Applying soft computing methods to improve the computational tractability of a subsurface simulation-optimization problem

Formal optimization strategies normally evaluate hundreds or even thousands of scenarios in the course of searching for the optimal solution to a give n management question. This process is extremely time-consuming when numeri c simulators of the subsurface are used to predict the efficacy of a scenar io. One solution is to train artificial neural networks (ANNs) to stand in for the simulator during the course of searches directed by some optimizati on technique such as the genetic algorithm (GA) or simulated annealing (SA) . The networks are trained from a representative sample of simulations, whi ch forms a re-useable knowledge base of information for addressing many dif ferent management questions. These concepts were applied to a water flood project at BP's Pompano Field. The management problem was to locate the combination of 1-4 injection loca tions that would maximize Pompano's simple net profit over the next 7 years . Using a standard industry reservoir simulator, a knowledge base of 550 si mulations sampling different combinations of 25 potential injection locatio ns was created. The knowledge base was first queried to answer questions co ncerning optimal scenarios for maximizing simple net profit over 3 and 7 ye ars. The answers indicated that a considerable increase in profits might be achieved by expanding from an approach to injection depending solely on co nverting existing producers to one involving the drilling of three to four new injectors, despite the increased capital expenses. Improved answers were obtained when the knowledge base was used as a source of examples for training and testing ANNs. ANNs were trained to predict pe ak injection volumes and volumes of produced oil and gas at 3 and 7 years a fter the commencement of injection. The rapid estimates of these quantities provided by the ANNs were fed into net profit calculations, which in turn were used by a GA to evaluate the effectiveness of different well-field sce narios. The expanded space of solutions explored by the OA contained new sc enarios that exceeded the net profits of the best scenarios found by simply querying the knowledge base. To evaluate the impact of prediction errors on the quality of solutions. th e best scenarios obtained in searches where ANNs predicted oil and pas prod uction were compared with the best scenarios found when the reservoir simul ator itself generated those predictions during the course of search. Despit e the several thousand CPU hours required to complete the simulator-based s earches, the resulting best scenarios failed to match the best scenarios un covered by the ANN-based searches. Lastly, results obtained from ANN-based searches directed by the GA were compared with ANN-based searches employing an SA algorithm. The best scenarios generated by both search techniques we re virtually identical.