Application of Improved Wavelet Neural Network in Fault Diagnosis in Frequency Conversion System

Currently,fault diagnosis for the frequency conversion system depends on the monitoring and analyzing of voltage and electric current signals.This paper presented a method that the eigenvectors of the fault signals of the frequency converser were extracted using the improved wavelet packet,and then the eigenvectors were used as learning samples for the training.The diagnosis results show that in comparison with the conventional neural network diagnosis method,the improved wavelet neural network dia-gnosis method has the advantages of high accuracy,fast diagnosis and good generalization ability.