基于无源自监督域适应的旋转机械跨域新故障诊断方法

doi:10.3901/JME.2025.03.067

机械工程学报 ›› 2025, Vol. 61 ›› Issue (3): 67-76.doi: 10.3901/JME.2025.03.067

• 特邀专栏：人机联合认知赋能的高端装备设计、制造与运维 • 上一篇

扫码分享

基于无源自监督域适应的旋转机械跨域新故障诊断方法

乐珂¹, 李霁蒲², 陈祝云^2,3, 何国林^2,3, 邓书涵¹, 李巍华^2,3

1. 华南理工大学吴贤铭智能工程学院广州 510640;
2. 华南理工大学机械与汽车工程学院广州 510640;
3. 琶洲实验室广州 510005

收稿日期:2024-02-19 修回日期:2024-08-01 发布日期:2025-03-12
作者简介:乐珂，女，1999年出生，博士研究生。主要研究方向为工业装备智能故障诊断与健康管理、工业大数据分析。E-mail:yonina25@163.com;李巍华(通信作者)，男，1973年出生，博士，教授，博士研究生导师。;主要研究方向为工业智能、工业大数据、装备智能运维、汽车智能驾驶。E-mail:whlee@scut.edu.cn
基金资助:
广东省重点领域研发计划(2023B0909050007,2021B0101200004)和国家自然科学基金(52275111,52075182,52205101)资助项目。

Cross-domain Emerging Fault Diagnosis of Rotating Machinery Using Source-free Self-supervised Domain Adaptation Network

YUE Ke¹, LI Jipu², CHEN Zhuyun^2,3, HE Guolin^2,3, DENG Shuhan¹, LI Weihua^2,3

1. Shien-Ming Wu School of Intelligent Engineering, South China University of Technology, Guangzhou 510640;
2. School of Mechanical & Automotive Engineering, South China University of Technology, Guangzhou 510640;
3. Pazhou Lab, Guangzhou 510005

Received:2024-02-19 Revised:2024-08-01 Published:2025-03-12

摘要/Abstract

摘要： 随着新一代人工智能技术的迅猛发展，数据驱动的智能故障诊断方法已经在机械装备中已取得广泛应用。智能诊断模型的高精度诊断通常依赖于大规模的已知标签数据。然而，机械装备在运行过程会发生不可预知的新故障，这使得已知故障数据训练的模型难以准确识别新出现的故障类型。此外，数据隐私保护限制着数据的可访问性，为诊断模型的域适应过程带来了巨大挑战。为此，提出一种基于无源自监督域适应的跨域新故障诊断方法，在源域数据仅参与模型预训练的情况下，实现对目标域中已知故障的精准分类和新故障的智能诊断。首先，利用带标签的源域样本预训练得到源域诊断模型；其次，利用基于不确定性熵的自监督伪标签技术筛选出高置信度的已知故障样本；最后，将目标域中带伪标签的已知故障样本与无标签的故障样本相结合，并通过对抗训练策略辨别新故障类型。利用实验室测试采集的汽车变速箱数据集开展跨域新故障诊断试验，对所提算法的有效性进行验证。试验结果表明，所提方法能够仅利用源域诊断模型和无标签目标域数据，在保障已知故障高精度诊断的同时，实现对目标域中新故障的精准诊断。

关键词: 智能故障诊断, 无源域适应, 新故障, 自监督学习, 迁移学习

Abstract: With the rapid advancement of next-generation artificial intelligence technologies, data-driven intelligent fault diagnosis methods have found widespread applications in mechanical equipment. High-precision diagnosis of intelligent diagnostic models typically relies on a large volume of labeled data. However, unpredictable new faults may occur during the operation of mechanical equipment, which makes it difficult to adopt the model trained on known samples to accurately identify newly occurring faults. Furthermore, data privacy restricts the accessibility of data, adding substantial complexity to the domain adaptation process for diagnostic models. To address these challenges, a source-free self-supervised domain adaptation network is proposed for cross-domain emerging fault diagnosis of rotating machinery, which enables diagnosis emerging fault in target domain without access to source domain. First, a source domain fault diagnosis model is established using labeled source samples. Subsequently, a self-supervised pseudo-labeling technique based on uncertainty information entropy is utilized to acquire a target domain dataset with high-quality pseudo-labels. Finally, we merge the pseudo-labeled fault dataset with unlabeled fault dataset, then train the model through an adversarial training strategy to realize emerging fault detection. The effectiveness of the proposed method is validated through experiments conducted on an automotive gearbox dataset. Experimental results show that the proposed method can accurately detect emerging faults while maintaining the performance in classifying known fault types.

Key words: intelligent fault diagnosis, source-free domain adaptation, emerging fault, self-supervised learning, transfer learning

中图分类号:

TG156

乐珂, 李霁蒲, 陈祝云, 何国林, 邓书涵, 李巍华. 基于无源自监督域适应的旋转机械跨域新故障诊断方法[J]. 机械工程学报, 2025, 61(3): 67-76.

