[1]毛文涛,钞龙,张子怡,等.高速电驱轴承早期故障预警的流数据鲁棒异常检测[J].郑州大学学报(工学版),2027,48(XX):1-10.[doi:10.13705/j.issn.1671-6833.2026.03.013]
　MAO Wentao,CHAO Long,ZHANG Ziyi,et al.Robust Anomaly Detection with Streaming Data for Incipient Fault Alarm of High-speed Electrical-Driven Bearing[J].Journal of Zhengzhou University (Engineering Science),2027,48(XX):1-10.[doi:10.13705/j.issn.1671-6833.2026.03.013]
点击复制

高速电驱轴承早期故障预警的流数据鲁棒异常检测()
分享到：

《郑州大学学报(工学版)》[ISSN:1671-6833/CN:41-1339/T]

卷:
48
期数:
2027年XX
页码:
1-10
栏目:
出版日期:
2027-12-10

文章信息/Info

Title:
Robust Anomaly Detection with Streaming Data for Incipient Fault Alarm of High-speed Electrical-Driven Bearing
作者:
毛文涛1,2, 钞龙1, 张子怡1, 邵燚博1, 仲志丹3
1. 河南师范大学 计算机与信息工程学院(人工智能学院) ,河南 新乡 453007;2. 智慧商务与物联网技术” 河南省工程实验室(河南师范大学) ,河南 新乡 453007;3. 河南科技大学 机电工程学院,河南 洛阳 471023
Author(s):
MAO Wentao1,2, CHAO Long1, ZHANG Ziyi1, SHAO Yibo1, ZHONG Zhidan3
1. School of Computer and Information Engineering ( School of Artifical Intelligence) , Henan Normal University, Xinxiang 453007, China; 2. Engineering Laboratory of Intelligence Business and Internet of Things, Xinxiang 453007, China; 3. School of Mechatronics Engineering, Henan University of Science and Technology, Luoyang 471023, China
关键词:
概念漂移;  流数据;  异常检测;  对比学习;  早期故障检测
Keywords:
concept drift;  streaming data;  anomaly detection;  contrastive learning;  early fault detection
分类号:
TP183 TH165+.3
DOI:
10.13705/j.issn.1671-6833.2026.03.013
文献标志码:
A
摘要:
针对新能源汽车高速电驱轴承在急变转速工况下振动信号呈现“阶梯状”分布变化、引发概念漂移，导致传统检测方法易误报的问题，本文提出一种融合概念漂移感知的流数据鲁棒异常检测方法。首先，构建结合对比学习与张量分解的预训练方法提取公共特征表征，进而构建基于概念漂移感知的深度支持向量数据描述模型对流数据进行快速微调，动态计算局部偏移得分，并设计基于滑动窗口和核密度估计的流数据分布感知机制计算概念漂移得分；合并两种得分判断是否需要更新模型，最终完成早期故障状态的识别。在自建新能源汽车电驱轴承退化寿命试验台采集的急加速试验数据上进行验证，结果表明所提方法能在变工况环境下有效排除概念漂移点，并准确识别真正早期故障，在误报警率保持为0的情况下较上位机报警提前10个样本点。
Abstract:
The vibration signals of high-speed electrical-driven bearings under rapidly-varying rotational speed are characterized by stepwise variations in their statistical characteristics, leading to concept drift in the data distribution. Current anomaly detection methods generally rely on static independent and identically distributed (i. i. d.) assumptions, but still struggle to recognize concept drift well, which further results in false alarms. To address these challenges, a concept drift-aware robust anomaly detection method with streaming data is proposed in this paper. First, an anomaly detection pre-training mechanism based on contrastive learning and tensor decomposition is designed to produce high-quality initial features with both generalization and discriminative capability. Second, a new concept drift-aware deep support vector data description (Deep SVDD) model is constructed to enable rapid fine-tuning of streaming data, while calculating the local deviation scores under hyper-sphere constraint. A distribution-aware mechanism using sliding windows and kernel density estimation (KDE) is also integrated to calculate concept drift scores. Finally, these two scores are evaluated together to determine whether the model update under new data distribution is required, with early fault occurrence precisely recognized. Experimental validation on our high-speed bearing testbed under varying operating conditions demonstrates that concept drift points can be accurately filtered out with real early fault identified. The proposed method provides an advance warning of 10 samples compared to the supervisory alarm, while maintaining a zero false alarm rate.

参考文献/References:

