基于通道注意力机制与多尺度减法轻量化网络的滚动轴承故障诊断

章力; 邓艾东; 王敏; 卞文彬; 张宇剑

doi:10.19805/j.cnki.jcspe.2025.240071

欢迎来到能源电力期刊网！

English Version 读者登录期刊登录

您当前的位置：

首页 >

文章列表页 >

基于通道注意力机制与多尺度减法轻量化网络的滚动轴承故障诊断

数字化与智能化 | 更新时间：2025-11-28

- 基于通道注意力机制与多尺度减法轻量化网络的滚动轴承故障诊断
- Fault Diagnosis of Rolling Bearing Based on Channel Attention Mechanism and Multi-scale Subtraction Lightweight Neural Network
- 动力工程学报 2025年45卷第4期页码：571-581
- 作者机构：
  
  1. 东南大学大型发电装备安全运行与智能测控国家工程研究中心,江苏,南京,210096
  2. 东南大学能源与环境学院,江苏,南京,210096
- 作者简介：
  
  [ "章力(1995—),男,浙江绍兴人,硕士研究生,主要从事旋转设备故障诊断方面的研究" ]
  [ "邓艾东(通信作者),男,教授,博士生导师,E-mail:dnh@seu.edu.cn" ]
- 基金信息：
  
  江苏省碳达峰碳中和科技创新专项资金资助项目(BA2022214);中央高校基本科研业务费专项资金资助项目(2242023K30011、2242023K30058)
- DOI：10.19805/j.cnki.jcspe.2025.240071
  中图分类号： TM315;TP183
- 网络出版：2025-04-28，
  
  纸质出版：2025
- 稿件说明：
移动端阅览
章力,邓艾东,王敏,卞文彬,张宇剑. 基于通道注意力机制与多尺度减法轻量化网络的滚动轴承故障诊断动力工程学报, 2025, 45(4): 571-581 https://doi.

org/10.19805/j.cnki.jcspe.2025.240071
章力,邓艾东,王敏,卞文彬,张宇剑. 基于通道注意力机制与多尺度减法轻量化网络的滚动轴承故障诊断动力工程学报, 2025, 45(4): 571-581 https://doi. DOI： 10.19805/j.cnki.jcspe.2025.240071.

org/10.19805/j.cnki.jcspe.2025.240071 DOI：

摘要

针对传统多尺度卷积神经网络模型存在的特征定位不精确、训练时间长、抗噪性能差等问题

提出了一种基于通道注意力机制与多尺度减法轻量化网络的滚动轴承故障诊断模型。首先

将滚动轴承的一维振动信号转换为二维灰度图作为输入

丰富特征信息;同时

构建多尺度减法神经网络模型

关注层级差异;其次

引入轻量化模块

减少内存访问;然后

结合通道注意力机制

调整特征权重;最后

将故障样本输入到网络模型中

实现精确分类。利用风电机组传动系统模拟实验台采集的样本数据进行诊断任务。结果表明:该故障诊断模型能够有效克服传统多尺度卷积神经网络模型网络层数多、参数量大所带来的弊端

能够充分关注各层级之间的差异信息

减少冗余信息的提取

精确定位故障特征

缩短模型训练时间

在恒定工况、变工况和强噪声工况下都具有较高的诊断精度。

Abstract

Aiming at the problems of inaccurate feature location

long training time

and poor anti-noise performance in traditional multi-scale convolutional neural network models

a fault diagnosis model for rolling bearings based on channel attention mechanism and multi-scale subtraction lightweight neural network was proposed. First

the one-dimensional vibration signal of rolling bearing was converted into a two-dimensional grayscale images as input to enrich feature information. Meanwhile

a multi-scale subtraction neural network model was constructed to focus on level differences. Secondly

a lightweight module was introduced to reduce memory access. Then

the feature weights were adjusted by combining with the channel attention mechanism. Finally

the fault samples were input into the network model to achieve accurate classification. The diagnosis tasks were conducted using the sample data collected from the wind turbine transmission system simulation test bench. Results show that this model has high diagnostic accuracy under constant

variable working conditions and strong noise working conditions

which can overcome the drawbacks caused by the large number of network layers and parameters in traditional multi-scale convolutional neural network models. It can fully focus on the difference information between the levels

reduce the extraction of redundant information

locate fault characteristics accurately

and shorten the model training time. It has high diagnostic accuracy under constant working condition

variable working conditions and strong noise conditions.

关键词

Keywords

references

陈雪峰, 郭艳婕, 许才彬, 等. 风电装备故障诊断与健康监测研究综述[J].中国机械工程, 2020, 31(2): 175-189. CHEN Xuefeng, GUO Yanjie, XU Caibin, et al. Review of fault diagnosis and health monitoring for wind power equipment[J]. China Mechanical Engineering, 2020, 31(2): 175-189.

邓子豪, 李录平, 刘瑞, 等. 基于SCADA数据特征提取的风电机组偏航齿轮箱故障诊断方法研究[J]. 动力工程学报, 2021, 41(1): 43-50. DENG Zihao, LI Luping, LIU Rui, et al. Research on diagnosis method of wind turbine yaw gearbox based on SCADA data feature extraction[J].Journal of Chinese Society of Power Engineering, 2021, 41(1): 43-50.