YUE Ke, LI Jipu, CHEN Zhuyun, HE Guolin, DENG Shuhan, LI Weihua. Cross-domain Emerging Fault Diagnosis of Rotating Machinery Using Source-free Self-supervised Domain Adaptation Network[J]. Journal of Mechanical Engineering, 2025, 61(3): 67-76.

参考文献

[1] 高金吉. 人工自愈概论[J]. 机械工程学报，2021，57(2):1-10. GAO Jinji，Overview on Artificial Self-recovery[J]. Journal of Mechanical Engineering， 2021，57(2):1-10.
[2] 王辉，徐佳文，严如强. 基于多尺度注意力深度强化学习网络的行星齿轮箱智能诊断方法[J]. 机械工程学报，2022，58(11):133-142. WANG Hui， XU Jiawen， YAN Ruqiang. Multi-scale attention based deep reinforcement learning for intelligent fault diagnosis of planetary gearbox[J]. Journal of Mechanical Engineering， 2022，58(11):133-142.
[3] 邵海东，林健，闵志闪，等. 分布外样本干扰下基于改进半监督原型网络的齿轮箱跨域故障诊断[J]. 机械工程学报:2024，60(4):212-221. SHAO Haidong， LIN Jian， MIN Zhishan， et al， Improved semi-supervised prototype network for cross-domain fault diagnosis of gearbox under out-of-distribution interference samples[J]. Journal of Mechanical Engineering，2024，60(4):212-221.
[4] XU Yadong，FENG Ke，YAN Xiaoan，et al. Cross-modal fusion convolutional neural networks with online soft label training strategy for mechanical fault diagnosis[J]. IEEE Transactions on Industrial Informatics，2024，20(1):73-84.
[5] 雷亚国，杨彬，李乃鹏，等. 跨设备的机械故障靶向迁移诊断方法[J]. 机械工程学报，2022，58(12):1-9. LEI Yaguo， YANG Bin， LI Naipeng， et al. Targeted transfer diagnosis method across different machines[J]. Journal of Mechanical Engineering，2022，58(12):1-9.
[6] ZHANG Liangwei，FAN Qi，LIN Jing，et al. A nearly end-to-end deep learning approach to fault diagnosis of wind turbine gearboxes under nonstationary conditions[J]. Engineering Applications of Artificial Intelligence，2023， 119:105735.
[7] SHI Mingkuan，DING Chuancang，WANG Rui，et al. Deep hypergraph autoencoder embedding:An efficient intelligent approach for rotating machinery fault diagnosis[J]. Knowledge-Based Systems，2023，260: 110172.
[8] PENG Binsen，XIA Hong，LÜ Xinzhi，et al. An intelligent fault diagnosis method for rotating machinery based on data fusion and deep residual neural network[J]. Applied Intelligence，2022，52(3):3051-3065.
[9] MIAN T，CHOUDHARY A，FATIMA S. Multi-sensor fault diagnosis for misalignment and unbalance detection using machine learning[J]. IEEE Transactions on Industry Applications，2023，59(5):5749-5759.
[10] TANG Shengnan，ZHU Yong，YUAN Shouqi. Intelligent fault diagnosis of hydraulic piston pump based on deep learning and Bayesian optimization[J]. ISA Transactions， 2022，129:555-563.
[11] LI Weihua，HUANG Ruyi，LI Jipu，et al. A perspective survey on deep transfer learning for fault diagnosis in industrial scenarios:Theories，applications and challenges[J]. Mechanical Systems and Signal Processing， 2022，167:108487.
[12] ZHANG Wei，LI Xiang，MA Hui，et al. Open-set domain adaptation in machinery fault diagnostics using instance-level weighted adversarial learning[J]. IEEE Transactions on Industrial Informatics，2021，17(11):7445-7455.
[13] 陈祝云，林慧斌，夏景演，等.基于选择性加权适配网络的多域新故障识别方法[J]. 仪器仪表学报，2022，43(10):270-279. CHEN Zhuyun， LIN Huibin， XIA Jingyan，et al. Selective weighted adaptive network for multi-domainemerging fault identification[J]. Chinese Journal of Scientific Instrument，2022，43(10):270-279.
[14] WEN Long，SU Shaoquan，WANG Bin，et al. A new multi-sensor fusion with hybrid convolutional neural network with wiener model for remaining useful life estimation[J]. Engineering Applications of Artificial Intelligence，2023，126:106934.
[15] 李响，付春霖，雷亚国，等. 保证数据隐私的装备协同智能故障诊断联邦迁移学习方法[J]. 机械工程学报，2023，59(6):1-9. LI Xiang，FU Chunlin，LEI Yaguo，et al. Federated transfer learning method for privacy-preserving collaborative intelligent machinery fault diagnostics[J]. Journal of Mechanical Engineering，2023，59(6):1-9.
[16] YUE Ke，LI Jipu，CHEN Zhuyun，et al. Multiple source-free domain adaptation network based on knowledge distillation for machinery fault diagnosis[J]. IEEE Transactions on Instrumentation and Measurement， 2023，72:3523111.
[17] KRAUSE A， PERONA P， GOMES R. Discriminative clustering by regularized information maximization[J]. Advances in Neural Information Processing Systems， 2010，23:775-783.
[18] SAITO K，YAMAMOTO S，USHIKU Y，et al. Open set domain adaptation by backpropagation[C]//Proceedings of the European conference on computer vision (ECCV). 2018:153-168.
[19] HAN Te，LIU Chao，YANG Wenguang，et al. A novel adversarial learning framework in deep convolutional neural network for intelligent diagnosis of mechanical faults[J]. Knowledge-Based Systems， 2019，165:474-487.
[20] LI Jipu，HUANG Ruyi，HE Guolin，et al. A deep adversarial transfer learning network for machinery emerging fault detection[J]. IEEE Sensors Journal，2020，20(15):8413-8422.
[21] LIANG Jian，HU Dapeng，FENG Jiashi. Do we really need to access the source data? source hypothesis transfer for unsupervised domain adaptation[C]//International conference on machine learning. PMLR，2020:6028-6039.

