Afshin Tatar - Department of Chemical Engineering, Sahand University of Technology, Tabriz, Iran
S.A. Tabatabaei Nejad - Department of Chemical Engineering, Sahand University of Technology, Tabriz, Iran
Elnaz Khodapanah - Department of Chemical Engineering, Sahand University of Technology, Tabriz, Iran
Mosayyeb Kamari - National Iranian South Oil Company (NISOC), Reservoir Evaluation department

Permeability is a key factor in fluid flow in porous media and is of great importance in petroleum industry. Numerous correlation and different methods to predict permeability signifies this fact. Direct methods to predict permeability such as nuclear magnetic resonance (NMR) log or core analysis is very expensive. It can be assumed that all the drilled wells have full set logs. Thus, it is convenient to develop a model to predict permeability using full set logs as input. One of the best tools to predict permeability is the neural networks. The purpose of this study is to construct a novel and efficient method to predict permeability based on intelligent methods. For this aim, full set logs and core permeability data were acquired from open literature and radial basis function neural network along with genetic algorithm were used to develop a novel method. The proposed model was validated using two different neural networks. The results show that the proposed model predicts the permeability values satisfactorily and is superior to other investigated neural networks.

Permeability, Radial basis function neural network, Full set logs, Machine learning, Artificial intelligence, NMR Log