Fault Diagnosis Method of Rolling Bearing Based on bCNN-LSTM Layout guide for Journal of Physics: Conference Series using Microsoft Word

Abstract To address the present methods based on Rectified Linear Unit (ReLU) were prone to mean shift, a fault diagnostic optimization method was proposed that combined the Convolutional Neural Network (CNN) of bReLU activation function and Long Short-Term Memory (LSTM). The CNN layer based on a no...

Full description

Saved in:

Bibliographic Details
Published in:	Journal of physics. Conference series 2022-07, Vol.2303 (1), p.12050
Main Authors:	Xiang, Xuan, Cao, Shaozhong, Yang, Yanhong, Wei, Xuhang
Format:	Article
Language:	eng
Subjects:	Artificial neural networks Fault diagnosis Feature extraction Journal bearings Optimization Physics Roller bearings
Online Access:	Get full text
Tags:	Add Tag No Tags, Be the first to tag this record!

Description
Summary:	Abstract To address the present methods based on Rectified Linear Unit (ReLU) were prone to mean shift, a fault diagnostic optimization method was proposed that combined the Convolutional Neural Network (CNN) of bReLU activation function and Long Short-Term Memory (LSTM). The CNN layer based on a novel activation function bReLU was utilized to adaptively extract features, and enhanced feature extraction speed by adding a batch standardization layer between the convolution layer and the activation function. In addition, the LSTM layer was employed to learn features to classify rolling bearing faults. The accuracy of the enhanced method was greater than 99% on the single-channel fault dataset of Case Western Reserve University (CWRU), and it also has a simpler model structure and takes less time to train than other CNN-LSTM ways.
ISSN:	1742-6588 1742-6596