Determination of Reservoir Facies with Seismic Data and Artificial Neural Networks

One of the purposes in seismic interpretation and reservoir modeling is mapping the changes of anisotropy in the subsurface layers. In this study, multi-attribute analyses were applied based on ANN methods and well logs data to determine the reservoir facies alteration and heterogeneity in the Ghar reservoir of the Hendijan oil field. The Sequential Indicator Simulation (SIS) algorithm coupled with the possible trend and indicator kriging was performed for each facies. Comparison of the mentioned models with core facies shows that the accuracy of SIS algorithm coupled with the possible trend and accuracy of indicator kriging are about 95% and 92%, respectively. The results showed reservoir quality with an average porosity of 0.18 and average gamma of 29 is from moderate to high.