[1] Schneider V, Behrendt C, Höltje P, et al. Electrical bearing damage, a problem in the nano- and macro-range[J]. Lubricants, 2022, 10(8): 194-
[2] Li Yongbo, Xu Minqiang, Liang Xihui, et al. Application of bandwidth EMD and adaptive multiscale morphology analysis for incipient fault diagnosis of rolling bearings[J]. IEEE Transactions on Industrial Electronics, 2017, 64(8): 6506-6517.
[3] Liu Yang, Li Lingjun, Wang Yu, et al. Fault feature extraction method of rolling bearings based on FIF-CYCBD[J]. Journal of Zhengzhou University (Engineering Science), 2022, 43(4): 35-40. [刘洋, 李凌均, 王宇, 等. 基于FIF-CYCBD的滚动轴承故障特征提取方法研究[J]. 郑州大学学报(工学版), 2022, 43(4): 35-40.]
[4] Khelif M A, Bendiabdelah A, Cherif B D E. A combined RMS-MEAN value approach for an inverter open-circuit fault detection[J]. Periodica Polytechnica Electrical Engineering and Computer Science, 2019, 63(3): 169-177.
[5] Pan Yongke, He Ziping, Xia Kewen, et al. Improved co-training semi-supervised SVM and its application in oil layer recognition[J]. Journal of Zhengzhou University (Engineering Science), 2022, 43(1): 14-19. [潘用科, 贺紫平, 夏克文, 等. 改进的协同训练半监督SVM在油层识别中的应用[J]. 郑州大学学报(工学版), 2022, 43(1): 14-19.]
[6] Kopp M, Pevný T, Holeňa M. Anomaly explanation with random forests[J]. Expert Systems with Applications, 2020, 149: 113187.
[7] Mao Wentao, Chen Zongtao, Zhang Yanna, et al. Tensor-DAAD: when tensor meets online early fault detection with transfer learning[J]. Measurement, 2023, 208: 112478.
[8] Hayashi T, Cimr D, Studnička F, et al. Distance-based one-class time-series classification approach using local cluster balance[J]. Expert Systems with Applications, 2024, 235: 121201.
[9] Jana D, Patil J, Herkal S, et al. CNN and Convolutional Autoencoder (CAE) based real-time sensor fault detection, localization, and correction[J]. Mechanical Systems and Signal Processing, 2022, 169: 108723.
[10] Ruff L, Vandermeulen R A, Goernitz N, et al. Deep one-class classification[C]//Proceedings of the International Conference on Machine Learning. Stockholm: PMLR, 2018: 4393-4402.
[11] Zhang Jian, Guo Lili, Gao Song, et al. Bayes-optimized adaptive growing neural gas method for online anomaly detection of industrial streaming data[J]. Applied Sciences, 2024, 14(10): 4139.
[12] Lin Y M, Kang Jialin, Yao Yuan. Handling domain drift and unknown fault detection in rotating machinery using few-shot learning with data scaling[J]. International Journal of Adaptive Control and Signal Processing, 2025, 39(7): 1591-1607.
[13] Hu Chenye, Wu Jingyao, Sun Chuang, et al. Unified flowing normality learning for rotating machinery anomaly detection in continuous time-varying conditions[J]. IEEE Transactions on Cybernetics, 2025, 55(1): 221-233.
[14] Guo Husheng, Li Hai, Ren Qiaoyan, et al. Concept drift type identification based on multi-sliding windows[J]. Information Sciences, 2022, 585: 1-23.
[15] Zhang Lingyu, Zhao Jiabao, Li Wei. Online and unsupervised anomaly detection for streaming data using an array of sliding windows and PDDs[J]. IEEE Transactions on Cybernetics, 2021, 51(4): 2284-2289.
[16] Grill J B, Strub F, Altché F, et al. Bootstrap your own latent a new approach to self-supervised learning[C]//Proceedings of the 34th International Conference on Neural Information Processing Systems. New York: ACM, 2020: 21271-21284.
[17] Shi Qiquan, Yin Jiaming, Cai Jiajun, et al. Block Hankel tensor ARIMA for multiple short time series forecasting[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2020, 34(4): 5758-5766.
[18] Nectoux P, Gouriveau R, Medjaher K, et al. PRONOSTIA: an experimental platform for bearings accelerated degradation tests[C]//Proceedings of the IEEE International Conference on Prognostics and Health Management (PHM). Piscataway: IEEE, 2012: 1-8.
[19] Wang Biao, Lei Yaguo, Li Naipeng, et al. A hybrid prognostics approach for estimating remaining useful life of rolling element bearings[J]. IEEE Transactions on Reliability, 2020, 69(1): 401-412.
[20] Lu Weining, Li Yipeng, Cheng Yu, et al. Early fault detection approach with deep architectures[J]. IEEE Transactions on Instrumentation and Measurement, 2018, 67(7): 1679-1689.
[21] Guo Xiaoping, Liu Shiyang, Li Yuan. Fault detection of multi-mode processes employing sparse residual distance[J]. Acta Automatica Sinica, 2019, 45(3): 617-625. [郭小萍, 刘诗洋, 李元. 基于稀疏残差距离的多工况过程故障检测方法研究[J]. 自动化学报, 2019, 45(3): 617-625.]
[22] Yoon S, Lee Y, Lee J G, et al. Adaptive model pooling for online deep anomaly detection from a complex evolving data stream[C]//Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining. New York: ACM, 2022: 2347-2357.

相似文献/References:

[1]白洋,王志海,孙艳歌.基于图的概念重现发现与预测[J].郑州大学学报(工学版),2017,38(4):57.[doi:10.13705/j.issn.1671-6833.2017.01.021]
　Baiyang,Wang Zhihai,Sun Yange.Recurring Concept Detection and Prediction Based on the Graph[J].Journal of Zhengzhou University (Engineering Science),2017,38(XX):57.[doi:10.13705/j.issn.1671-6833.2017.01.021]

备注/Memo

备注/Memo:
收稿日期:2025-10-10;修订日期:2025-11-21
基金项目:国家自然科学基金资助项目(62472146) ;河南省科技研发计划联合基金( 产业类) 资助项目( 225101610001) ;河南师范大学研究生科研与实践创新项目( YZ202504)
作者简介:毛文涛(1980— ) ,男,河南新乡人,河南师范大学教授,博士,博士生导师,主要从事机器学习、智能故障预测及工业大数据分析等方面的研究,E-mail:maowt@ htu. edu. cn。
更新日期/Last Update: 2026-04-17