Modified Morlet Wavelet Neural Networks for Fault Detection

Abstract

Wavelet neural networks (WNN) combining the properties of the wavelet transform and the advantages of artificial neural networks (ANNs) have attracted great interest and become a popular tool for various fields of mathematics and engineering. We describe here the application of the modified Morlet based WNN to the fault detection of rotating machinery. The activation functions of the wavelet nodes in the hidden layer are derived from a modified Morlet mother wavelet. In this paper, the wavelet network architecture for intelligent fault detection is first introduced. Then an optimization method of neural network and a training algorithm is developed. Finally, feedforward backpropagation neural network (BP) and wavelet neural networks are compared for fault detection. The aim of this study is to examine the effective of the modified Morlet WNN model for fault detection. Experiment results shows that the modified Morlet WNN has advantages of rapid training, generality and accuracy over feedforward backpropagation neural network.