A new approach to improve permeability prediction of petroleum reservoirs using neural network adaptive wavelet (wavenet)

• Introducing a new simulation method and its acceleration by wavelet transform.
• Comparison of the new methods with some of the past efficient methods.
• Identification of the optimal parameters of the method.

Permeability is one of the most important characteristics of hydrocarbon bearing formations. An accurate knowledge of permeability provides petroleum engineers with a tool for efficiently managing the production process of a field. Furthermore, it is one of the most important pieces of information in the design and management of enhanced recovery operations. Formation permeability is often measured in the laboratory from cores or evaluated from well test data. Core analysis and well test data, however, are only available from a few wells in a field, while the majority of wells are logged. Therefore, applying an efficient method, which can model this important parameter, is necessary. One of these methods, which recently have been used frequently, is artificial neural networks (ANNs), which have a significant ability to find the complex spatial relationship in the existence parameters of reservoir. Despite all of the applications of ANNs, most of them need a time-consuming procedure of architecture design and the problem of local minima and slow convergence. In this paper, an alternative method of permeability prediction, which is based on integration between wavelet theory and Artificial Neural Network (ANN) or wavelet network (wavenet), is presented. In this study, different wavelets are applied as activation functions to predict the permeability for well logging. Wavenet parameters such as dilation and translation are fixed and only the weights of the network are optimized during its learning process. The efficacy of this type of network in function learning and estimation is compared with ANNs. The results showed that the wavelet network (WNN, Morlet) with 92% correlation coefficient for permeability would be an appropriate substitute for artificial neural network with 89% correlation coefficient.