Permeability prediction in carbonate reservoirs using specific area, porosity and water saturation

This paper presents a comparative study of the capabilities of Extreme Learning Machines (ELM), Decision Trees (DT) and Artificial Neural Networks (ANN), in the prediction of permeability from specific surface area, porosity and water saturation. ANN has been applied in the prediction of various oil and gas properties but with limitations such as computational instability due to its lack of global optima. ELM and DT are recent advances in Artificial Intelligence with improved architectures and better performance. The techniques were optimized and applied to the same carbonate reservoir field dataset. Following the popular convention and to ensure fairness, a stratified sampling approach was used to randomly extract 70% of the dataset for training while the remaining 30% was used for testing. The results showed that ELM performed best with the highest correlation coefficient, lowest root mean square error and shortest execution time. This agrees perfectly with the literature that ELM has a more compact architecture optimized for faster execution than the original ANN. DT was also found to be a promising technique for reservoir modeling. The results showed that ELM performed best with the highest correlation coefficient, lowest root mean square error and shortest execution time. This agrees perfectly with the literature that ELM has a more compact architecture optimized for faster execution than the original ANN. DT was also found to be a promising technique for reservoir modeling.