Most of the Malaysian oil fields are mature and the need for IOR/EOR driven development planning is vital to maintain the national production. The most common EOR methods employed in Malaysia are immiscible WAG and CEOR. However, one of the main problems in WAG is how to control gas mobility to minimize the conformance issue and maximize the 3-phase region in the reservoir. Using Foam is a relatively cost effective way of controlling gas mobility by means of low concentration of surfactant. However, the practical implementation has been hindered due to the lack of general understanding of the complex process of foam phase behavior and flow dynamics in porous media, the absence of reliable predictive models of multi-phase foam flow, and also the challenges in upscaling the lab results to the field application.The data integration, fundamental understanding, input evaluation and predictive modeling of WAG and Foam Assisted WAG (FAWAG) have been reviewed in this work using two extensive core flooding data of one of the Malaysian oil field in two different simulators. Review of the laboratory procedure showed the importance of taking proper pre-design measures to be able to achieve conclusive results. The modeling practice shows that output parameters from coreflood experiments might be affected by forced history matching and consequently unrealistic simulation input parameters as well as the shortcomings of the predictive capability of different simulators.The study showed that in the core flood experiments and simulations, various aspects need to be carefully considered including:• Reliable data integration and understanding of the physics/dynamics involved in the process are important for proper evaluation of the foam effect. • Simulation studies to be used for the pre-design and optimization of the coreflood experiment and parameters (i.e., rate, cycles, pore volume, etc.). • The measured rates and pressures must be of consistent resolution to be able to capture the exact phase breakthroughs and precise fractional flows of the experiment. • The SCAL endpoints and WAG parameters must be measured and fixed prior to FAWAG experiment and simulation.• Capillary end effect, formation water salinity effect on contact angle and capillary pressure impact on coreflood results needs to be carefully considered in both laboratory and modeling stages. • Different commercial simulators may results in different prediction on foam output and the effect, hence proper evaluation, de-risking and sensitivity analysis should be performed to capture the foam EOR window of opportunity.

Most of the Malaysian oil fields are mature and the need for IOR/EOR driven development planning is vital to maintain the national production. The most common EOR methods employed in Malaysia are immiscible WAG and CEOR. However, one of the main problems in WAG is how to control gas mobility to minimize the conformance issue and maximize the 3-phase region in the reservoir. Using Foam is a relatively cost effective way of controlling gas mobility by means of low concentration of surfactant. However, the practical implementation has been hindered due to the lack of general understanding of the complex process of foam phase behavior and flow dynamics in porous media, the absence of reliable predictive models of multi-phase foam flow, and also the challenges in upscaling the lab results to the field application. The data integration, fundamental understanding, input evaluation and predictive modeling of WAG and Foam Assisted WAG (FAWAG) have been reviewed in this work using two extensive core flooding data of one of the Malaysian oil field in two different simulators. Review of the laboratory procedure showed the importance of taking proper pre-design measures to be able to achieve conclusive results. The modeling practice shows that output parameters from coreflood experiments might be affected by forced history matching and consequently unrealistic simulation input parameters as well as the shortcomings of the predictive capability of different simulators. The study showed that in the core flood experiments and simulations, various aspects need to be carefully considered including:Reliable data integration and understanding of the physics/dynamics involved in the process are important for proper evaluation of the foam effect.Simulation studies to be used for the pre-design and optimization of the coreflood experiment and parameters (i.e., rate, cycles, pore volume, etc.).The measured rates and pressures must be of consistent resolution to be able to capture the exact phase breakthroughs and precise fractional flows of the experiment.The SCAL endpoints and WAG parameters must be measured and fixed prior to FAWAG experiment and simulation.Capillary end effect, formation water salinity effect on contact angle and capillary pressure impact on coreflood results needs to be carefully considered in both laboratory and modeling stages.Different commercial simulators may results in different prediction on foam output and the effect, hence proper evaluation, de-risking and sensitivity analysis should be performed to capture the foam EOR window of opportunity.

