SPE 90362 Use of Material Balance to Enhance 3D Reservoir Simulation: A Case Study

Eugene Esor, SPE, Nigerian Agip Oil Co., Stefano Dresda, SPE, and Carlo Monico, ENI E & P Division, Italy

Copyright 2004, Society of Petroleum Engineers Inc. This paper was prepared for presentation at the SPE Annual Technical Conference and Exhibition held in Houston, Texas, U.S.A., 2629 September 2004. This paper was selected for presentation by an SPE Program Committee following review of information contained in a proposal submitted by the author(s). Contents of the paper, as presented, have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material, as presented, does not necessarily reflect any position of the Society of Petroleum Engineers, its officers, or members. Papers presented at SPE meetings are subject to publication review by Editorial Committees of the Society of Petroleum Engineers. Electronic reproduction, distribution, or storage of any part of this paper for commercial purposes without the written consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to a proposal of not more than 300 words; illustrations may not be copied. The proposal must contain conspicuous acknowledgment of where and by whom the paper was presented. Write Librarian, SPE, P.O. Box 833836, Richardson, TX 75083-3836, U.S.A., fax 01-972-952-9435.

compartmentalisation in 2 of the reservoirs; communication across a fault in 1 reservoir; and homogeneity in the remaining 6 reservoirs. This information was applied to the 3D simulation models by assigning appropriate transmissibility values across the relevant faults. The results also established the need to adjust Hydrocarbon contact in one of the reservoirs. Furthermore, the established historical drive mechanisms helped in obtaining faster and more reliable 3D history match. Background The material balance evaluation was carried out in order to provide input for proper 3D initialisation and history match of the main reservoirs in a broader study. The primary benefits consist in narrowing down uncertainties regarding reservoir connectivity and associated hydrocarbon volumes, hydrocarbon water contact and reservoir drive mechanism. The study was planned for a group of fields with a total of 53 reservoirs. The study has, as part of its objectives, the establishment of hydrocarbon volumes in the affected fields through 3D numerical simulation of the main reservoirs. This implies, by definition, that the oil reservoirs with existing STOIIP > 30 MMSTB or gas reservoirs with existing GIIP >115 Bscf were to be simulated using a 3D numerical simulation software in order to re-establish hydrocarbon initially in place volumes and reserves under various forecast scenarios. The screening results of all 53 reservoirs indicated that 12 of them fall within the category of major reservoirs as defined above. Out of these, 9 had adequate pressure and historical production data with most of the reservoirs having production history of about 30 years each. The existence of structural, fluid contact and drive mechanism uncertainties in the major reservoirs served as primary motivating factors to carry out the MBEs, the results from which is to guide in the construction of more realistic 3D numerical simulation models. Specifically, the uncertainties to be addressed included: connected hydrocarbon volumes; reservoir compartmentalisation; likely hydrocarbon water contact in hydrocarbon-down-to situations in which there are no RFT measurements; and reservoir drive mechanisms. These uncertainties have significant impact on established hydrocarbon volumes and reserves.

Abstract Material balance evaluation (MBE) has long been identified as a useful means of establishing connected hydrocarbon volume initially in place and reservoir drive mechanism1,2. Analysis of MBE results can also prove invaluable in addressing questions on reservoir compartmentalisation and fluid contacts. This paper demonstrates the application of MBE in narrowing down the above uncertainties prior to a full 3D numerical simulation study. The scope covers 9 reservoirs in the Niger delta area of Nigeria as part of a broader study. The aims of the MBE were to guide the construction of more realistic 3D simulation models by establishing connected hydrocarbon volumes and addressing uncertainties in reservoir compartmentalisation and hydrocarbon contacts. Furthermore, the MBE was aimed at facilitating more efficient history matching by establishing the reservoir drive mechanisms beforehand. The candidate reservoirs, which range from undersaturated oil to gas condensate with varying aquifer strengths, have long history of production and adequate pressure measurements, making them ideal candidates for reliable MBE. The MBALTM software was used to carry out the MBE. A combination of the analytical non-linear regression and the linear (P/Z, Havlena-Odeh etc) methods was used in establishing the correct material balance model, thereby giving confidence in the obtained results. The paper presents the study background, introduction, methodology, results, conclusions and recommendations. Results from the MBE showed the connected hydrocarbon volumes for all the reservoirs and established

