Application of neural networks for analysis, imitation modelling and optimal control of oil fields

It is necessary to solve monitoring problems of oil recovery, analysis problems of oil reservoir conditions, problems of fluid filtration processes modelling and to form forecasts for effective oil field development. Results of these researches are the information and intellectual base for decision making in optimal control of oil field development systems. However, traditional methods of analysis and mathematical modelling of oil reservoirs and their software implementation are not always suitable for the use by oil producers and engineers and modelling itself is more "state of art" than professional. These traditional analysis methods are based on solving of differential equations and boundary problems of mathematical physics. The development of new technologies of imitation modelling is necessary for solving existing mathematical modelling problems. In this report the authors discuss possibilities of neural network technology application for solving some live problems of analysis and modelling problems of systems with hydrodynamic interacting wells and oil reservoirs.