Accurate Estimation of Petrophysical Indexes by RBF, ANFIS, and MLP Networks

Abstract

Production and management of oil and gas in today’s highly competitive environment require the use of high tech tools. These tools provide the means by which the cost of exploration, production, and management of hydrocarbon resources may be reduced. Petrophysical characteristics of underlying formation have an important role in reservoir management and drilling wells. One of the most common ways to reach this information is well log analysis. Meanwhile, sometimes well logging does not implement well or some log data are accompanied by many errors. Thus, highly skilled experts and laboratory information are needed for interpretation and evaluation of data. Therefore, designing a model that is able to evaluate the petrophysical index using well log data without laboratory information will be very economical. In this study, after selecting the best logs (minimum information and maximum accuracy) some parameters, such as porosity, saturation, sonic, and density logs, were predicted by adaptive neuro fuzzy intelligence system, radial basis function, and artificial neural network models. Finally, the best models for each parameter were optimized and optimal epoch, neuron, function, and spread clarified. In fact, due to these new models, some important parameters of formations were predicted well and cost and time of data gathering were reduced.

Keywords:

• adaptive neuro fuzzy intelligence system
• artificial neural network
• petrophysical index
• radial basis function
• well logging

ACKNOWLEDGMENTS

The authors would like to thank the National Iranian Oil Company (NIOC) for supporting this study and M. Rostami and M. Behzadijo for their cooperation.