately after the liberation of Kuwait in 1990). The model was calibrated to reservoir pressures up to 1990. This avoided periods of high production from nearby Minagish reservoirs and also the blow-out period. The results of this “tank” model can be found in Figure-2. This Figure shows historical and calculated average reservoir pressures and historical production up to the present time. We note an increasing gap between calculated and historical pressures starting from around year 2002. The discrepancy eventually builds up to approximately 150 psi.

A possible explanation of this discrepancy is cross-flow. The material balance model was re-run incorporating additional productions attributed to cross flow and the level of such production was adjusted to improve the match with pressure data. A good match was achieved by assuming 25 Mbd of cross-flow since 2002. The results of this model are shown in Figure-3.

The scope for cross flow in wells drilled to the Minagish or deeper horizons was reviewed. There was no evidence of cross flow from temperature surveys of the wells. The Minagish oil is distinct from the oil in the shallower producing horizons. There was no evidence of cross-flow to these horizons based on geochemical analyses of the oil produced from the shallower horizons.

Investigation of communication through the aquifer The other possible scenario to explain the rapid decline in reservoir pressure is offset production from other reservoirs in a common aquifer. A series of geological and geophysical tasks were carried out to determine whether this was plausible. These included:

(i) A review of scope for major faults disrupting flow at Minagish level. (ii) A review of major faults providing a means of inter-connection between the Minagish aquifer and other aquifer. (iii) Preparation of some simple maps of the most likely views of the Minagish reservoir thickness, porosity and

permeability based on the data from KOC producing fields and regional data. (iv) Evaluation of the uncertainty in the views of the regional aquifer. Based on these tasks between several fields through

a common aquifer seemed to be plausible.

Engineering data from the Burgan Minagish and nearby Minagish formation reservoirs were reviewed. A review of initial pressures from these fields was consistent with the view that they could be in communication. These data were, however, inaccurate; pressures were from mechanical pressure gauges. The pressure and production data from the various fields were also compared. These data were highly suggestive of communication through the aquifer. Figure-4 shows the production and pressure data from the Burgan Minagish reservoir and from a nearly reservoir, the Um Gadir Minagish. These data would suggest that the anomalous decrease in Burgan Minagish pressure could be related to the increase in Um Gadir Minagish production in 2002. A “proof of principle” simulation model, using single cells to model the Burgan and Minagish reservoir and a uniform rectangular aquifer, gave increased confidence that this explanation was the correct one.

Based on the finding, a decision was made to embark on a more comprehensive and realistic study of pressure

Page 3: Spe 163367-ms-p Modelling of regional aquifer.....Burgan Field Minagish Reservoir, Kuwait

SPE 163367 3

communication through the Minagish aquifer to help support the development planning work of the Burgan Minagish reservoir. In particular there was a desire to evaluate the requirements for water injection and to determine if water injection would be needed prior to the availability of the planned new production facilities The remainder of this paper describe the an outsourced study carried out to address these issues.

Static model A static model covering an area of approximately 100 km x 100 km (see Figure-5) was prepared for this purpose. The structural grid was assembled from multiple sources of data to create a realistic combined aquifer model for the Burgan, Umm Gudair (UG), Minagish (MN), and Wafra (WF) fields. The top Ahmadi structure (Carmen, 1996), was utilized as a trend surface set to the elevation of the Umm Gudair and Burgan fields for the north part of the study area and the Ratawi structure (Nelson, 1968) for the southern part of the study area (Figure 6). The final grid is 1000x1000 m, with 100 cells N-S and 100 cells E-W for a total coverage of 10,000 km2 as the model area. The structure of the top horizon, MN100, was used as the main structural horizon and the grid was completed with the available isochores of the subsequent layers. By including fifteen Minagish horizons, a final structural grid of 140,000 cells was created (Figure 77). Five facies are carried in the model: mudstone-marl, bioturbated mudstone-wackestone, pelletal packstone, oolitic grainstone, and pelletal mixed grainstone. These form the framework of the petrophysical model. Regional Aquifer Dynamic Model The regional aquifer model was primarily calibrated by historical observed static pressures with the use of a computer assisted history matching (CAHM) application. To this end, the properties of the oil in the individual fields are made consistent with the latest information related to the PVT data of the individual fields. On the other hand, matching historical fluid ratios (water cut and GOR) is of secondary importance as only pseudo wells were used to carry entire historical productions of individual fields. These wells are essentially long horizontal production and injection wells (Figure 88). In addition, significant tar mat layers have been observed from petrophysical logs acquired in BG and MN fields. These may act as significant barriers to reservoir fluid flows. In the model, partial flow barriers immediately beneath the oil zones of BG and MN fields were added to the model to mimic their effects. These represent the tar mats observed in these fields and due to the uncertainties involved, the extents to which they impeded flows are used as parameters for calibrating the model to production data. These parameters are in the form of transmissibility multipliers.

