Fault Detection of Modular Dispensing Based on Improved Wavelet-BP Neural Network
FAN Jian;LI Zhigang;LI Liguo;School of Mechanical Engineering,Nanjing University of Science and Technology;Inner Mongolia North Heavy Industry Group Co.Ltd.;
In WNN( wavelet neural network),the additional momentum method was adopted to adjust the network parameters and the learning rate by self-adapting learning rate method. The characteristic points of the current curve of dispensing motor are used as inputs,the convergence rate can be greatly improved by using the function approximation of WNN. The experimental results show that compared with the traditional BP neural network,the improved wavelet-BP neural network has better network performance and detection accuracy.
【CateGory Index】： E924;TP183