周文宣, 刘洋, 邓敏强, 等. 基于CAE和CNN的变工况下滚动轴承智能故障诊断研究[J]. 动力工程学报, 2022, 42(1): 43-48. ZHOU Wenxuan, LIU Yang, DENG Minqiang, et al. Research on intelligent fault diagnosis of rolling bearing under variable conditions based on CAE and CNN[J]. Journal of Chinese Society of Power Engineering, 2022, 42(1): 43-48.

孙康, 金江涛, 李春, 等. 优化变分模态分解与深度学习融合的滚动轴承故障诊断策略[J]. 动力工程学报, 2023, 43(6): 749-758. SUN Kang, JIN Jiangtao, LI Chun, et al. Fault diagnosis strategy for rolling bearing based on optimal fusion of variational modal decomposition and deep learning[J]. Journal of Chinese Society of Power Engineering, 2023, 43(6): 749-758.

JIAO Jinyang, ZHAO Ming, LIN Jing, et al. A comprehensive review on convolutional neural network in machine fault diagnosis[J]. Neurocomputing, 2020, 417: 36-63.

葛江华, 刘奇, 王亚萍, 等. 支持张量机与KNN-AMDM决策融合的齿轮箱故障诊断方法[J]. 振动工程学报, 2018, 31(6): 1093-1101. GE Jianghua, LIU Qi, WANG Yaping, et al. Fault diagnosis method of gearbox supporting tension machine and KNN-AMDM decision fusion[J]. Journal of Vibration Engineering, 2018, 31(6): 1093-1101.

UNAL M, ONAT M, DEMETGUL M, et al. Fault diagnosis of rolling bearings using a genetic algorithm optimized neural network[J].Measurement, 2014, 58: 187-196.

DING Yao, ZHANG Zhili, ZHAO Xiaofeng, et al. Multi-feature fusion: graph neural network and CNN combining for hyperspectral image classification[J]. Neurocomputing, 2022, 501: 246-257.

黄雅静, 廖爱华, 丁亚琦, 等. 多尺度CNN结合自注意力特征融合机制的轴承故障诊断方法[J].智能计算机与应用, 2022, 12(9): 37-44. HUANG Yajing, LIAO Aihua, DING Yaqi, et al. Bearing fault diagnosis based on multi-scale convolutional neural network and self-attention feature fusion mechanism[J]. Intelligent Computer and Applications, 2022, 12(9): 37-44.

陈奇, 陈长征, 安文杰. 基于多通道融合多尺度自适应残差学习的行星齿轮箱故障诊断研究[J]. 机电工程, 2023, 40(7): 1031-1038. CHEN Qi, CHEN Changzheng, AN Wenjie. Fault diagnosis of planetary gearbox based on multi-channel fusion and multi-scale adaptive residual learning[J]. Journal of Mechanical & Electrical Engineering, 2023, 40(7): 1031-1038.

王妮妮, 马萍, 张宏立, 等. 基于多尺度深度卷积网络特征融合的滚动轴承故障诊断[J].太阳能学报, 2022, 43(4): 351-358. WANG Nini, MA Ping, ZHANG Hongli, et al. Fault diagnosis of rolling bearing based on feature fusion of multi-scale deep convolutional network[J]. Acta Energiae Solaris Sinica, 2022, 43(4): 351-358.

CHEN Yahui. Convolutional neural network for sentence classification[D]. Canada: University of Waterloo, 2015.

LIANG Haopeng, CAO Jie, ZHAO Xiaoqiang. Multi-scale dynamic adaptive residual network for fault diagnosis[J]. Measurement, 2022, 188: 110397.

HUANG Wenyi, CHENG Junsheng, YANG Yu, et al. An improved deep convolutional neural network with multi-scale information for bearing fault diagnosis[J]. Neurocomputing, 2019, 359: 77-92, doi: 10.1016/j.neucom.2019.05.052.

宫文峰, 陈辉, 张美玲, 等. 基于深度学习的电机轴承微小故障智能诊断方法[J].仪器仪表学报, 2020, 41(1): 195-205. GONG Wenfeng, CHEN Hui, ZHANG Meiling, et al. Intelligent diagnosis method for incipient fault of motor bearing based on deep learning[J]. Chinese Journal of Scientific Instrument, 2020, 41(1): 195-205.

HU Jie, SHEN Li, SUN Gang. Squeeze-and-excitation networks[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City, UT, USA: IEEE, 2018: 7132-7141.

HOWARD A, SANDLER M, CHEN Bo, et al. Searching for mobileNetV3[C]//Proceedings of 2019 IEEE/CVF International Conference on Computer Vision. Seoul, Korea (South): IEEE, 2019: 1314-1324.

SANDLER M,HOWARD A, ZHU Menglong, et al. Mobilenetv2: inverted residuals and linear bottlenecks[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, Salt Lake City, UT, USA:IEEE, 2018.

MA Ningning, ZHANG Xiangyu, ZHENG Haitao, et al. ShuffleNet v2: practical guidelines for efficient CNN architecture design[C]//Proceedings of the 15th European Conference on Computer Vision (ECCV). Munich, Germany: Springer, 2018: 116-131.

HAN Kai, WANG Yunhe, TIAN Qi, et al. GhostNet: more features from cheap operations[C]//Proceedings of 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Seattle, WA, USA: IEEE, 2020: 1580-1589.

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

基于不平衡样本数据的轴承故障诊断模型与方法研究