DOI:
https://doi.org/10.64539/sjer.v2i2.2026.413Keywords:
Anomaly Learning, Federated learning, Temporal graph neural networks, Weakly supervised anomaly detection, Bearing prognostics, Leave-one-bearing-outAbstract
Industrial bearing health monitoring is hindered by four interrelated challenges: high class imbalance, the absence of fault-type annotations, stringent data privacy constraints prohibiting centralized aggregation, and non-independent and identically distributed (non-IID) degradation dynamics across geographically dispersed assets. To address these, we propose Fed-TGCN, a novel weakly supervised federated learning framework grounded in temporal graph neural networks. Each client represents a leave-one-bearing-out fold, comprising two training bearings, one validation bearing, and one held-out test bearing, constructs a hybrid spatio-temporal graph from six physics-informed statistical features derived from raw vibration signals; edges encode both sequential dependencies and feature-space similarity via k-nearest neighbors. Pseudo-anomaly labels are generated locally through adaptive thresholding of a degradation score using exponentially weighted moving average, eliminating reliance on expert annotations. Under a strict leave-one-bearing-out protocol on the NASA IMS dataset (12 bearings), local Temporal Graph Convolutional Networks are trained in isolation and aggregated globally via FedAvg. Our method achieves an Average Precision of 0.675 ± 0.276 and Matthews Correlation Coefficient of 0.636 ± 0.285, maintains stronger performance consistency across heterogeneous bearing conditions than isolated and non-graph baselines (ΔMCC = +0.130, p < 0.01). Ablation studies confirm the necessity of temporal modeling (MCC drops by 0.069 without GRU). To the best of our knowledge, this is the first work integrating weakly supervised, graph-based federated learning for bearing prognostics under, demonstrating that parameter coordination but not the data sharing which enables degradation-invariant representation learning across heterogeneous assets.
References
[1] Y. Lei, N. Li, L. Guo, N. Li, T. Yan, and J. Lin, “Machinery health prognostics: A systematic review from data acquisition to RUL prediction,” Mechanical Systems and Signal Processing, vol. 104, pp. 799–834, May 2018, https://doi.org/10.1016/j.ymssp.2017.11.016.
[2] T. Sun, C. Yin, H. Zheng, and Y. Dong, “An unsupervised framework for dynamic health indicator construction and its application in rolling bearing prognostics,” Reliability Engineering & System Safety, vol. 260, pp. 111039–111039, Mar. 2025, https://doi.org/10.1016/j.ress.2025.111039.
[3] X. Li, C. Cheng, and Z. Peng, “Unsupervised construction of health indicator for rotating machinery via multi-criterion feature selection and attentive variational autoencoder,” Science China Technological Sciences, vol. 67, no. 5, pp. 1524–1537, Apr. 2024, https://doi.org/10.1007/s11431-023-2610-4.
[4] T. Li, A. K. Sahu, A. Talwalkar, and V. Smith, “Federated Learning: Challenges, Methods, and Future Directions,” IEEE Signal Processing Magazine, vol. 37, no. 3, pp. 50–60, May 2020, https://doi.org/10.1109/msp.2020.2975749.
[5] S. Wang, Y. Vidal, and F. Pozo, “An unsupervised approach to early fault detection and performance degradation assessment in bearings,” Advanced Engineering Informatics, vol. 68, pp. 103620–103620, Jul. 2025, https://doi.org/10.1016/j.aei.2025.103620.
[6] S. Zhang, F. Ye, B. Wang, and T. Habetler, “Few-Shot Bearing Fault Diagnosis Based on Model-Agnostic Meta-Learning,” IEEE Transactions on Industry Applications, pp. 1–1, 2021, https://doi.org/10.1109/tia.2021.3091958.
[7] A. Deng and B. Hooi, “Graph Neural Network-Based Anomaly Detection in Multivariate Time Series,” Proceedings of the AAAI Conference on Artificial Intelligence, vol. 35, no. 5, pp. 4027–4035, May 2021, https://doi.org/10.1609/aaai.v35i5.16523.
[8] W. Caesarendra, G. Niu, and B.-S. Yang, “Machine condition prognosis based on sequential Monte Carlo method,” Expert Systems with Applications, vol. 37, no. 3, pp. 2412–2420, Mar. 2010, https://doi.org/10.1016/j.eswa.2009.07.014.
[9] R. Zhao, R. Yan, Z. Chen, K. Mao, P. Wang, and R. X. Gao, “Deep learning and its applications to machine health monitoring,” Mechanical Systems and Signal Processing, vol. 115, pp. 213–237, Jan. 2019, https://doi.org/10.1016/j.ymssp.2018.05.050.
[10] An, N. Ho Kim, and J.-H. Choi, “Options for Prognostics Methods: A review of data-driven and physics- based prognostics,” Annual Conference of the PHM Society, vol. 5, no. 1, Oct. 2013, https://doi.org/10.36001/phmconf.2013.v5i1.2184.
[11] P. Jieyang, A. Kimmig, W. Dongkun, Z. Niu, F. Zhi, W. Jiahai, X. Liu, J. Ovtcharova, “A systematic review of data-driven approaches to fault diagnosis and early warning,” Journal of Intelligent Manufacturing, Sep. 2022, https://doi.org/10.1007/s10845-022-02020-0.
[12] H. Wang, J. Wang, Y. Zhao, Q. Liu, M. Liu, and W. Shen, “Few-Shot Learning for Fault Diagnosis With a Dual Graph Neural Network,” IEEE Transactions on Industrial Informatics, vol. 19, no. 2, pp. 1559–1568, Feb. 2023, https://doi.org/10.1109/tii.2022.3205373.
