Fuzzy logic-driven and SVM-driven hybrid computational intelligence models applied to oil and gas reservoir characterization

This work demonstrates the capabilities of two hybrid models as Computational Intelligence tools in the prediction of two important oil and gas reservoir properties, viz., porosity and permeability. The hybrid modeling was based on the combination of three existing Artificial Intelligence techniques: Functional Networks, Type-2 Fuzzy Logic System, and Support Vector Machines, using six datasets by utilizing the functional approximation capabilities of Functional Networks, the ability of Type-2 Fuzzy Logic to handle uncertainties and the scalability and robustness of Support Vector Machines in handling small and high-dimensional data. The hybridization was done in a way that allows one technique to further improve on the output of the other. Various Artificial intelligence techniques have been used in the prediction of oil and gas reservoir properties but each technique have exhibited specific capabilities, demonstrated certain limitations and posed a number of challenges. They have proven clearly that no single technique is perfect in all situations; hence the need for hybrid models that will combine the best characteristics of each technique in a single package and in the process, increase the confidence in the prediction of various oil and gas reservoir properties. This will result in increased production of more crude oil and hydrocarbons to meet the increasing world's demand. The results showed that the hybrid models perform better with higher correlation coefficients than the individual techniques when used alone for the same sets of data. In terms of execution time, the hybrid models took less time for both training and testing than the Type-2 Fuzzy Logic, but more time than Functional Networks and Support Vector Machines. This could be the price to pay for having better and more robust models. This work has demonstrated a successful application of the hybridization of three Artificial Intelligence techniques in one of the real-life problems encountered in oil and gas production where high quality information and accurate predictions are required for better and more efficient exploration, resource evaluation and their management.