No abstract

The objective of this paper is to address the challenges that are frequently encountered in simulation studies when using local grid refine (LGR) within upscaled models. The difficulties mainly arise due to the unreliability of populating the fine grids with reservoir properties and attributes. Dynamic modeling of a pilot is an important task to predict fluid flow and reservoir behavior which is a major step of pilot design. Dynamic models usually have many limitations when it comes to geological description due to the upscaling of fine-grid static model. Using local grid refinement (LGR) alone for the pilot area within a coarse dynamic model also would not enhance reservoir description of the pilot area without a realistic reservoir description. This work was aimed to provide an improved method for the proper simulation of pilot project in order to optimize the design of the pilot injector, borehole location and length. Furthermore, the model would be used to plan an efficient reservoir monitoring program including an optimized well data gathering with sponge coring for defining the remaining oil saturation. To overcome these limitations, the proposed method introduces a special fine scale LGR covering the pilot area within the upscaled dynamic model. Whereas the upscaled model has 36 layers, the LGR contains exactly the same fine-scale layering scheme and reservoir properties as the static model with 160 layers. Thus, it eliminates the upscaling process within the LGR. This process will ensure a better quality history match, for example the TDTs and RSTs derived saturations when compared against the fine layer LGR from the dynamic model. The dynamic model in this study is a large sector model (Nx 167, Ny 98 and Nz 36) with a detailed LGR (Nx 130, Ny 145 and Nz 160). The fluid properties were calculated from a 10-components equation of state (EOS). After the addition of the LGR, the model history match was updated. Several prediction cases were then studied to optimize well location (injector and observation wells) in order to convert the existing inverted five-spot gas pattern into water injection line-drive. Several saturation maps and profiles were generated to predict the breakthrough time for each observer and utilized to design the future pilot monitoring program. The new data will be utilized for future updates of the history match and performance of pilot under the optimum scheme of sweep and thus well spacing. Introduction This study simulates a giant carbonate reservoir (1200 sq. km) anticline structure the field is situated onshore in UAE. It has been on production for more than 45 years. The reservoir comprises a series on interbedded shallow marine bioclastic carbonate of Lower Cretaceous Thamama Group. The average thickness of the reservoir is about 50 meters (+160 feet) at the crest. The porosity range is 8% to 35%. The permeability range is 10 to 1000+ mD in the upper part of the reservoir and 1 to 10 mD in the lower part of the reservoir. Porosity and permeability decreases from crest to flank, good reservoir quality are related to depositional lithofacies and favorable diagenetic process. The reservoir contains a light oil of an average API of 37°. The main production mechanism is peripheral and mid-dip pattern water injection along with crestal gas injection in the gas cap and pattern gas injection in the north area of the field. The pilot is situated in one of the nine inverted five-spot patterns located in the north where gas injection commenced in 1997.

The objective of this paper is to address the challenges that are frequently encountered in simulation studies when using local grid refine (LGR) within upscaled models. The difficulties mainly arise due to the unreliability of populating the fine grids with reservoir properties and attributes. Dynamic modeling of a pilot is an important task to predict fluid flow and reservoir behavior which is a major step of pilot design. Dynamic models usually have many limitations when it comes to geological description due to the upscaling of fine-grid static model. Using local grid refinement (LGR) alone for the pilot area within a coarse dynamic model also would not enhance reservoir description of the pilot area without a realistic reservoir description. This work was aimed to provide an improved method for the proper simulation of pilot project in order to optimize the design of the pilot injector, borehole location and length. Furthermore, the model would be used to plan an efficient reservoir monitoring program including an optimized well data gathering with sponge coring for defining the remaining oil saturation. To overcome these limitations, the proposed method introduces a special fine scale LGR covering the pilot area within the upscaled dynamic model. Whereas the upscaled model has 36 layers, the LGR contains exactly the same fine-scale layering scheme and reservoir properties as the static model with 160 layers. Thus, it eliminates the upscaling process within the LGR. This process will ensure a better quality history match, for example the TDTs and RSTs derived saturations when compared against the fine layer LGR from the dynamic model. The dynamic model in this study is a large sector model (Nx 167, Ny 98 and Nz 36) with a detailed LGR (Nx 130, Ny 145 and Nz 160). The fluid properties were calculated from a 10-components equation of state (EOS). After the addition of the LGR, the model history match was updated. Several prediction cases were then studied to optimize well location (injector and observation wells) in order to convert the existing inverted five-spot gas pattern into water injection line-drive. Several saturation maps and profiles were generated to predict the breakthrough time for each observer and utilized to design the future pilot monitoring program. The new data will be utilized for future updates of the history match and performance of pilot under the optimum scheme of sweep and thus well spacing. Introduction This study simulates a giant carbonate reservoir (1200 sq. km) anticline structure the field is situated onshore in UAE. It has been on production for more than 45 years. The reservoir comprises a series on interbedded shallow marine bioclastic carbonate of Lower Cretaceous Thamama Group. The average thickness of the reservoir is about 50 meters (+160 feet) at the crest. The porosity range is 8% to 35%. The permeability range is 10 to 1000+ mD in the upper part of the reservoir and 1 to 10 mD in the lower part of the reservoir. Porosity and permeability decreases from crest to flank, good reservoir quality are related to depositional lithofacies and favorable diagenetic process. The reservoir contains a light oil of an average API of 37°. The main production mechanism is peripheral and mid-dip pattern water injection along with crestal gas injection in the gas cap and pattern gas injection in the north area of the field. The pilot is situated in one of the nine inverted five-spot patterns located in the north where gas injection commenced in 1997.