Abstract
The bearing state signal collected by the vibration sensor contains a large amount of environmental noise in actual processes, which leads to a reduction in the accuracy of the convolutional network in identifying bearing faults. To solve this problem, a one-dimensional convolutional neural network with a multiscale kernel (MSK-1DCNN) is proposed for the classification information enhancement of the input. A two-layer multiscale convolution structure (MSK) is used at the front of the network. MSK has five convolutional kernels with different sizes, and those kernels are used to extract features with varying resolutions in the original signal. In the multiscale convolution structure, the ELU activation function is used instead of the ReLU function to improve the antinoise ability of MSK-1DCNN, also by adding pepper noise to the training set data to destroy the input data and forcing the network to learn more representative features to improve the robustness of the network. Experimental results illustrate that the improved methods proposed in this paper effectively enhance the diagnostic performance of MSK-1DCNN under intense noise, and the diagnostic accuracy is higher than that of other comparison algorithms.
Highlights
Rolling bearings are an essential component and the main factor leading to system failures in rotating machinery. 45%– 55% of equipment failures are caused by bearing damage [1]
Introduction of Convolutional Neural Networks e convolutional neural network is a multilevel feed-forward neural network, which is usually composed of three types of layers: a convolutional layer, pooling layer, and fully connected layer. e convolutional layer and the pooling layer extract the characteristics of input data through convolution calculation and downsampling operations
BP neural network and SAE’s full connection structure leads to serious network overfitting, so even if the SNR is high, the diagnostic accuracy of them is low. e diagnosis accuracy of the multiscale kernel feature extraction structure (MSK)-1DCNN fault diagnosis model we proposed at low SNR is significantly higher than that of other models, which proves the effectiveness of improvements made in this paper of bearing faults diagnosis under the noise environment
Summary
Rolling bearings are an essential component and the main factor leading to system failures in rotating machinery. 45%– 55% of equipment failures are caused by bearing damage [1]. TICNN directly extracts the fault characteristics from the original vibration signal without additional data preprocessing It has made the following improvements: (1) Convolution kernel dropout is used in the first convolutional layer; (2) small batch training is used in the optimization algorithm, and ensemble learning is used to improve the stability of the network. Erefore, most models have not achieved good diagnostic accuracy in the presence of noise To address this problem, we propose a one-dimensional convolutional neural network with multiscale convolution kernels (MSK1DCNN). E activation function is usually used to implement a nonlinear transformation on the output of convolution calculation to obtain a nonlinear representation of input data, thereby improving the feature-learning ability of the network. Where qli(t) represents the output of the tth neuron in the ith feature map of the layer l, t ∈ [(j − 1)W + 1, jW], Wis the width of the pooled area, and Pli+1(j) is the pooled value of the corresponding neuron in the layer l + 1
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