www.petroman.ir

SPE 90362

Introduction Material balance has long been established as an interpretative and diagnostic tool for establishing connected hydrocarbon volumes originally in place and reservoir drive mechanisms. The application of this technique prior to a full 3D numerical simulation can provide valuable insight that would guide in the proper initialisation, history matching and production forecast of a 3D numerical simulation study thereby increasing confidence in the obtained results. The choice of carrying out MBE prior to the 3D numerical simulation was necessitated by the following considerations: Direct construction of 3D numerical simulation models is associated with lots of uncertainties related to reservoir geometry (from geological maps); compart-mentalisation and petrophysics. These translate to uncertainties in the established hydrocarbon volumes and their recovery. MBE, however, is based purely on the relationship between production, pressure and PVT data as regards the law of conservation of mass. This implies that results obtained from MBE using adequate and reliable production, pressure and PVT data are less prone to these uncertainties. Direct construction of 3D numerical simulation models does not capture the relative energy contributions from the gas cap, aquifer, fluid expansion, pore volume compressibility and gas injection. Without prior material balance evaluation, the order of magnitude of these various reservoir energy contributors has to be inferred by applying rule of thumb or at best some personal judgement. The implication of this is that more time is spent during the history matching phase as there is no prior quantitative knowledge of the effect of the changes in the various reservoir parameters on the history match. Prior MBE solves this problem by establishing the historical drive indices of the reservoir thereby presenting a quantitative picture of the energy sources of the reservoir recovery process and also giving direction in the history matching process. The Material Balance Evaluation Methodology. The MBALTM material balance tool was used for the evaluation. The adopted general workflow is shown in Fig. 1 below and consists of the reservoir type definition, data preparations/consistency checks, history matching/simulation and results analyses/ documentation.
Data preparation & consistency checks History matching & simulation

the reservoir can be defined as oil, gas or retrograde condensate. In the first step, the fluid type for each reservoir was defined by the integration of available engineering and petrophysical information. Data Preparation/Consistency Checks. Preparation of data and consistency checks were performed on the PVT, production history, average reservoir pressure, and reservoir and aquifer data. PVT Data. After a validation process, the 9 reservoirs were confirmed to have valid PVT study reports suggesting the suitability of the data for the evaluation. For the black oil reservoirs without flash PVT experiments, the oil formation volume factor (Bo) and solution gas oil ratio (Rs) from differential experiments had to be converted to flash data which is the form required by the MBALTM for PVT modelling. The following equations have been adopted for this conversion where necessary: Oil Formation Volume Factor:

Bo = Bod

Bobf Bobd

.......(1)

Where, Bo = Flash Oil Formation Volume Factor (FVF) at each stage of pressure depletion (rb/stb) Bod = Differential Oil FVF at each stage of pressure depletion (rb/stb) Bobf = Flash FVF of bubble point oil (rb/stb)

Bobd = Differential FVF of bubble point oil (rb/stb)

Solution Gas Oil Ratio:

Bobf Rs = Rsf (Rsbd Rsd ) ..(2) Bobd

Where Rs = Flash Solution GOR at each stage of pressure depletion (scf/stb) Rsf = Flash Solution GOR of bubble point oil (scf/stb)

Rsbd = Differential Solution GOR of bubble point oil (scf/stb) Rsd = Differential Solution GOR at each pressure stage
(scf/stb) The input PVT data were matched before being used for the evaluation. Production Data. About 30 years of production data each was available for most of the reservoirs. Monthly historical cumulative production and injection data were exported from the OFMTM production database and reformatted with the correct units before importing into MBALTM.

