An Efficient Methodology for Performance Optimization and Uncertainty Analysis in a CO2 EOR Process

A pragmatic technique is proposed and successfully applied to determine the optimal production-injection scheme in a CO2 flooding reservoir under uncertainty. Well rates of injectors and bottomhole pressures of producers are chosen as the controlling variables. Geological uncertainty is accounted for using the multiple reservoir models. An objective function associated with net present value (NPV) is defined, and a modified genetic algorithm is employed to determine the optimal production-injection scheme. It is shown from a field case study that the optimized scheme can not only increase the expected oil recovery and NPV by 7.8 and 6.6%, respectively, but can also achieve a considerably small range of possible NPVs.