APPLICATION OF IMPROVED WAVELET NEURAL NETWORK TO FAULT DIAGNOSIS OF PUMPING WELLS

摘要

The paper introduced an improved wavelet neural network model based on the certain link between signs and failure of pumping wells. We increased the part of the direct connection between input and output node, and highlighted the con-elation of signs input and fault output. The experiment results show that using the improved wavelet neural network for fault diagnosis of pumping wells will greatly enhance the accuracy and speed of pumping wells fault diagnosis.