Research on Fault Diagnosis of Anode Effect based on Wavelet Elman Neural Network

摘要

Anode effect is a common fault in aluminum electrolysis. Fault prediction is difficult as it occurs abruptly. A new fault diagnosis approach for anode effect based on Elman neural network is proposed in this papaer. According to the characteristics when anode effect occurs, the network is optimized by wavelet theory in order to raise the accuracy of fault diagnosis. Simulation results show that wavelet Elman neural network can predict anode effect accurately, and it has a certain value in engineering applications.