Ensuring Water Saturation Consistency Between Static (Fine-Grid) and Dynamic (Upscaled) Models—A Case Study of the North Kuwait Jurassic Complex

Abstract

Abstract The North Kuwait Jurassic Complex (NKJC) consists of six fields with four identified reservoirs in the naturally fractured Jurassic carbonate section. An integrated model incorporating seismic, geological, petrophysical, and engineering data is built to estimate the hydrocarbon volumes and provide multiscenario production forecasts for the NKJC. Water saturation is populated in the static (fine-grid) model using a saturation height function (SHF), which is a continuous function of porosity and height above free-water level (HAFWL). The SHF is equivalent to a capillary pressure (at reservoir conditions), which is a function of porosity and water saturation. Capillary pressure curves based on the SHF are used to initialize water saturation in the dynamic (upscaled) model. Given the fact that reservoir simulators cannot handle capillary pressure curves as continuous functions of porosity, discretization of the porosity-dependent capillary pressure function was performed. For this purpose, 15 capillary pressure curves (corresponding to 15 porosity categories) were used. Populating initial water saturation in the dynamic model using the SHF (or the equivalent capillary pressure curves) results in a substantial discrepancy in terms of water saturation between the static model and the dynamic model. The discrepancy is mainly because of the nonlinear dependency of the SHF versus porosity and the averaging of HAFWL. Upscaling tends to eliminate the high and low porosity values in favor of the average porosity, which leads to substantial changes in the resulting water saturation. Because the calculated water saturation is highly dependent on the porosity and HAFWL, upscaling results in a mismatch between static and dynamic models in terms of hydrocarbon estimates. To overcome this problem, the initial water saturation in the simulation model is populated using the pore volume–weighted, upscaled water saturation from the static model. The endpoint-scaling functionality in the simulator is then used to modify the discrete set of capillary pressure curves to support the provided initial water saturation. Results show that the correction of the capillary pressure curves needed by the simulator to support the provided initial water saturation values is relatively minor while full consistency between the static and dynamic models in terms of water saturation is obtained.