Abstract：Industrial motor fault detection faces significant challenges in identifying subtle failure features from sequential data and reducing reliance on expert knowledge. Traditional methods like principal component analysis (PCA) and autoencoders often struggle with limited fault sensitivity and require extensive labeled data for training. This study proposes a residual inner product-driven fault detection model (RID-FDM). This fully data-driven framework achieves high-precision detection of three critical motor faults—localized bearing damage, stator inter-turn short circuits, and rotor bar breakage—using only healthy data for training. By leveraging sequence windowing to partition long-term temporal data into localized fragments, the model introduces a novel residual inner product mechanism to amplify fault-sensitive features while maintaining robustness to normal operational variations. The core innovations of RID-FDM lie in its three-tiered technical framework. First, the residual inner product analysis (RIPA) framework employs a segment residual generator (SRG) to produce orthogonal residual representations of input fragments, enforced by a global residual inner product (GRIP) loss function that drives the residual inner product sequence (RIPS) of healthy data toward zero. Second, the windowed residual propagation mechanism decomposes long sequences into localized segments, enabling the convolutional SRG to capture spatial-temporal fault patterns while accumulating abnormal evidence through the temporal evolution of RIPS. Third, the model demonstrates universal adaptability to diverse fault types by leveraging the generic feature representation capacity of residual inner products, requiring hyperparameter adjustments rather than architectural redesign. This design significantly enhances engineering applicability in industrial settings with limited fault samples and stringent real-time requirements. Extensive experiments validate RID-FDM's superiority across benchmark datasets. The model achieves 100% detection accuracy for bearing faults, 96.07% for stator faults, and 93.33% for rotor faults in zero-shot scenarios. Notably, the method does not require prior knowledge of fault during training and operates in real time with minimal computational overhead. The proposed framework offers a scalable solution for predictive maintenance systems where labeled fault data is scarce and rapid response is critical.
[1] 迟连强, 张殿海, 赵俊清, 等. 旋转电机轴承电蚀损伤机理与缓解措施研究进展[J]. 电工技术学报, 2024, 39(20): 6409-6430. Chi Lianqiang, Zhang Dianhai, Zhao Junqing, et al.Research progress on the mechanism and mitigation measure of electrical corrosion damage in rotating motor bearings[J]. Transactions of China Electrotechnical Society, 2024, 39(20): 6409-6430. [2] 宋向金, 孙文举, 刘国海, 等. 深度子领域自适应网络电机滚动轴承跨工况故障诊断[J]. 电工技术学报, 2024, 39(1): 182-193. Song Xiangjin, Sun Wenju, Liu Guohai, et al.Across working conditions fault diagnosis for motor rolling bearing based on deep subdomain adaption network[J]. Transactions of China Electrotechnical Society, 2024, 39(1): 182-193. [3] Liu Zepeng, Zhang Long.A review of failure modes, condition monitoring and fault diagnosis methods for large-scale wind turbine bearings[J]. Measurement, 2020, 149: 107002. [4] Yan Ruqiang, Shang Zuogang, Xu Hong, et al.Wavelet transform for rotary machine fault diagnosis: 10 years revisited[J]. Mechanical Systems and Signal Processing, 2023, 200: 110545. [5] 张辉, 戈宝军, 韩斌, 等. 基于GAF-CapsNet的电机轴承故障诊断方法[J]. 电工技术学报, 2023, 38(10): 2675-2685. Zhang Hui, Ge Baojun, Han Bin, et al.Fault diagnosis method of motor bearing based on GAF-CapsNet[J]. Transactions of China Electrotechnical Society, 2023, 38(10): 2675-2685. [6] 姚鹏. 人工智能技术在交流异步电机故障诊断中的应用[J]. 电机与控制应用, 2022, 49(4): 1-9. Yao Peng.Application of artificial intelligence technology in fault diagnosis of AC asynchronous motor[J]. Electric Machines & Control Application, 2022, 49(4): 1-9. [7] 樊红卫, 孟瑾, 任众孚, 等. 一种变转速电机转子-轴承系统故障智能诊断方法[J]. 电机与控制学报, 2024, 28(11): 195-210. Fan Hongwei, Meng Jin, Ren Zhongfu, et al.Intelligent fault diagnosis method of motor rotorbearing system under variable speed conditions[J]. Electric Machines and Control, 2024, 28(11): 195-210. [8] Li Jun, Liu Yongbao, Li Qijie.Intelligent fault diagnosis of rolling bearings under imbalanced data conditions using attention-based deep learning method[J]. Measurement, 2022, 189: 110500. [9] Gewers F L, Ferreira G R, De Arruda H F, et al. Principal component analysis[J]. ACM Computing Surveys, 2022, 54(4): 1-34. [10] Dong Wei, Woźniak M, Wu Junsheng, et al.Denoising aggregation of graph neural networks by using principal component analysis[J]. IEEE Transactions on Industrial Informatics, 2023, 19(3): 2385-2394. [11] 詹诗语, 高国强, 曹炳磊, 等. 