Integrating Neural, Fuzzy Logic, and Nero-fuzzy Approaches Implementing Ant Colony Optimization Routing Algorithm to Determine Reservoir Facies

Determining the reservoir facies and areas with high-quality reservoirs play a pivotal role in reservoir modeling as well as future drilling in developing oilfields. As an index which varies in line with changes in the reservoir characteristics, Flow Zone Indicator (FZI) could be an influential factor in dividing the facies. The present study attempts to propose an advanced, optimized model through integrating the intelligent systems to estimate the FZI in the whole oilfield. This Committee Machine (CM) integrates the predicted results obtained from the intelligent neural, fuzzy logic, and Nero-Fuzzy systems with defined weights. Optimized weights for each method are determined using the Ant Colony Optimization Routing (ACOR) Algorithm. In this study, to apply the methods, well log and seismic data were used from one of the oilfields in South Iran. At the first stage, seismic attributes which were far more associated with the target data (FZI) were selected by stepwise regression. Subsequently, a 3D cube flow indicator in the whole field was estimated with intelligent systems. Finally, various reservoir facies were classified by the means of Fuzzy C-Mean Algorithm. The results illustrate that the committee machine which utilizes ACOR outperforms other individual systems acting alone.