This paper presents an optimization system for the development of petroleum fields. Developing a petroleum field consists of choosing an alternative exploitation of an already known and delimited petroleum reservoir allowing the maximum hydrocarbon production within the physical and economical limitations i.e., maximizing the net present value (NPV). The net present value is calculated according to the oil production, which is obtained with the use of a reservoir simulator. Each reservoir simulation can take from few seconds to several hours, depending on the complexity of the reservoir being modeled. This reduces the total number of configurations that can be generated and evaluated by the user in search for the best solution. Therefore, this work proposes and evaluates a new intelligent, optimization system that employs genetic algorithms (GA), cultural algorithms (CA), and co-evolution in order to search for an optimal development alternative in a parallel computing environment for reservoir simulations and NPV calculation. The proposed system provides the user, in a reasonable time, with the optimum (or sub-optimum) configuration for the development of the petroleum field. The results obtained in the case studies demonstrate that the proposed system, based on intelligent techniques, enable good configurations for the development of petroleum fields with a great reduction in computational time. This reduction is obtained from the computational power of the parallel computing environment and from the expert knowledge, through the initial configuration of the optimizing system (initial seed).
Genetic algorithms; Cultural algorithms; Co-evolution; Optimization; Petroleum reservoir development; Reservoir engineering; Distributed genetic algorithm; Parallel computing
