在实际工业生产中，工况的不同会导致数据分布差异，这使得不同工作条件下的轴承故障诊断成为一个挑战。针对上述问题，提出了一种基于多对抗和平衡分布自适应的故障诊断方法。首先，通过改进的残差网络直接从原始振动信号中提取域不变特征，提高特征提取效率的同时保留了丰富的故障特征信息。其次，提出了关联对齐与多对抗域自适应相结合的域自适应方法，同时对齐源域和目标域的边缘分布和条件分布以最小化域间数据分布差异。再次，对平衡分布自适应方法进行改进，设计了一种平衡因子为自适应过程中的边缘分布和条件分布分配权重，增强跨域故障诊断效果。最后，通过公开轴承故障数据集验证所提方法的有效性，实验结果表明，相较于流行的域自适应方法，所提方法具有更高的故障诊断精度，在不同工作条件下轴承的故障诊断任务中具有实际的应用价值。
Abstract
In actual industrial production, different operating conditions lead to variations in data distribution, posing a challenge for bearing fault diagnosis under different working conditions. To address this issue, a fault diagnosis method based on multi-adversarial and balanced distribution adaptation was proposed. Firstly, an improved residual network was used to directly extract domain-invariant features from the original vibration signals, enhancing feature extraction efficiency while preserving rich fault feature information. Secondly, a domain adaptation method combining correlation alignment and multi-adversarial domain adaptation was proposed, which can simultaneously align marginal distribution and conditional distribution of source domain and target domain to minimize data distribution differences between domains.Thirdly, the balanced distribution adaptation method was improved with designing a balance factor to allocate weights to the marginal distribution and conditional distribution in the adaptation process, so as to enhance cross-domain fault diagnosis effect. Finally, the effectiveness of the proposed method was validated using publicly available bearing fault datasets. Experimental results show that compared to popular domain adaptation methods, the proposed method achieves higher fault diagnosis accuracy, showing practical application value in bearing fault diagnosis tasks under different working conditions.
关键词
轴承故障诊断 /
多对抗域自适应 /
平衡分布自适应 /
残差网络 /
关联对齐
{{custom_keyword}} /
Key words
bearing fault diagnosis /
multi-adversarial domain adaptation /
balanced distribution adaptation /
residual network /
correlation alignment
{{custom_keyword}} /
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] LIU B, YAN C F, LIU Y F, et al. ISEANet: An interpretable subdomain enhanced adaptive network for unsupervised cross-domain fault diagnosis of rolling bearing [J]. Advanced Engineering Informatics, 2024, 62:102610.
[2] XU Y, LI Z X, WANG S Q, et al. A hybrid deep-learning model for fault diagnosis of rolling bearings [J]. Measurement, 2021, 169:14.
[3] CHEN H M, MENG W, LI Y J, et al. An anti-noise fault diagnosis approach for rolling bearings based on multiscale CNN-LSTM and a deep residual learning model [J]. Measurement Science and Technology, 2023, 34:045013.
[4] ALNFIAI M M. Improved symbiotic organism search with deep learning for industrial fault diagnosis [J]. Cmc-Computers Materials & Continua, 2023, 74:3763-3780.
[5] TANG Y H, ZHANG C Y, WU J Z, et al. Deep learning-based bearing fault diagnosis using a trusted multiscale quadratic attention-embedded convolutional neural network [J]. IEEE Transactions on Instrumentation and Measurement, 2024, 73:3513215.
[6] SHAO H D, DING Z Y, CHENG J S, et al. Intelligent fault diagnosis among different rotating machines using novel stacked transfer auto-encoder optimized by PSO [J]. Isa Transactions, 2020, 105:308-319.
[7] XU G W, HUANG C X, SILVA D S D, et al. A compressed unsupervised deep domain adaptation model for efficient cross-domain fault diagnosis [J]. Ieee Transactions on Industrial Informatics, 2023, 19:6741-6749.
