基于双路并行多尺度ResNet的滚动轴承故障诊断方法

摘要
图/表
参考文献(25)
相关文章 (15)

全文: PDF (4248 KB) HTML (1 KB)
输出: BibTeX | EndNote (RIS)

摘要针对传统方法在滚动轴承故障诊断中无法自适应提取有效特征信息，且滚动轴承在强环境噪声干扰、复杂变工况等因素影响下诊断效果不佳，抗噪性和泛化性下降的问题，本文提出了一种双路并行多尺度的改进残差神经网络的方法。该方法设计了多尺度的残差Inception模块，可以有效提取特征信息，同时加入注意力机制解决了数据的突变性和差异性，此外还使用多个空洞卷积的残差块扩大感受野，有助于提取更多特征信息，实现准确故障诊断。利用凯斯西储大学轴承数据集和东南大学变速箱数据集分别训练并测试了诊断效果，将本文方法与其它卷积神经网络的方法在变噪声、变工况情况下作了对比，诊断准确率最高达到99.73%，平均准确率也在95%以上，均高于其它比较方法。结果表明，本文方法在复杂多变的工况下具有较好的故障识别能力和泛化能力。

	服务

	把本文推荐给朋友
	加入我的书架
	加入引用管理器
	E-mail Alert
	RSS
	作者相关文章
	赵小强1
	2
	3
	张毓春1

关键词 ：故障诊断, 滚动轴承, 变工况, 注意力机制, 多尺度ResNet

Abstract：Aiming at the problems that traditional methods cannot adaptively extract effective feature information in the fault diagnosis of rolling bearings, and the diagnosis effect of rolling bearings is not good under the influence of strong environmental noise interference, complex variable working conditions and other factors, the noise resistance and generalization is reduced, This paper proposes a dual-path parallel multi-scale method to improve the residual neural network. This method designs a multi-scale residual Inception module, which can effectively extract feature information. At the same time, the attention mechanism is introduced to solve the mutability and difference of the data. In addition, it also uses the residual blocks of multiple dilated convolutions to expands the receptive field, which helps to extract more characteristic information and realize accurate fault diagnosis. The diagnosis effects are trained and tested by using the Case Western Reserve University bearing data set and Southeast University gearbox data set. The proposed method is compared with other methods of convolutional neural network under the conditions of variable noise and variable working conditions, the highest rate of the proposed method is 99.73%, and the average accuracy rate is also above 95%, which is higher than other comparison methods. The results show that the proposed method has better fault identification ability and generalization ability under complex and changeable working conditions.

Key words： Fault diagnosis Rolling bearing Variable working condition Attention mechanism Multiscale ResNet

收稿日期: 2021-09-23 出版日期: 2023-02-15

引用本文:

赵小强1,2,3，张毓春1. 基于双路并行多尺度ResNet的滚动轴承故障诊断方法[J]. 振动与冲击, 2023, 42(3): 199-208.
ZHAO Xiaoqiang1,2,3, ZHANG Yuchun1. Fault diagnosis method of rolling bearing based on dual-path parallel multi-scale ResNet method. JOURNAL OF VIBRATION AND SHOCK, 2023, 42(3): 199-208.

链接本文:

http://jvs.sjtu.edu.cn/CN/ 或 http://jvs.sjtu.edu.cn/CN/Y2023/V42/I3/199

[1] 王国彪, 何正嘉, 陈雪峰, 等. 机械故障诊断基础研究“何去何从”[J]. 机械工程学报, 2013, 49(01):63-72.
Wang Guobiao, He Zhengjia, Chen Xuefeng, et al. Basic research on mechanical fault diagnosis "Where to go"[J]. Chinese Journal of Mechanical Engineering, 2013, 49(01): 63-72.
[2] 李晗, 萧德云. 基于数据驱动的故障诊断方法综述[J]. 控制与决策, 2011, 26(01):1-9+16.
Li Han, Xiao Deyun. Summary of fault diagnosis methods based on data drive[J]. Control and Decision, 2011, 26(01):1-9+16.
[3] 赵志宏. 基于振动信号的机械故障特征提取与诊断研究[D]. 北京交通大学,2012.
Zhao Zhihong. Research on Feature Extraction and Diagnosis of Mechanical Fault Based on Vibration Signal[D]. Beijing Jiaotong University, 2012.
[4] 崔江, 王友仁. 基于支持向量机与最近邻分类器的模拟电路故障诊断新策略[J]. 仪器仪表学报, 2010, 31(01):45-50.
Cui Jiang,Wang Youren.A new analog circuit fault diagnosis strategy based on support vector machine and nearest neighbor classifier[J].Chinese Journal of Scientific Instrument,2010,31(01):45-50.
[5] 陈二恒, 贺德强, 刘建仁, 等. 基于小波包和贝叶斯分类的机车走行部滚动轴承故障诊断研究[J]. 铁道科学与工程学报, 2015, 12(03):636-642.
Chen Erheng, He Deqiang, Liu Jianren, et al. Study on fault diagnosis of rolling bearing of locomotive running gear based on wavelet packet and Bayes classification[J]. Journal of Railway Science and Engineering, 2015, 12(03):636-642.
[6] 周小勇, 叶银忠. 故障信号检测的小波基选择方法[J]. 控制工程, 2003(04):308-311.
Zhou Xiaoyong, Ye Yinzhong. Wavelet basis selection method for fault signal detection[J]. Control engineering, 2003 (04): 308-311.
[7] 杨宇,于德介,程军圣. 基于EMD与神经网络的滚动轴承故障诊断方法[J]. 振动与冲击, 2005(01):87-90+138-139.
Yang Yu, Yu Dejie, Cheng Junsheng. Fault diagnosis method of rolling bearing based on EMD and neural network[J]. Vibration and Shock, 2005(01):87-90+138-139.
[8] 胡爱军,孙敬敬,向玲.经验模态分解中的模态混叠问题[J].振动.测试与诊断, 2011, 31(04):429-434+532-533.
Hu Aijun,Sun Jingjing, Xiang Ling. The problem of modal aliasing in empirical mode decomposition[J]. Vibration,Test and Diagnosis, 2011, 31(04):429-434+532-533.
[9] 闫文源. 基于变分模态分解和SVM的滚动轴承故障诊断[D]. 河南理工大学, 2017.
Yan Wenyuan. Fault diagnosis of rolling bearing based on variational modal decomposition and SVM[D]. Henan University of Technology, 2017.
[10] 王博, 郭继昌, 张艳. 基于深度网络的可学习感受野算法在图像分类中的应用[J]. 控制理论与应用, 2015, 32(08):1114-1119.
Wang Bo, Guo Jichang, Zhang Yan. Application of Learning Receptive Field Algorithm Based on Deep Network in Image Classification[J]. Control Theory and Applications,2015,32(08):1114-1119.
[11] 张立智, 徐卫晓, 井陆阳, 等. 基于二维深度卷积网络的旋转机械故障诊断[J]. 机械强度, 2020, 42(05): 1039-1044.
Zhang Lizhi, Xu Weixiao, Jing Luyang, et al. Rotating machinery fault diagnosis based on two-dimensional deep convolutional network[J]. Mechanical Strength, 2020, 42(05):1039-1044.
[12] 肖雄, 王健翔, 张勇军,等. 一种用于轴承故障诊断的二维卷积神经网络优化方法[J]. 中国电机工程学报, 2019, 39(15):4558-4568.
Xiao Xiong, Wang Jianxiang, Zhang Yongjun, et al. A two-dimensional convolutional neural network optimization method for bearing fault diagnosis[J]. Proceedings of the CSEE, 2019, 39(15): 4558-4568.
[13] 张建付, 宋雨, 李刚,等. 基于长短时记忆神经网络的风电机组滚动轴承故障诊断方法[J]. 计算机测量与控制, 2017, 25(01):16-19.