Output to 3D Simulation

Result analysis & documentation

Fig. 1: General workflow of the MBE Reservoir type Definition. The MBALTM tool allows the definition of reservoir type in terms of fluid content. That is,

www.petroman.ir

SPE 90362

Pressure Data. Each reservoir had adequate and distributed pressure measurements. For each reservoir, validity checks were made on the average reservoir pressure data by plotting these on Excel spreadsheet. Invalid pressure points were excluded. In situations where more than one valid pressure record existed within a month, an arithmetic average was calculated and this was recorded against the corresponding monthly average production data. Reservoir and Aquifer Data. Reservoir data such as temperature, initial pressure, porosity and connate water saturation were obtained from existing records. Aquifer properties were estimated based on known reservoir properties and evidence from geological maps. Because of the higher level of uncertainty attached to the aquifer parameters, these were more adjusted during the history match phase. History Matching. Two methods were utilized in carrying out the history match namely: Analytical method Graphical method In the analytical method, non-linear regression was used to estimate unknown reservoir and aquifer parameters during the history match. The method is graphically interactive and is based on fitting a model to the trend of pressure vs cumulative production of the predominant reservoir fluid as changes are made to the input parameters. A typical analytical history match plot for one of the reservoirs (G) is shown in Fig. 2 below. The points represent historical data whereas the line represents the simulated trend.

different combinations of hydrocarbon volume and aquifer model. Verification of the right model is achieved through the graphical method. Here, the results of the analytical approach are visualised with the corresponding straight-line method like the P/Z (for gas reservoir), the Havlena-Odeh (for oil reservoirs), etc. The graphical method has a synchronized output with the analytical method. The analytical model is considered adequate if the graphical method produces a good match which guarantees that the hydrocarbon volumes and aquifer strength applied are correct. A typical graphical history match plot for one of the reservoirs (G) is shown in Fig. 3 below. The blue squares represent values of the plotted parameters at time steps where there are pressure measurements whereas the associated numbers represent the time step in months at which the plotted parameters are calculated.

Fig. 3: Graphical history match plot for reservoir G History Match Simulation. After obtaining a history match, the validity of the match was established by running a simulation with the final material balance model. The results obtained from the simulation were compared with the historical input data of pressure, cumulative oil, gas and water production, and finally cumulative gas injection (for one of the reservoirs with gas injection history). Connected Hydrocarbon Volume and Reservoir Compartmentalisation. The connected hydrocarbon volume was established after obtaining the right reservoir model. The established volumes are those of the predominant hydrocarbon in the reservoir i.e. original oil in place for oil reservoirs and original gas in place for gas condensate reservoirs. This parameter was used to investigate reservoir compartmentalisation in 3 of the studied reservoirs namely B, C and H, where the status of some faults posed major communication uncertainties. In two of the reservoirs (B and H), the completed and produced wells existed in the northern fault block, whereas the southern fault block, though hydrocarbon-bearing, had no completed or produced well. As shown in Figure 4 below, reservoir B has one well completed in the northern block.

Fig. 2: Analytical history match plot for reservoir G

A match obtained through the analytical method alone does not guarantee that the right model has been used especially where there is some aquifer or gas cap support or both. This is mainly due to the compensatory nature of the effects of hydrocarbon volume and aquifer strength/size on the historical pressure trend of the reservoir. That is, based on the analytical method alone, it is possible to obtain history matches at

www.petroman.ir

SPE 90362

Through the MBE, it was established that the fault separating the north from the south is extended and sealing since the connected hydrocarbon volume from material balance is about one-third the volume obtained from volumetric evaluation of the entire structure. A similar result was obtained for reservoir H.
Completed and producing well

