A feasibility study on modeling and prediction of production behavior in naturally fractured shale reservoirs

The Objective of this study is to demonstrate the feasibility of predicting production characteristics in a Devonian Shale reservoir. This paper discusses the use of a multidimensional, equation of state compositional simulator to develop a prototypical anisotropic shale reservoir. The production data from a group of single well simulations is used to train an Artificial Neural Network to generate production indicators for a well location based on geopositional parameters. This study was conducted to ascertain if an Artificial Neural Network could be adapted to the random uncertainty in naturally fractured shale reservoirs and whether this approach could further the understanding of this unconventional source for natural gas.