Zhang Jianfu, Song Yu, Li Gang, et al. Fault diagnosis method of wind turbine rolling bearing based on long and short-term memory neural network[J]. Computer Measurement and Control, 2017, 25(01):16-19.
[14] 曲建岭, 余路, 袁涛, 等. 基于一维卷积神经网络的滚动轴承自适应故障诊断算法[J]. 仪器仪表学报, 2018, 39(07):134-143.
Qu Jianling, Yu Lu, Yuan Tao, et al. An adaptive fault diagnosis algorithm for rolling bearings based on one-dimensional convolutional neural network[J]. Chinese Journal of Scientific Instrument, 2018, 39(07):134-143.
[15] 章立军, 徐金梧, 阳建宏, 等. 自适应多尺度形态学分析及其在轴承故障诊断中的应用[J]. 北京科技大学学报, 2008(04):441-445.
Zhang Lijun, Xu Jinwu, Yang Jianhong, et al. Adaptive multi-scale morphological analysis and its application in bearing fault diagnosis[J]. Journal of University of Science and Technology Beijing, 2008(04):441-445.
[16] 洪奇峰, 施伟斌, 吴迪, 等. 深度卷积神经网络模型发展综述[J]. 软件导刊, 2020, 19(04): 84-88.
Hong Qifeng, Shi Weibin, Wu Di, et al. Summary of the development of deep convolutional neural network models[J]. Software Guide, 2020, 19(04): 84-88.
[17] 孔子宇, 王海瑞. 基于Inception-ResNet模型的轴承故障分类[J]. 电子测量技术, 2021, 44(05):55-62.
Kong Ziyu, Wang Hairui. Bearing fault classification based on Inception-ResNet model[J]. Electronic Measurement Technology, 2021, 44(05): 55-62.
[18] 雷景生, 闫晨阳, 杨忠光. 基于Inception模块的卷积自编码器图像去噪[J]. 计算机应用与软件, 2021, 38(02): 221-226+322.
Lei Jingsheng, Yan Chenyang, Yang Zhongguang. Image denoising based on Inception module with convolutional autoencoder[J]. Computer Applications and Software, 2021, 38(02):221-226+322.
[19] 郭玥秀,杨伟,刘琦,等. 残差网络研究综述[J].计算机应用研究, 2020, 37(05):1292-1297.
Guo Yuexiu, Yang Wei, Liu Qi, et al. Summary of residual network research[J]. Application Research of Computers, 2020, 37(05):1292-1297.
[20] 李彦冬, 郝宗波, 雷航. 卷积神经网络研究综述[J]. 计算机应用, 2016, 36(09):2508-2515+2565.
Li Yandong, Hao Zongbo, Lei Hang. Summary of Convolutional Neural Network Research[J]. Computer Applications, 2016, 36(09):2508-2515+2565.
[21] 朱张莉, 饶元, 吴渊, 等. 注意力机制在深度学习中的研究进展[J]. 中文信息学报, 2019, 33(06):1-11.
Zhu Zhangli, Rao Yuan, Wu Yuan, Qi Jiangnan, Zhang Yu. Research progress of attention mechanism in deep learning[J]. Journal of Chinese Information Processing, 2019, 33(06):1-11.
[22] 张神林. 基于卷积神经网络的滚动轴承及行星齿轮箱故障诊断方法[D]. 安徽工业大学,2018.
Zhang Shenlin. Fault diagnosis method of rolling bearing and planetary gearbox based on convolutional neural network[D]. Anhui University of Technology, 2018.
[23] 赵小强, 张青青, 陈鹏, 等. 基于PSO-BFA和改进Alexnet的滚动轴承故障诊断方法[J]. 振动与冲击, 2020, 39(07):21-28.
Zhao Xiaoqiang, Zhang Qingqing, Chen Peng, et al. Fault diagnosis method of rolling bearing based on PSO-BFA and improved Alexnet[J]. Journal of Vibration and Shock, 2020, 39(07):21-28.
[24] 赵小强, 梁浩鹏. 使用改进残差神经网络的滚动轴承变工况故障诊断方法[J]. 西安交通大学学报, 2020, 54(09):23-31.
Zhao Xiaoqiang, Liang Haopeng. Fault Diagnosis Method of Rolling Bearing under Variable Condition Using Improved Residual Neural Network[J]. Journal of Xi'an Jiaotong University, 2020, 54(09):23-31.
[25] 饶雷, 唐向红, 陆见光. 基于CNN-SVM和特征融合的齿轮箱故障诊断[J]. 组合机床与自动化加工技术, 2020(08): 130-133+142.
Rao Lei, Tang Xianghong, Lu Jianguang. Gearbox fault diagnosis based on CNN-SVM and feature fusion[J]. Modular Machine Tool and Automatic Manufacturing Technology, 2020(08):130-133+142.