The larger gas volume from the material balance of block A was further supported by the fact that at the end of production history, about 16.5 Bscf of gas had already been produced from both wells 1 & 5 (i.e. 94% recovery based on volumetrics of block A) and currently well 5 is still producing over 2.5 MMscf/d of gas. Moreover, the pressure trends for both blocks do not show any disparity in depletion rates further highlighting the existence of communication between blocks A and B. Another MBE was therefore carried out considering the entire reservoir as a single tank in communication. The new model gave a gas initially in place volume for the entire structure. By lowering the contact in block A it was possible to obtain a new gas condensate-water contact that gives the volume obtained from material balance. Utilizing the result from material balance, the appropriate 3D initialisation was performed on both blocks. Reservoir Drive Mechanism. The reservoir drive mechanism is established after obtaining the right reservoir model through the history match procedure stated above. A graphical representation of the historical fractional contributions of energy from the gas cap, aquifer, fluid expansion, pore volume compressibility and gas injection are generated. The plot of historical drive mechanism for one of the reservoirs (G) with four primary energy sources is shown in Fig. 6 below:
Water Influx PV Compressibility

Fault confirmed extended and sealing

Fig. 4: Top structure map of reservoir B showing the fault confirmed sealing from the MBE Fluid Contact Evaluation. In reservoir C, both reservoir compartmentalisation and possible hydrocarbon-water contact were to be addressed. Reservoir C is a gas condensate reservoir divided into two blocks A and B by a fault. See structural map in Fig. 5 below:
BLOCK A Wells-1 & 5 GCDT = 2626.1 mssl BLOCK B Well-2 GCWC = 2651.1 mssl

Non-sealing fault confirmed from MBE Fig. 5: Top structure map of reservoir C Wells 1 and 5 are completed in block A with an original GCDT seen by well 1 at 2626.1 mss (meters sub-sea). Well 2, however, is completed in block B and saw an original gas GCWC at 2651.1 mss. Two uncertainties were identified in this reservoir namely the situation of the fault (i.e. whether it is sealing or nonsealing) and the position of the gas condensate water contact in block A. In order to address these uncertainties, MBE was used to investigate the volume of hydrocarbon seen by wells 1 & 5 in block A. A tank model was constructed considering only production from these two wells. The result showed that the hydrocarbon volume in communication with the wells 1 and 5 in block A is much more than is suggested from the geometry of the fault block and the existing GCDT. A hydrocarbon volume of 40.2 Bscf was obtained from material balance against a volume of 17.5 Bscf obtained from volumetrics using the GCDT as seen by well 1. This result suggested two things: that the fault is communicating and that the GCDT should be lowered.
Gas Cap Expansion

Fluid Expansion

Fig. 6: Typical plot of historical drive mechanism (reservoir G) In order to provide a quick knowledge of the historical drive indices for each reservoir, an average value was calculated with which a plot was generated for all the reservoirs. The plot of the average historical drive indices for all the reservoirs is shown in Fig. 7 below. The plot shows that reservoirs A, B, D, H and I are predominantly driven by fluid expansion whereas reservoirs C and G are predominantly driven by water influx and gas cap expansion respectively.

www.petroman.ir

SPE 90362

Fig. 7: Average % historical drive indices for the studied reservoirs. The usefulness of fig. 7 is that it presents a historical summary of the drive indices for each reservoir thereby giving a quantitative idea of the relative contributions of the various energy sources to the recovery process. This conversely means that one is able to identify beforehand which parameters would have the greatest effect on the history match of a given reservoir. For instance, adjusting the aquifer parameters for reservoir C would be more effective in obtaining the desired history match. Results. The results obtained from this evaluation adopting the above methodology established key parameters that gave a better picture of the reservoirs which guided the study team in constructing more realistic 3D numerical simulation models and effectively carrying out history matching on each. The key parameters established include: Connected hydrocarbon volume Reservoir compartmentalisation Possible hydrocarbon-water contact Historical reservoir drive mechanism From the connected hydrocarbon volume, the presence or absence of reservoir compartmentalisation was deduced. This was particularly important for 3 of the studied reservoirs where the existence of some faults presented major communication uncertainties. The historical drive mechanism quantitatively established each reservoirs energy comportment and this guided and also facilitated more effective history matching during the 3D numerical simulation. Conclusions. The following conclusions are drawn from the material balance evaluation: This Study. Material balance evaluation has been performed for 9 reservoirs and the results integrated into the work process of a full 3D reservoir simulation study. Connected hydrocarbon volumes and drive mechanisms have been established through material balance evaluation for the 9 reservoirs planned for 3D numerical simulation in the study. Owing to the availability of long production history, adequate and reliable PVT and pressure data, the volumes established are indications of the original hydrocarbon in