Although the model grid covers quite a large area (100 km x 100 km), pressure declines along its boundary are still inevitable throughout the simulated period of 64 years. An external analytical (Carter Tracy) aquifer was attached to the model as a way of approximating the extensive aquifer system. The way it was connected to the model was made a parameter during the history matching process and Figure 88 shows its final configuration in the base reservoir realization.

The model covers a vast area and many of the properties are subject to high level of uncertainty due to the very limited number of control points. Some of these quantities (e.g. aquifer rock properties between field areas) may never be measured and their values chosen for use in the simulation work are likely to be incorrect. The static model provides estimates of how these properties are distributed throughout the area. These are adjusted when the model is calibrated by using the historical production data as additional constraints.

During this process, multipliers of the two rock properties, namely, porosity and permeability, in the static model are adjusted separately within and outside each field area. In other words, porosity and permeability in the eight regions (Figure 99) were used as parameters. Porosity multipliers in the regions (1 to 4) covering field areas typically vary between 0.5 and 1.5 during the process whereas permeability multipliers vary between 0.2 and 5.0. The full list of parameters that have been adjusted within the range of uncertainty to calibrate the dynamic model:

Porosity multipliers Permeability multipliers Initial datum pressures in UG and MN to compensate for the structural uncertainties. Due to the lack of information,

the area around the WF field shares the same PVT data and datum pressure as BG. Oil-water contact depths in UG, MN and WF to compensate for the formation thickness and structural uncertainties Reservoir radius for adjusting the influx constant of the analytical aquifer Aquifer permeability for adjusting the time constant of the analytical aquifer Average MN production rate during the early 90s

Once the model was deemed satisfactorily matching field data (Figure 10), it was interrogated for information that is of interest. Figure 1111 shows pressure distributions at four specific times in the past. This Figure allows the general aquifer flow directions during the development of the fields to be inferred. It also shows pressure declines along the model boundary.

One of the primary objectives in the fast loop phase was to estimate the historical aquifer flow among the Minagish reservoirs in these fields. This will be an invaluable piece of information for the work in the Slow Loop phase. The interference each field imposed on neighboring fields was very much the result of their individual development activities. These activities (production, injection and the abnormal flow during the early 90s) led to fluctuations in the estimated aquifer flows. The exact magnitudes of these aquifer flows are also very much dependent on the locations of the boundaries for quantifying them. By setting two arbitrary boundaries, one midway between BG and UG and another one between BG and

Page 4: Spe 163367-ms-p Modelling of regional aquifer.....Burgan Field Minagish Reservoir, Kuwait

4 SPE 163367

WF, the model estimates history aquifer flow profiles as shown in

Figure 1212. The Figure shows a very significant outflow from the BG area considering the peak BG MN production (excluding the early 90s) is approximately 45,000 b/d. In fact, the production activities of other fields account for most of the pressure drop there.

The high level of uncertainty associated with a number of parameters (e.g. aquifer reservoir properties) makes the model

calibration process described above very non-unique indeed. In other words, it is very likely that multiple versions of the model possibly with very different input parameters also match the same set production and pressure data to similar degrees. The identification of these alternative realizations was achieved by continuing the calibration process with the CAHM application. This time, the BG area (region 1) porosity multiplier has been excluded in order to maintain a constant fluid volume there. Table 1 summarizes the parameters of the regional aquifer realizations identified following this process. The ‘Base’ case refers to the one described in the previous section. Using these combinations, Figure 1313 shows the pressures in the model compared to historical data.

