This paper presents a field case study describing how 3-D geological data are used in reservoir simulation to improve performance forecast and help optimize the development of the Serang Field, East Kalimantan, Indonesia.
The Serang field consists of complex stacked fluvial and deltaic channel sands under a large gas cap and over a strong aquifer. Development of the field is a challenge with the reservoir heterogeneity nature of the sands and the dynamic movement of fluid flow. Reservoir management is even more difficult as the originally thick oil columns have now become thinner with long-term production as adding infill wells or planning workovers require more accuracy for target locations. The study demonstrates the workflow involved in integrating the 3D geological data for constructing a reservoir model for field management. The model constructed helps explain the reservoir fluid flow behavior, identify by-passed oil under reservoir heterogeneity, plan for infill locations and workovers and long-term development strategy of the field. It shows the key benefits include a faster work process of building a reliable model that accounts for crucial heterogeneity, flexibility in varying reservoir properties with new data, and quicker history matching of the production data. These allow for quick turn-around to evaluate multiple development plans while improving production forecast.
The paper also discusses the practical and operational use of the model results in planning, implementing and operating wells for optimizing oil and gas recoveries.
The case study presented in this paper is from the Serang Field located approximately 16 km offshore East Kalimantan in 325-ft water (Figure 1). The field, composed of complex, stacked channel sandstones deposited in the ancient Mahakam Delta of the Kutei Basin, contains a large gas cap, thick oil column and strong surrounding aquifer. As a result of production from continuous drillings and workovers since late 1993, fluid distribution in the reservoirs has changed significantly from the initial stage and has become extremely dynamic even during a short-term production period. Specific issues and challenges important for managing and optimizing the field development are (i) locating current fluid contacts for optimal placement and timing of new wells, (ii) selecting proper completion techniques, (iii) identifying potential workover candidates, (iv) anticipating fluid movement and (v) predicting the remaining reserves. Reservoir models have been used as a monitoring and planning tool in the field life1–2 from the beginning to address these issues. As the field grows mature, their use becomes more critical as the remaining oil column becomes thinner and gas/water coning turns severe2.
The conventional reservoir models2 are constructed based on 2-D maps (Geographix or Z-Map) and then exported as Ascii data into a commercial simulator (DT-VIP3). Relevant maps include top/bottom of structure, gross/net sands, and occasionally manually-contoured geological properties. Afterwards, the fluid-flow simulation grids are inserted on these maps using gridding modules available in the simulator. Through this process, some editing and converting effort is required using special script files since exported 2-D data are not always directly readable by the simulator. From the reservoir constructing standpoint, this workflow is simple but inflexible, and at times requires simplifying the reservoir heterogeneity as it is time consuming to hand-contour reservoir properties.