place for those reservoirs with good connectivity namely A, C, D, E, F, G and I. Exceptions are reservoirs B and H in which the separating faults are confirmed sealing. Communication has been established to occur between blocks A and B of reservoir C. Hence the fault separating these blocks is non-sealing. The GCDT seen in block A of reservoir C can be lowered to a possible gas condensate- water contact based on the new established volume from the MBE. No communications exist between the northern and southern blocks of reservoirs B and H. Fluid expansion is the predominant drive mechanism for 5 of the 9 studied reservoirs namely A, B, D, H and I. For reservoirs C and G, water influx and gas cap expansion are respectively the predominant drive mechanisms. Reservoirs E and G, however have more distribution of drive mechanism.

General. Material Balance Evaluation (MBE) is a handy tool for quickly defining reservoir drive mechanism, possible hydrocarbon-water contact, hydrocarbon in place and connectivity for reservoirs with adequate and reliable production, pressure and PVT data. The results from an MBE can be integrated into the work process of a full 3D reservoir simulation study leading to more realistic initialisation, faster history match and, consequently, more representative production forecast. Recommendations. The following recommendations are made: This Study. During the 3D initialisation, the fault separating the eastern and western parts of reservoir C should be kept partially open. However the faults separating reservoirs B and H into northern and southern blocks should be closed. The volumes established for reservoirs A, C, D, E, F, G and I should be used as a check during the 3D initialisation. In case of major disparity in volumes, a review of the 3D initialisation parameters should be carried out. The GCDT seen in block A of reservoir C should be lowered to a possible gas condensate-water contact that gives the same hydrocarbon volume obtained from material balance. During the 3D history match, more attention should be paid to the parameters affecting the predominant drive mechanism for each reservoir as changes in those parameters would be more effective in obtaining a history match. General. It is recommended to carry out MBE prior to any 3D simulation study provided there is adequate and reliable pressure, production and PVT data. The result from the MBE should be intregrated into the 3D model for better simulation.

www.petroman.ir

SPE 90362

Acknowledgements The authors wish to express their profound appreciation to the management of the Nigerian Agip Oil Company Ltd. for granting the permission to publish this paper. Special thanks also go to the entire study team and staff of Eni E & P, Milan for their co-operation and contributions. Nomenclature Bscf = Billion standard cubic feet E&P = Exploration & Production FVF = Formation Volume Factor GCDT = Gas condensate-down-to GCWC = Gas condensate water contact GIIP = Gas Initially In Place MBE = Material Balance Evaluation MMscf = Million standard cubic feet

mssl PVT rb RFT Rs Scf stb STOIIP 3D

= Meters sub-sea level = Pressure-Volume-Tempearture = Reservoir barrel = Repeat Formation Tester = Solution gas/oil ratio = Standard cubic feet = Stock tank barrels = Stock Tank Oil Initially In Place = Three-Dimensional

References
1. Dake, L.P.: The Practice of Reservoir Engineering, Revised Edition, Elsevier, Amsterdam, (2001) 73. 2. Craft, B.C. and Hawkins, M.: Applied Petroleum Reservoir Engineering, Second Edition, Prentice-Hall Inc., Englewood Cliffs, NJ (1991) 64.

www.petroman.ir