Development and application of the artificial neural network based technical screening guide system to select production methods in a coalbed methane reservoir

摘要

The technical screening guide system was developed using and artificial neural network (ANN) to assist in the selection of production methods such as drilling, completion, and stimulation in a coalbed methane (CBM) reservoir. The ANN was trained with a Bayesian regularization algorithm utilizing field data obtained from the various CBM projects. To develop the system, the field database and the ANN model were constructed. Based on the literatures, the factors and ranges affecting the decision of CBM production methods were determined. The optimum system architecture was designed by conducting a sensitivity analysis with the training algorithm and proper number of hidden layers and neurons. The results from the ANN evaluation model indicated that the test was successful, yielding a correlation coefficient of 0.99. The system was also utilized to evaluate the field application in North American basins, and positive results could be obtained. It was confirmed that the technical screening guide system can be successful in the prediction of a proper CBM production method.