Prediction of the productivity of horizontal wells based on gray-relation analysis and neural network
Minfeng Chen · 2009
The productivity of horizontal wells is determined by complicated factors. The dominant factors forming the various influential factors have been found put via gray-relation analysis. Based on the analysis, we have been able to predict the productivity of horizontal wells by BP neural network. The method is to focus on predicting initial or current production. The neural network model is good in stability and its computation is precise. The error committed by the model is between ±10% which was acceptable in field application.