[13] Q. Zhou, W. Ma, Y. Zhang, and J. Guo, “Bearing fault diagnosis for variable working conditions via lightweight transformer and homogeneous generalized contrastive learning with inter-class repulsive discriminant,” Engineering Applications of Artificial Intelligence, vol. 139, pp. 109548–109548, Oct. 2024, https://doi.org/10.1016/j.engappai.2024.109548.
[14] V. Jorry, Z.-S. Duma, Tuomas Sihvonen, Satu-Pia Reinikainen, and Lassi Roininen, “Statistical batch-based bearing fault detection,” Journal of Mathematics in Industry, vol. 15, no. 1, Feb. 2025, https://doi.org/10.1186/s13362-025-00169-w.
[15] Z. Wang, Z. Xu, C. Cai, X. Wang, J. Xu, K. Shi, X. Zhong, Z. Liao, Q. Li, “Rolling bearing fault diagnosis method using time-frequency information integration and multi-scale TransFusion network,” Knowledge-Based Systems, vol. 284, p. 111344, Jan. 2024, https://doi.org/10.1016/j.knosys.2023.111344.
[16] A. A. Soomro, M. B. Muhammad, A. A. Mokhtar, M. H. M. Saad, N. Lashari, M. Hussain, U. Sarwar, A. S. Palli, “Insights into modern machine learning approaches for bearing fault classification: A systematic literature review,” Results in Engineering, vol. 23, p. 102700, Sep. 2024, https://doi.org/10.1016/j.rineng.2024.102700.
[17] T. N. Kipf and M. Welling, “Semi-Supervised Classification with Graph Convolutional Networks,” ArXiv (Cornell University), Jan. 2016, https://doi.org/10.48550/arxiv.1609.02907.
[18] Z. Wu, S. Pan, F. Chen, G. Long, C. Zhang, and P. S. Yu, “A Comprehensive Survey on Graph Neural Networks,” IEEE Transactions on Neural Networks and Learning Systems, vol. 32, no. 1, pp. 1–21, 2020, https://doi.org/10.1109/TNNLS.2020.2978386.
[19] M. Wang, J. Yu, H. Leng, X. Du, and Y. Liu, “Bearing fault detection by using graph autoencoder and ensemble learning,” Scientific Reports, vol. 14, no. 1, p. 5206, Mar. 2024, https://doi.org/10.1038/s41598-024-55620-6.
[20] P. Kairouz and H. B. McMahan, “Advances and Open Problems in Federated Learning,” Foundations and Trends® in Machine Learning, vol. 14, no. 1, 2021, https://doi.org/10.1561/2200000083.
[21] W. Song, D. Wu, W. Shen, and B. Boulet, “Early fault detection for rolling bearings: A meta‐learning approach,” IET Collaborative Intelligent Manufacturing, vol. 6, no. 2, May 2024, https://doi.org/10.1049/cim2.12103.
[22] R. B. Randall and J. Antoni, “Rolling element bearing diagnostics—A tutorial,” Mechanical Systems and Signal Processing, vol. 25, no. 2, pp. 485–520, Feb. 2011, https://doi.org/10.1016/j.ymssp.2010.07.017.
[23] J. Ji and H. Dong, “Spatio-Temporal Graph Convolutional Networks for Traffic Prediction Considering Multiple Spatio-Temporal Information,” 2024 20th International Conference on Mobility, Sensing and Networking (MSN), pp. 730–737, Dec. 2024, https://doi.org/10.1109/msn63567.2024.00103.
[24] C. Feng, C. Liu, and D. Jiang, “Unsupervised anomaly detection using graph neural networks integrated with physical-statistical feature fusion and local-global learning,” Renewable Energy, vol. 206, pp. 309–323, Apr. 2023, https://doi.org/10.1016/j.renene.2023.02.053.
[25] J. Wang, B. Liang, Z. Zhu, E. Thepie Fapi, and H. Dalal, “Communication-Efficient Network Topology in Decentralized Learning: A Joint Design of Consensus Matrix and Resource Allocation,” IEEE Transactions on Networking, vol. 33, no. 2, pp. 761–776, Apr. 2025, https://doi.org/10.1109/tnet.2024.3511333.
[26] J. Lee, H. Qiu, G. Yu, J. Lin, and Rexnord Technical Services, "Bearing data set," NASA Ames Prognostics Data Repository, NASA Ames Research Center, Moffett Field, CA, USA, 2007.
[27] G. Li, M. Wei, D. Wu, Y. Cheng, J. Wu, and J. Yan, “Zero-Sample fault diagnosis of rolling bearings via fault spectrum knowledge and autonomous contrastive learning,” Expert Systems with Applications, vol. 275, p. 127080, May 2025, https://doi.org/10.1016/j.eswa.2025.127080.
[28] Yasir Saleem Afridi, L. Hasan, R. Ullah, Z. Ahmad, and J.-M. Kim, “LSTM-Based Condition Monitoring and Fault Prognostics of Rolling Element Bearings Using Raw Vibrational Data,” Machines, vol. 11, no. 5, pp. 531–531, May 2023, https://doi.org/10.3390/machines11050531.
[29] M. Xu, P. Guan, X. Shi, R. Jiang, J. Tian, J. Geng, G. Xiong, “Research on Bearing Fault Diagnosis Methods Based on Various Convolutional Neural Network Architectures,” IEEE Access, vol. 13, pp. 44445–44465, 2025, https://doi.org/10.1109/access.2025.3548693.
[30] F. T. Liu, K. M. Ting, and Z.-H. Zhou, “Isolation Forest,” 2008 Eighth IEEE International Conference on Data Mining, Dec. 2008, https://doi.org/10.1109/icdm.2008.17.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 Khagendra Darlami, Lalit Awasthi
This work is licensed under a Creative Commons Attribution 4.0 International License.