基于无源自监督域适应的旋转机械跨域新故障诊断方法

Cross-domain Emerging Fault Diagnosis of Rotating Machinery Using Source-free Self-supervised Domain Adaptation Network

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

本文评价

[1]	马帅, 冷杰武, 陈祝云, 李巍华, 李波, 刘强. 基于数字孪生和深度迁移学习的电主轴热误差建模方法[J]. 机械工程学报, 2025, 61(3): 52-66.
[2]	邓聪颖, 邓子豪, 林丽君, 陈翔, 马莹, 禄盛. 基于迁移学习的变刀具-刀柄夹持状态下数控铣削稳定性预测研究[J]. 机械工程学报, 2024, 60(3): 296-304.
[3]	李响, 陈欣瑞, 雷亚国, 李乃鹏, 杨彬, 俞舒鹏. 动态视觉赋能的非接触式装备迁移诊断[J]. 机械工程学报, 2024, 60(24): 1-10.
[4]	赵轲, 叶敏, 王瑞欣, 陆海, 刘孟孟, 邵海东. 基于模糊域自适应的源自由域旋转机械故障诊断方法[J]. 机械工程学报, 2024, 60(18): 43-52.
[5]	朱永超, 朱才朝, 谭建军, 冉峯, 宋朝省. 考虑数据特征差异的风电齿轮箱群组健康状态预测研究[J]. 机械工程学报, 2024, 60(18): 64-75.
[6]	温楷儒, 陈祝云, 黄如意, 李东鹏, 李巍华. 基于可解释时空图卷积网络的多传感数据融合诊断方法[J]. 机械工程学报, 2024, 60(12): 158-167.
[7]	李响, 付春霖, 雷亚国, 李乃鹏, 杨彬. 保证数据隐私的装备协同智能故障诊断联邦迁移学习方法[J]. 机械工程学报, 2023, 59(6): 1-9.
[8]	林京, 焦金阳. 可解释机械智能诊断技术的研究进展与挑战[J]. 机械工程学报, 2023, 59(20): 215-224.
[9]	刘佳宁, 曹宏瑞, 闫鹏辉, 姬汶辰. 变工况下磨削振纹的深度迁移智能辨识方法[J]. 机械工程学报, 2023, 59(16): 128-136.
[10]	王文东, 李杰, 张俊博, 秦雷, 袁小庆. 基于迁移学习的手部离散动作识别方法研究[J]. 机械工程学报, 2022, 58(7): 12-19.
[11]	宰红斌, 刘建国, 张文岗, 封士永, 祖国强. 基于深度迁移学习的配电线路绝缘子状态监测方法^*[J]. 电气工程学报, 2022, 17(4): 275-281.
[12]	杨淞元, 田勇, 田劲东. 基于iCEEMDAN和迁移学习的锂离子电池SOH估计^*[J]. 电气工程学报, 2022, 17(4): 2-10.
[13]	鲍劲松, 张荣, 李婕, 陆玉前, 彭涛. 面向人-机-环境共融的数字孪生协同技术[J]. 机械工程学报, 2022, 58(18): 103-115.
[14]	邓柯楠, 高栋, 马守东, 路勇. 基于迁移学习的铣削机器人定位误差补偿方法[J]. 机械工程学报, 2022, 58(14): 170-180.
[15]	雷亚国, 杨彬, 李乃鹏, 李响, 武通海. 跨设备的机械故障靶向迁移诊断方法[J]. 机械工程学报, 2022, 58(12): 1-9.