Global Optimization Strategies for History Matching & Optimal Well Placement
The aim is to test and improve existing global optimization algorithms for applications to reservoir management problems. Algorithms to be tested and improved include but are not limited to genetic algorithm (GA), differential evolution (DE) particle swarm optimization (PSO), ant colony optimization (ACO), natural evolutionary strategies (NES), memetic algorithms, etc. Areas of Application are mainly in well placement (and rate) optimization (Fig. 3a) during water flooding and enhanced oil recovery, history matching of limited reservoir parameters and large-scale reservoir parameter estimation. Commercial simulators such as Eclipse are utilized for this research. In some cases, the need to use personal simulators tailored to solving specific flow problems may arise. Recently, our lab has developed an enhanced form of the PSO called the PSOAF. The PSOAF has shown better performance over existing algorithms such as DE, NES and the standard PSO (Fig. 3b) in the estimation of limited history-matched reservoir parameters.
Fig. 3a: Well placement in a reservoir.
Fig. 3b: Objective function during history matching of reservoir parameters