CASE NAME

Units Base E263 E276 E281 E287 E313 E346 E354 E358 E359 E365 E367 E373 E376 E379 Min Max Av

Region 2 porosity multiplier 0.85 0.75 0.75 0.77 0.78 0.75 0.95 0.75 0.75 0.75 0.75 0.75 0.75 0.75 0.75 0.75 0.95 0.77

Region 3 porosity multiplier 1.00 1.10 1.10 1.10 1.10 1.10 1.10 1.10 1.10 1.10 1.10 1.10 1.10 1.00 0.93 0.93 1.10 1.07

Region 4 porosity multiplier 1.20 1.40 1.00 1.01 1.00 1.00 1.40 1.00 1.00 1.00 1.31 1.00 1.31 1.13 1.38 1.00 1.40 1.14

Region 5 porosity multiplier 0.50 0.61 0.30 0.44 0.69 0.30 0.44 0.30 0.48 0.30 0.59 0.39 0.58 0.42 0.30 0.30 0.69 0.44

Region 6 porosity multiplier 0.50 0.33 0.30 0.54 0.30 0.32 0.30 0.30 0.53 0.40 0.54 0.50 0.30 0.70 0.30 0.30 0.70 0.41

Region 7 porosity multiplier 0.10 0.20 0.18 0.20 0.12 0.18 0.20 0.20 0.13 0.20 0.09 0.19 0.12 0.16 0.20 0.09 0.20 0.17

Region 8 porosity multiplier 1.00 0.87 1.04 1.20 1.20 0.84 1.20 0.80 0.90 0.80 1.20 1.18 1.20 1.19 0.92 0.80 1.20 1.04

Region 1 perm multiplier 3.00 5.00 3.56 3.71 5.00 4.62 5.00 4.82 2.82 2.90 2.42 1.58 3.35 4.05 2.43 1.58 5.00 3.62

Region 2 perm multiplier 2.50 5.00 2.00 3.32 5.00 2.87 5.00 5.00 2.00 2.00 4.32 3.75 5.00 5.00 2.00 2.00 5.00 3.65

Region 3 perm multiplier 1.20 1.67 1.73 1.47 0.50 0.50 0.57 1.55 0.50 0.75 0.50 1.07 2.00 0.50 1.99 0.50 2.00 1.10

Region 4 perm multiplier 1.68 3.46 1.76 1.70 2.48 1.70 3.50 1.70 3.50 1.70 2.37 1.70 2.69 3.50 2.34 1.68 3.50 2.39

Region 5 perm multiplier 2.00 3.00 2.23 3.00 1.61 2.90 1.31 1.91 3.00 3.00 1.68 2.26 2.31 3.00 2.66 1.31 3.00 2.39

Region 6 perm multiplier 1.80 0.80 1.97 0.94 1.05 1.36 0.83 1.10 2.43 1.53 1.01 1.27 0.78 0.63 1.37 0.63 2.43 1.26

Region 7 perm multiplier 1.20 0.82 1.50 1.25 0.82 1.30 1.32 0.30 1.21 1.50 1.50 0.30 1.41 1.21 1.50 0.30 1.50 1.14

Region 8 perm multiplier 0.06 0.09 0.07 0.05 0.09 0.05 0.05 0.05 0.07 0.05 0.04 0.04 0.06 0.05 0.04 0.04 0.09 0.06

MN tar mat trans multiplier 0.003 0.002 0.003 0.001 0.035 0.036 1.000 0.001 0.305 0.024 0.040 0.013 0.002 0.371 0.002 0.001 1.000 0.142

BG tar mat trans multiplier 0.006 0.063 0.040 0.027 0.073 0.072 0.005 0.032 0.024 1.000 0.020 0.001 0.026 0.049 0.019 0.001 1.000 0.137

UG datum pres psia 4125 4019 4004 4131 4200 4020 4123 4000 4022 4000 4116 4000 4200 4200 4064 4000 4200 4082

MN datum pres psia 4647 4600 4847 4847 4847 4847 4847 4847 4847 4647 4847 4847 4847 4843 4606 4600 4847 4788

UG OWC depth ftSS 8450 8440 8462 8444 8440 8443 8440 8440 8454 8458 8440 8440 8440 8440 8454 8440 8462 8446

MN OWC depth ftSS 9780 9756 9750 9750 9786 9779 9791 9758 9783 9764 9777 9761 9750 9827 9776 9750 9827 9773

WF OWC depth ftSS 6470 6457 6521 6491 6492 6498 6452 6492 6495 6498 6460 6492 6458 6461 6456 6452 6521 6480

