Reservoir Characterization by a Combination of Fuzzy Logic and Genetic Algorithm

摘要

Reservoir characterization is the basis of almost all aspects of well drilling and oilfield managements and therefore the precise determination of the reservoir properties has a massive profound impact on the success of the operations. Different methods have been suggested by previous authors for this purpose but complexity of the procedures, lack of precision, and time-consuming nature are the major weak points of such procedures, especially when uncored intervals are intended. In order to defeat such problems, a novel hybrid computational intelligence system is developed through which the optimum inputs (features) are investigated by a combination of genetic algorithm and adaptive network-based fuzzy inference system and then the cluster center's range of influence in each of the data dimension is optimized by the same algorithm. The resulting performances of the modeled cases are then compared by other methods. The considerable improvement in the system accuracies and achieving mean standard error of .00007 and .00033 for the porosity and water saturation models, substantiates the power, efficiency, and precision of the stated novel method.