| Heavy oil and bitumen recovery processes need to be optimized in order to increase the recovery, reduce costs, and minimize the environment impact. Most of the optimization studies published in petroleum engineering literature focus on a few design parameters by combining the elements of numerical flow simulation with graphical or analytical techniques. Limited efforts, particularly in the areas of enhanced heavy oil recovery design, combine global optimization techniques with flow simulation to achieve better performance and design. The challenge remains because of high computational costs and slow convergence efficiency of the algorithms. In this research, genetic algorithm and simulated annealing are considered first as a single optimization technique. Then, the hybridization of these with the orthogonal arrays and response surface proxy techniques are tested. Savings up to 85% on the execution time are obtained for steam and solvent applications in oilsands and fractured carbonates. |
