船电技术2025，Vol.45Issue(7)：7-10,4.
基于卷积神经网络的滚动轴承故障诊断方法研究
Research on rolling bearing fault diagnosis method based on convolutional neural network
蒋炜 1王永兴1
作者信息
- 1. 武汉威迈新能源动力有限公司,武汉 430064
- 折叠
摘要
Abstract
To improve the diagnostic accuracy and reduce the reliance on manual feature extraction,this paper proposes a rolling bearing fault diagnosis method based on Convolutional Neural Network(CNN).As a critical component,rolling bearing failures may lead to equipment malfunctions or accidents.Traditional methods have limitations,whereas CNN can automatically extract signal features for efficient fault identification.This study utilizes the CWRU dataset,preprocesses vibration signals(denoising,data augmentation,and time-frequency transformation),and constructs a multi-layer CNN trained with the ReLU activation function and Adam optimizer.Experimental results show that this method achieves a classification accuracy of over 98%across 10 fault categories and demonstrates strong robustness under complex working conditions.Compared with traditional methods,CNN enables more accurate and efficient fault diagnosis while reducing human intervention,highlighting its promising engineering applications.关键词
滚动轴承/故障诊断/卷积神经网络Key words
rolling bearing/fault diagnosis/convolutional neural network分类
计算机与自动化引用本文复制引用
蒋炜,王永兴..基于卷积神经网络的滚动轴承故障诊断方法研究[J].船电技术,2025,45(7):7-10,4.
