Application of gradient-based control methods in efficient oil well placement through dynamic fuzzy neural network modeling

Abstract

This study presents an approach for determining optimal locations of oil wells in an oil field such that a predetermined production policy can be achieved. In the first step, Dynamic Fuzzy Neural Network (DFNN) is employed to generate an analytical and dynamic model of the reservoir. The model would be updated during the process due to current reservoir information. In order to determine optimal weights for DFNN, Orthogonal Least Square method is modified and applied. It will be shown that the method has more computational efficiency in comparison with other approaches. In the second stage, gradient-based approaches as one of the most common methods in control field are employed to determine well locations based on the generated model. The method can perform search without being obliged to use the simulator several times. Finally, simulation results show the abilities of the proposed procedure in both modeling and tracking control. © 2011 Springer-Verlag.