Full text loading...
-
Fracture Characterizations from Well Testing Data Using Artificial Neural Networks
- Publisher: European Association of Geoscientists & Engineers
- Source: Conference Proceedings, 72nd EAGE Conference and Exhibition incorporating SPE EUROPEC 2010, Jun 2010, cp-161-00786
- ISBN: 978-90-73781-86-3
Abstract
Dual porosity model refers to those reservoirs which have two different media. The interporosity flow coefficient (λ), and storativity ratio (ω). Well testing analysis is used to estimate reservoir parameters that are used in the reservoirs description. Pressure derivative plots corresponding to different value of λ and ω, are dissimilar. Derivative plots are used to design a model based on ANN to estimate λ and ω. In this study the capability of Artificial Neural Network to estimate λ and ω from well testing data has been investigated. Well testing data for dual porosity reservoir have been generated and converted to derivative plots. The best configuration of ANN has been selected by a trial and error procedure through applying different training algorithms and changing the number of neurons in the hidden layer. Using this procedure, a two-layer ANN model has been found as an efficient tool to estimate ω and λ. The trained ANN has been validated using the test data not been used in the training data set. The results have shown that the ANN is capable of estimating λ and ω using derivative plot obtained from the reservoir simulation as well as the information obtained from the literatures.