Aquifer int. radius ft 2E+7 4E+5 1E+8 4E+5 1E+8 9E+5 1E+8 3E+6 1E+6 2E+6 2E+6 1E+8 1E+8 1E+8 8E+5 4E+5 1E+8 4E+7

Aquifer perm md 600 800 800 800 658 313 420 399 310 271 418 599 725 109 350 109 800 505

MN early 90s rate mbo/d 400 410 200 210 410 410 410 410 410 290 410 410 263 410 410 200 410 364

Table 1 – Model input parameters in alternative regional aquifer realizations

The average BG MN pressures predicted by the model with these 15 realizations are shown in Figure 144. In all these cases, the pressure decline is predicted to continue for roughly another 10 years before beginning the flattening trend. As the pressure drops, the rate at which fluid leaves the Burgan area is slowed gradually and this resulted in the flattening trend. The Figure shows the estimated average BG MN pressure in 16 years to be estimated to be between 2,450 psi and 2,580 psi if the productions at all fields were to be maintained at their rates in 2011.

Obviously, the low resolution of the model grid casts doubt on performance prediction results obtained from the model. In

this case, simple scenarios are simulated to determine the impact of implementation of water injection and the relative location for the injection to take place. These cases (Table 2) are mostly for determining the effects of injection at a moderate rate (80 Mbwpd) and at a high rate (160 Mbwpd) on BG MN pressure.

Page 5: Spe 163367-ms-p Modelling of regional aquifer.....Burgan Field Minagish Reservoir, Kuwait

SPE 163367 5

Case

BG Injection BG Prod

1 80 Mbd above tar mat from 1/1/2014 Continue at current rate 2 80 Mbd above tar mat from 1/1/2014 Shut down BG MN production till 2015 3 160 Mbd above tar mat from 1/1/2014 Continue at current rate 4 160 Mbd above tar mat from 1/1/2014 Shut down BG MN production till 2015 4a 160 Mbd below tar mat from 1/1/2014 Shut down BG MN production till 2015

Table 2 – Simple prediction scenarios

In addition, the non-uniqueness of the model calibration process as illustrated above prompts the need for estimating the range of uncertainty associated with the predicted results in these cases. The one parameter that could potentially have a significant impact on the effect of water injection in BG MN is the extent of its tar mat that was observed near or at the oil-water contact at a number of wells. Given this, each of the development scenarios listed in Table 2 is tested with three model realizations (see Table 1 for their input parameters) with very different reductions in flow transmissibility (or transmissibility multiplier) across the tar mat

No reduction (E359 case) – no tar mat transmissibility multiplier Moderate reduction (Base case) – tar mat transmissibility multiplier at 0.006 Severe reduction (E367 case) – tar mat transmissibility multiplier at 0.001

It should be noted that the model is calibrated to the same set of observed historical pressure and production data in these

realizations. With these extreme transmissibility multiplier values, the other parameters are set to different values (listed in Table 1) as a result of the calibration process for the model to match the same observed data in a similar fashion.

Results of these simulations with the nine development scenarios are presented in Figure 155 to Figure 188. In each of these Figures, results of two of these cases are compared to those in the case in which all voidage production and injection rates are assumed to be maintained at their levels in September, 2011 (“do nothing” case). Figure 155 and Figure 166 show how the predicted average BG MN datum pressure in these cases and based on the three realizations described above. The following observations can be made from this Figure:

1. The effect of suspending BG production (until 2015) on the long term predicted pressure is relatively insignificant. This is because production activities in other neighboring fields have a more pronounced effect on BG MN reservoir pressure than even its own production.

2. As expected, the average BG MN pressure is very sensitive to the transmissibility across the tar mat when injecting water inside the tar mat

Table 3 shows a snapshot of predicted pressure values in 2024 in these cases. The estimated pressure increase values at that

time are also presented in Figure 177. Given this, the reservoir pressure should be closely monitored in any future water injection operations. This would be useful for reducing the uncertainty on how the tar mat affects the way the BG MN reservoir interacts with the surrounding aquifer. Figure 188 shows how the placement of the water injection well (inside or outside the tar mat) impacts the predicted average pressure in BG MN. As expected, the contrast is more prominent if the tar mat is relatively impervious to flow across it.