基于深度学习的局部放电分类识别研究综述[J]. 电气工程学报, 2025, 20(2): 371-382. Zhan Shiyu, Gao Guoqiang, Cao Binglei, et al.Review on classification and recognition of partial discharge rest on deep learning[J]. Journal of Electrical Engineering, 2025, 20(2): 371-382. [12] 卢太武, 马洪波, 王先芝, 等. 时变工况下基于精细复合多尺度散度熵的旋转机械故障诊断方法[J]. 振动与冲击, 2023, 42(21): 211-218. Lu Taiwu, Ma Hongbo, Wang Xianzhi, et al.Fault diagnosis method for rotating machinery based on fine composite multi-scale divergence entropy under time-varying working conditions[J]. Journal of Vibration and Shock, 2023, 42(21): 211-218. [13] 叶旭, 杨炯, 丘志力, 等. 结合拉曼光谱主成分分析-线性判别进行蛇纹石玉产地溯源的探索[J]. 光谱学与光谱分析, 2024, 44(9): 2551-2558. Ye Xu, Yang Jiong, Qiu Zhili, et al.An exploration of geographic determination of serpentine jade by Raman spectroscopy combined with principal component analysis and linear discriminant analysis[J]. Spectroscopy and Spectral Analysis, 2024, 44(9): 2551-2558. [14] Zhu Fa, Gao Junbin, Yang Jian, et al.Neighborhood linear discriminant analysis[J]. Pattern Recognition, 2022, 123: 108422. [15] Amin M T, Khan F, Ahmed S, et al.A data-driven Bayesian network learning method for process fault diagnosis[J]. Process Safety and Environmental Protection, 2021, 150: 110-122. [16] Wei Jihong, Chammam A, Feng Jianqin, et al.Power system monitoring for electrical disturbances in wide network using machine learning[J]. Sustainable Computing: Informatics and Systems, 2024, 42: 100959. [17] Jin Xiaohang, Fan Jicong, Chow T W S. Fault detection for rolling-element bearings using multivariate statistical process control methods[J]. IEEE Transactions on Instrumentation and Measurement, 2019, 68(9): 3128-3136. [18] Choudhary A, Goyal D, Letha S S.Infrared thermography-based fault diagnosis of induction motor bearings using machine learning[J]. IEEE Sensors Journal, 2021, 21(2): 1727-1734. [19] Wang Zhile, Yang Jianhua, Guo Yu.Unknown fault feature extraction of rolling bearings under variable speed conditions based on statistical complexity measures[J]. Mechanical Systems and Signal Processing, 2022, 172: 108964. [20] Jia Linshan, Chow T W S, Yuan Yixuan. GTFE-Net: a gramian time frequency enhancement CNN for bearing fault diagnosis[J]. Engineering Applications of Artificial Intelligence, 2023, 119: 105794. [21] Li Pengzhi, Pei Yan, Li Jianqiang.A comprehensive survey on design and application of autoencoder in deep learning[J]. Applied Soft Computing, 2023, 138: 110176. [22] Yan Shen, Shao Haidong, Min Zhishan, et al.FGDAE: a new machinery anomaly detection method towards complex operating conditions[J]. Reliability Engineering & System Safety, 2023, 236: 109319. [23] 李兵, 梁舒奇, 单万宁, 等. 基于改进正余弦算法优化堆叠降噪自动编码器的电机轴承故障诊断[J]. 电工技术学报, 2022, 37(16): 4084-4093. Li Bing, Liang Shuqi, Shan Wanning, et al.Motor bearing fault diagnosis based on improved sine and cosine algorithm for stacked denoising autoencoders[J]. Transactions of China Electrotechnical Society, 2022, 37(16): 4084-4093. [24] 雷蕾潇, 何怡刚, 姚其新, 等. 基于变权属性矩阵的变压器零样本故障诊断技术[J]. 电工技术学报, 2024, 39(20): 6577-6590. Lei Leixiao, He Yigang, Yao Qixin, et al.Zero-shot fault diagnosis technique of transformer based on weighted attribute matrix[J]. Transactions of China Electrotechnical Society, 2024, 39(20): 6577-6590. [25] Zhang Jiusi, Zhang Ke, An Yiyao, et al.An integrated multitasking intelligent bearing fault diagnosis scheme based on representation learning under imbalanced sample condition[J]. IEEE Transactions on Neural Networks and Learning Systems, 2024, 35(5): 6231-6242. [26] Yu Jianbo, Zhou Xingkang.One-dimensional residual convolutional autoencoder based feature learning for gearbox fault diagnosis[J]. IEEE Transactions on Industrial Informatics, 2020, 16(10): 6347-6358. [27] 梁郑秋, 郝亮亮, 周艳真, 等. 基于卷积神经网络的核电多相无刷励磁系统旋转整流器故障诊断[J]. 电工技术学报, 2023, 38(20): 5458-5472. Liang Zhengqiu, Hao Liangliang, Zhou Yanzhen, et al.Fault diagnosis of rotating rectifier in nuclear multiphase brushless excitation system based on convolutional neural network[J]. Transactions of China Electrotechnical Society, 2023, 38(20): 5458-5472. [28] Sonoda S, Murata N.Neural network with unbounded activation functions is universal approximator[J]. Applied and Computational Harmonic Analysis, 2017, 43(2): 233-268. [29] Yarotsky D.Error bounds for approximations with deep ReLU networks[J]. Neural Networks, 2017, 94: 103-114.
2026, Vol. 41 
