Connectionist model predicts the porosity and permeability of petroleum reservoirs by means of petro-physical logs: Application of artificial intelligence

Abstract In this paper, a new approach based on artificial intelligence concept is evolved to monitor the permeability and porosity of petroleum reservoirs by means of petro-physical logs at various conditions. To address the referred issue, different artificial intelligence techniques including fuzzy logic (FL) and least square support vector machine (LSSVM) were carried out. Potential application of LSSVM and FL optimized by genetic algorithm (GA) is proposed to estimate the permeability and porosity of petroleum reservoirs. The developed intelligent approaches are examined by implementing extensive real field data from northern Persian Gulf oil fields. The results obtained from the developed intelligent approaches are compared with the corresponding real petro-physical data and gained outcomes of the other conventional models. The correlation coefficient between the model estimations and the relevant actual data is found to be greater than 0.96 for the GA–FL approach and 0.97 for GA–LSSVM. The results from this research indicate that implication of GA–LSSVM and GA–FL in prediction can lead to more reliable porosity/permeability predictions, which can lead to the design of more efficient reservoir simulation schemes.