[8] QIAN Q, QIN Y, LUO J, et al. Deep discriminative transfer learning network for cross-machine fault diagnosis [J]. Mechanical Systems and Signal Processing, 2023, 186:109884.
[9] XIANG L, ZHANG X Y, ZHANG Y, et al. A novel method for rotor fault diagnosis based on deep transfer learning with simulated samples [J]. Measurement, 2023, 207:112350.
[10] XU H Y, LIU J, PENG X Y, et al. Deep dynamic adaptation network: a deep transfer learning framework for rolling bearing fault diagnosis under variable working conditions [J]. Journal of the Brazilian Society of Mechanical Sciences and Engineering, 2023, 45:41.
[11] WANG C X, WU S K, SHAO X. Unsupervised domain adaptive bearing fault diagnosis based on maximum domain discrepancy [J]. Eurasip Journal on Advances in Signal Processing, 2024, 2024(1):11.
[12] XU X F, YANG X, QIAO Z J, et al. Intelligent fault diagnosis for variable working conditions based on SAAFN and BICP [J]. IEEE Sensors Journal, 2024, 24(7):10841-10852.
[13] ZHANG Y, ZHANG S Z, ZHU Y H, et al. Cross-domain bearing fault diagnosis using dual-path convolutional neural networks and multi-parallel graph convolutional networks [J]. ISA Transactions, 2024, 152:129-142.
[14] WANG Z C, XIA H, YIN W Z, et al. An improved generative adversarial network for fault diagnosis of rotating machine in nuclear power plant [J]. Annals of Nuclear Energy, 2023, 180:109434.
[15] GUO J X, CHEN K, LIU J H, et al. Bearing fault diagnosis based on deep discriminative adversarial domain adaptation neural networks [J]. Computer Modeling in Engineering & Sciences, 2024, 138(3):2619-2640.
[16] HE K M, ZHANG X Y, REN S Q, et al. Deep residual learning for image recognition [C]//2016 IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas:CVPR, 2016.
[17] SUN B C, FENG J S, SAENKO K. Return of frustratingly easy domain adaptation [C]//30th Association for the Advancement of Artificial Intelligence Conference on Artificial Intelligence. Phoenix:AAAI, 2016.
[18] PEI Z Y, CAO Z J, LONG M S, et al. Multi-adversarial domain adaptation [C]//32nd AAAI Conference on Artificial Intelligence. New Orleans:AAAI, 2018.
[19] WANG J D, CHEN Y Q, HAO S J, et al. Balanced distribution adaptation for transfer learning [C]//17th IEEE International Conference on Data Mining. New Orleans:ICDM, 2017.
[20] ZHANG J H, ZHU Y G, LI W Y, et al. DRNet: a deep neural network with multi-layer residual blocks improves image denoising [J]. Ieee Access, 2021, 9:79936-79946.
[21] VERSTRAETE D, FERRADA A, DROGUETT E L. Deep learning enabled fault diagnosis using time-frequency image analysis of rolling element bearings [J]. Shock and Vibration, 2017, 2017:17.
[22] LONG M S, CAO Y, WANG J M, et al. Learning transferable features with deep adaptation networks [C]//32nd International Conference on Machine Learning. Lille:ICML, 2015.
[23] LONG M S, ZHU H, WANG J M, et al. Deep transfer learning with joint adaptation networks [C]//34th International Conference on Machine Learning. Sydney:ICML, 2017.
[24] GANIN Y, LEMPITSKY V. Unsupervised domain adaptation by backpropagation [C]//32nd International Conference on Machine Learning. Lille:ICML, 2015.
[25] LAURENS V D M, HINTON G. Visualizing data using t-SNE [J]. Journal of Machine Learning Research, 2008, 9:2579-2605.
[26] HUANG H, BADDOUR N. Bearing vibration data collected under time-varying rotational speed conditions [J]. Data Brief, 2018, 21: 1745-1749.
{{custom_fnGroup.title_cn}}
脚注
{{custom_fn.content}}
PDF(3093 KB)
