针对强噪音、时变转速下滚动轴承振动信号数据特征分布不一致及待监测故障样本不含标签的问题，提出一种残差注意力机制和子领域自适应的无监督迁移学习滚动轴承故障诊断方法。首先，为充分发挥卷积神经网络的图像分类能力，将时变转速下采集到的一维时域故障振动信号采用连续小波变换(continuous wavelet transform，CWT)转换成二维灰度图，作为本文模型的输入；其次，为更好提取源域与目标域的通用特征，特征提取器利用本文提出的残差通道注意力弱共享网络模型，该模型采用了残差网络的跨层连接方式和通道注意力机制，并弱化了传统网络模型强共享条件；再次，为匹配源域与目标域的条件分布差异，网络自适应层选择局部最大均值差异（local maximum mean discrepancy，LMMD）度量准则；最后，采用时变转速滚动轴承公开故障数据集进行试验验证与分析，结果表明本文提出的方法在强噪音、时变转速下平均识别精度达到93%以上，相比于传统卷积神经网络模型具有较好的泛化性、鲁棒性。
关键词：故障诊断；无监督迁移学习；残差注意力弱共享；时变转速；强噪音
Abstract
Aiming at the problem of inconsistent distribution of rolling bearing vibration signal data characteristics under strong noise and time-varying speed and the failure samples to be tested do not contain labels, a residual attention mechanism and sub-domain adaptive unsupervised transfer learning rolling bearing fault diagnosis method was proposed. Firstly, to give full play to the image classification capabilities of the convolutional neural network (CNN), the one-dimensional time-domain fault vibration signal under the time-varying speed was converted into a two-dimensional grayscale image using continuous wavelet transform (CWT), which was used as the input of the model in this article; Secondly, in order to better extract the common features of the source and target domains, the feature extractor used the residual channel attention weak sharing network model proposed in this paper, which used the cross-layer connection method of the residual network and the channel attention mechanism, and weakened the structural conditions of the traditional strong sharing network model; Thirdly, in order to match the conditional distribution difference between the source domain and the target domain, the network adaptation layer embedded the local maximum mean discrepancy (LMMD) measurement criterion; Finally, the time-varying speed rolling bearing public fault data set was used for experimental verification and analysis. The results show that the method proposed in this paper achieves an average recognition accuracy of more than 93% under strong noise and time-varying speed, which has better generalization and robustness than traditional convolutional neural network models.
Key words: Fault diagnosis；Unsupervised transfer learning； Weak sharing of residual attention； Time-varying speed； Strong noise
关键词
故障诊断 /
无监督迁移学习 /
残差注意力弱共享 /
时变转速 /
强噪音
{{custom_keyword}} /
Key words
Fault diagnosis /
Unsupervised transfer learning /
Weak sharing of residual attention /
Time-varying speed /
Strong noise
{{custom_keyword}} /
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] 董绍江，裴雪武，吴文亮，等. 基于多层降噪技术及改进卷积神经网络的滚动轴承故障诊断方法[J]. 机械工程学报，2021，57(01)：148-156.
DONG Shaojiang, PEI Xuewu, WU Wenliang, et al. Rolling bearing fault diagnosis method based on multi-layer noise reduction technology and improved convolutional neural network[J]. Chinese Journal of Mechanical Engineering, 2021, 57(01): 148-156.
[2] LU S L, WANG X X. A New Methodology to Estimate the Rotating Phase of a BLDC Motor with Its Application in Variable-Speed Bearing Fault Diagnosis[J]. IEEE Transactions on Power Electronics, 2017, 33(04): 3399-3410.
[3] JIANG X X, LI S M. A dual path optimization ridge estimation method for condition monitoring of planetary gearbox under varying-speed operation[J]. Measurement, 2016, 94: 630 - 644.
[4] 高冠琪，黄伟国，李宁，等. 基于时频挤压和阶比分析的变转速轴承故障检测方法[J]. 振动与冲击，2020，39(03)：205-210+226.
GAO Guanqi, HUANG Weiguo, Li Ning, et al. Fault detection method of variable speed bearing based on time-frequency extrusion and order analysis[J]. Vibration and Shock, 2020, 39(03): 205-210+226.
[5] 赵德尊，王天杨，褚福磊. 基于自适应广义解调变换的滚动轴承时变非平稳故障特征提取[J]. 机械工程学报，2020，56(03)：80-87.
ZHAO Dezun, WANG Tianyang, CHU Fu-lei. Feature extraction of rolling bearing time-varying non-stationary faults based on adaptive generalized demodulation transform[J]. Chinese Journal of Mechanical Engineering, 2020, 56(03): 80-87.
[6] AN Z H, LI S M, WANG J R, et al. A novel bearing intelligent fault diagnosis framework under time-varying working conditions using recurrent neural network[J]. ISA Transactions, 2020, 100: 155 - 170.
[7] HAN B K, JI S S, WANG J R, et al. An intelligent diagnosis framework for roller bearing fault under speed fluctuation condition[J]. Neurocomputing, 2021, 420: 171 - 180.
[8] SUN B, SAENKO K. Deep CORAL: Correlation Alignment for Deep Domain Adaptation[C]// Springer International Publishing. Springer International Publishing, 2016: 443-450.
[9] LI X, ZHANG W, DING Q, et al. Multi-Layer Domain Adaptation Method for Rolling Bearing Fault Diagnosis[J]. Signal Processing, 2019, 157: 180 - 197.
[10] SHEN C Q, WANG X, WANG D, et al. Dynamic Joint Distribution Alignment Network for Bearing Fault Diagnosis Under Variable Working Conditions[J]. IEEE Transactions on Instrumentation and Measurement, 2021, 70: 3510813.
[11] LI X D, HU Y, ZHENG J H, et al. Central Moment Discrepancy Based Domain Adaptation for Intelligent Bearing Fault Diagnosis[J]. Neurocomputing, 2020, 429: 12 - 24.
[12] HE K M, ZHANG X Y, REN S Q, et al. Deep Residual Learning for Image Recognition[C]// IEEE Conference on Computer Vision & Pattern Recognition. New York: IEEE Computer Society, 2016: 770 - 778.
[13] ZHU J, CHEN N, SHEN C Q. A New Deep Transfer Learning Method for Bearing Fault Diagnosis under Different Working Conditions[J]. IEEE Sensors Journal, 2020, 20(15): 8394 - 8402.
[14] HU J, SHEN L, ALBANIE S, et al. Squeeze-and -Excitation Networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2019, 42(08): 2011-2023.
[15] ZHU Y C, ZHUANG F Z, WANG J D, et al. Deep Subdomain Adaptation Network for Image Classification[J]. IEEE Transactions on Neural Networks and Learning Systems，2020, 32(04): 1713 - 1722.
[16] https://data.mendeley.com/datasets/v43hmbwxpm/2.
{{custom_fnGroup.title_cn}}
脚注
{{custom_fn.content}}
PDF(1396 KB)
