Recognition Method Of Non-Stationary Mechanical Vibration Signal Based On Convolution Neural Network

Abstract

In order to realize the accurate recognition of mechanical vibration signal, a method of Non-Stationary Mechanical vibration signal recognition based on convolution neural network is proposed. The characteristic values of Non-Stationary Mechanical vibration signals are collected, and the convolution neural network parameters are processed by combining the characteristic values of multi frequency mechanical vibration signals. Based on emd-svd-de software, the source number of mechanical vibration signals is estimated and the signal is decomposed. The signal characteristics of different mechanical vibration frequencies are analyzed to realize the recognition of Non-Stationary Mechanical vibration signals. Finally, the experimental results show that the research method has higher accuracy and effectiveness, and fully meets the research requirements.