Scenario

Realization

No Tar Mat Mult = 0.006 Mult = 0.001 Average Pressure,

psia

Do nothing 2614 2597 2587

Injecting at 80 Mbwd 2722 2756 2970

Injecting at 160 Mbwd 2829 2892 3414 Increase,

psi Injecting at 80 Mbwd 107 159 383

Injecting at 160 Mbwd 214 296 827 Table 3 – Predicted average BG MN pressures and pressure increases in 2024

Page 6: Spe 163367-ms-p Modelling of regional aquifer.....Burgan Field Minagish Reservoir, Kuwait

6 SPE 163367

Conclusions A detailed study has been carried out to address the extent of pressure communication between the Burgan Minagish reservoir and other nearby reservoirs producing from the Minagish formation. This has shown that the pressure performance of the Burgan Minagish reservoir can be explained by communication through the aquifer.

The properties of the aquifer are relatively poorly known, as are the properties of Tar mats in the producing reservoirs. A range of models were conditioned to the available data. Although these all match the historical data they predicted future pressure performance differs considerably. This range of models has been used to estimate the timing and volume of the required water injection.

Acknowledgments The authors thank Kuwait Oil Company and Schlumberger for their permission to publish this paper. We also wish to thank KOC Manager West Kuwait, Mr. Hasan Bunain, and his staff for their support and advice.

References 1. Carman, G.J.: “Structural Elements of Onshore Kuwait”, presented at the 2nd Middle East Geosciences Conf. and

Exhib., GEO96, Bahrain, 15-17 April 1996.

2. Nelson, P.H.: “Wafra Field Kuwait – Saudi Arabia Neutral Zone”, Presented at the Regional Technical Symposium, Dhahran, Saudi Arabia, 27-29 March 1968.

Page 7: Spe 163367-ms-p Modelling of regional aquifer.....Burgan Field Minagish Reservoir, Kuwait

SPE 163367 7

Figures

Figure 1 – Pressure and production performance of Burgan Minagish reservoir

Figure 2 – Material Balance model results

Page 8: Spe 163367-ms-p Modelling of regional aquifer.....Burgan Field Minagish Reservoir, Kuwait

8 SPE 163367

Figure 3 – Material Balance results assuming extrap efflux

Figure 4 – Relation of Burgan Minagish and Umm Gadir production and pressures

Page 9: Spe 163367-ms-p Modelling of regional aquifer.....Burgan Field Minagish Reservoir, Kuwait

SPE 163367 9

Figure 5 – Study model area

Figure 6 – Adjusted Ahmadi (blue), Ratawi (pink), and MN100 structural contours (green) and MN100 tops (pink dots).

Page 10: Spe 163367-ms-p Modelling of regional aquifer.....Burgan Field Minagish Reservoir, Kuwait

10 SPE 163367

Figure 7 – Final structural grid demonstrating the zones (10x exaggeration)

Figure 8 – Tar mat boundaries and external aquifer connections

Page 11: Spe 163367-ms-p Modelling of regional aquifer.....Burgan Field Minagish Reservoir, Kuwait

SPE 163367 11

Figure 9 – Regions for porosity and permeability adjustments

Figure 10 – Pressure match of calibrated model

Page 12: Spe 163367-ms-p Modelling of regional aquifer.....Burgan Field Minagish Reservoir, Kuwait

12 SPE 163367

Figure 11 – Estimated historical regional aquifer pressure distributions

Figure 12 – Historical aquifer flow profiles estimated by the base case model

Page 13: Spe 163367-ms-p Modelling of regional aquifer.....Burgan Field Minagish Reservoir, Kuwait

SPE 163367 13

Figure 13 – Regional aquifer model pressure match (15 realizations)

Figure 14 – Predicted BG MN pressures in the continuing current operation scenario

Page 14: Spe 163367-ms-p Modelling of regional aquifer.....Burgan Field Minagish Reservoir, Kuwait

14 SPE 163367

Figure 15 – Case 1, Case 2 and “Do-nothing” case predicted pressures

Figure 16 – Case 3, Case 4 and “Do-nothing” case predicted pressures

Page 15: Spe 163367-ms-p Modelling of regional aquifer.....Burgan Field Minagish Reservoir, Kuwait

SPE 163367 15

Figure 17 – Predicted pressure increase (2024) in BG MN (w.r.t. do-nothing case) with water injection inside tar mat

Figure 18 – Case 4, Case 4A and “Do-nothing” case predicted pressures