A Convolutional Neural Network with Whale Optimization and Batch Normalization for the Denoising of Vibration Signal
Due to the random selection of the initial weights,traditional convolutional neural network is prone to trap in local optimum,which leads to the difficulties in the extraction of clean signals from noisy vibration signals.To address this problem,a convolutional neural network with whale optimization and batch normalization(WO-BN-CNN)is proposed.A batch normalization layer is added after the convolu-tional neural network to normalize the parameter distribution in the hidden layer.The whale optimization algorithm is applied to optimize the network weight parameters.The amplitude spectrum of the noisy vi-bration signal and the time-domain waveform of the noisy signal are taken as the training features and the training target respectively to fully utilize the distribution characteristics of vibration signals in the time-frequency domain and realize the denoising by residual learning.The experimental results show that the proposed method improves the signal-to-noise ratio and reduces the mean square error and mean abso-lute error compared with wavelet threshold denoising,EMD,convolutional neural network,which enhances the denoising ability while preserving the original features of the vibration signal.
国家科技期刊平台
NSTL
万方数据
