RESEARCH PAPER
A Physics-Constrained Spectral Generative Adversarial Network for Rolling Bearing Fault Diagnosis
1
School of Mathematics, Physics and Statistics, Shanghai University of Engineering Science, China
2
School of Mechanical and Automotive Engineering, Shanghai University of Engineering Science, China
3
Office of Scientific Research, Shanghai University of Medicine & Health Sciences, China
A – Conceptualization; B – Methodology; C – Software; D – Validation; E – Formal analysis; F – Investigation; G – Resources; H – Data curation; I – Writing – original draft; J – Writing – review & editing; K – Visualization; L – Supervision; M – Project administration; N – Funding acquisition
Submission date: 2026-01-10
Final revision date: 2026-04-03
Acceptance date: 2026-06-25
Online publication date: 2026-07-10
Corresponding author
Guoqiang Wang
School of Mathematics, Physics and Statistics, Shanghai University of Engineering Science, China
School of Mathematics, Physics and Statistics, Shanghai University of Engineering Science, China
KEYWORDS
Generative Adversarial NetworksData AugmentationRolling Bearing Fault DiagnosisPhysics-Informed Generative ModelsTime-Frequency Spectral Consistency
TOPICS
ABSTRACT
Accurate bearing fault diagnosis is often hindered by sample scarcity and class imbalance. This paper proposes the Physics-Constrained Spectral Generative Adversarial Network (PCS-GAN), which embeds prior mechanical knowledge into adversarial training to enhance physical interpretability. The PCS-GAN model employs a time-frequency dual-discriminator structure to enforce statistical and spectral realism via physics-informed loss terms. To ensure stability, it integrates WGAN-GP with a curriculum-based loss scheduling strategy. Evaluations on CWRU and MFPT datasets confirm that the PCS-GAN model achieves high spectral integrity while reducing peak GPU memory consumption by 28% compared to Transformer-based architectures. Furthermore, the model exhibits robust noise resilience and elevates macro-F1 scores to over 0.94 in extreme imbalance scenarios. These results demonstrate that the PCS-GAN model provides a computationally efficient and reliable solution for fault diagnosis in data-scarce industrial settings.
REFERENCES (41)
1.
Zhang Y, Su C, Wu J, Liu H, Xie M. Trend-augmented and temporal-featured transformer network with multi-sensor signals for remaining useful life prediction. Reliability Engineering & System Safety 2024; 241: 109662. https://doi.org/10.1016/j.ress....
2.
Chen S, Li CL, Yoder N, Arik SÖ, Pfister T. TSMixer: An all-MLP architecture for time series forecasting. Transactions on Machine Learning Research 2023. https://doi.org/10.48550/arXiv....
3.
Gao Y, Wen Y, Wu J. A neural network-based joint prognostic model for data fusion and remaining useful life prediction. IEEE Transactions on Neural Networks and Learning Systems 2021; 32(1): 117–127. https://doi.org/10.1109/TNNLS.....
4.
Shi J, Zhong J, Zhang Y, Xiao B, Xiao L, Zheng Y. A dual attention LSTM lightweight model based on exponential smoothing for remaining useful life prediction. Reliability Engineering & System Safety 2024; 243: 109821. https://doi.org/10.1016/j.ress....
5.
Zhong J, Jiang H, Gu K, Zhong J, Zhong S. Remaining useful life prediction of rolling bearing based on multi-region hypergraph self-attention network. Mechanical Systems and Signal Processing 2025; 225: 112331. https://doi.org/10.1016/j.ymss....
6.
Wu C, He J, Wang L, Ma C, Liu Y. A dual-path architecture based on time series decomposition and degradation correction for remaining useful life prediction of aero-engine. Computers & Industrial Engineering 2025; 203: 110964. https://doi.org/10.1016/j.cie.....
7.
Zhang J, Tian J, Li M, Leon JI, Franquelo LG, Luo H, Yin S. A parallel hybrid neural network with integration of spatial and temporal features for remaining useful life prediction in prognostics. IEEE Transactions on Instrumentation and Measurement 2023; 72: 1–12. https://doi.org/10.1109/TIM.20....
8.
Yang S, Tang B, Wang W, Yang Q, Hu C. Physics-informed multi-state temporal frequency network for RUL prediction of rolling bearings. Reliability Engineering & System Safety 2024; 242: 109716. https://doi.org/10.1016/j.ress....
9.
Wang Y, Li M, Zheng L, Shi M, Zheng Z, Pei X. Phyformer: A degradation physics-informed self-data driven approach to machinery prognostics. Advanced Engineering Informatics 2024; 62: 102772. https://doi.org/10.1016/j.aei.....
10.
de Beaulieu M, Jha MS, Garnier H, Cerbah F. Remaining useful life prediction based on physics-informed data augmentation. Reliability Engineering & System Safety 2024; 252: 110451. https://doi.org/10.1016/j.ress....
11.
Xiong J, Fink O, Zhou J, Ma Y. Controlled physics-informed data generation for deep learning-based remaining useful life prediction under unseen operation conditions. Mechanical Systems and Signal Processing 2023; 197: 110359. https://doi.org/10.1016/j.ymss....
12.
Chen J, Huang R, Chen Z, Mao W, Li W. Transfer learning algorithms for bearing remaining useful life prediction: A comprehensive review from an industrial application perspective. Mechanical Systems and Signal Processing 2023; 193: 110239. https://doi.org/10.1016/j.ymss....
13.
Lin J, Shao H, Min Z, Luo J, Xiao Y, Yan S, Zhou J. Cross-domain fault diagnosis of bearing using improved semi-supervised meta-learning towards interference of out-of-distribution samples. Knowledge-Based Systems 2022; 252: 109493. https://doi.org/10.1016/j.knos....
14.
Xiang S, Li P, Luo J, Qin Y. Micro transfer learning mechanism for cross-domain equipment RUL prediction. IEEE Transactions on Automation Science and Engineering 2025; 22: 1460–1470. https://doi.org/10.1109/TASE.2....
15.
Li ZJ, Cheng DJ, Zhang HB, Zhou KL, Wang YF. Multi-feature spaces cross adaptation transfer learning-based bearings piece-wise remaining useful life prediction under unseen degradation data. Advanced Engineering Informatics 2024; 60: 102413. https://doi.org/10.1016/j.aei.....
16.
Lyu Y, Wen Z, Chen A. A novel transfer learning approach based on deep degradation feature adaptive alignment for remaining useful life prediction with multi-condition data. Journal of Intelligent Manufacturing 2023; 36(1): 619–637. https://doi.org/10.1007/s10845....
17.
Deng Y, Huang D, Du S, Li G, Zhao C, Lv J. A double-layer attention based adversarial network for partial transfer learning in machinery fault diagnosis. Computers in Industry 2021; 127: 103399. https://doi.org/10.1016/j.comp....
18.
Pan T, Chen J, Liu Z. A meta-weighted network equipped with uncertainty estimations for remaining useful life prediction of turbopump bearings. Expert Systems with Applications 2024; 252: 124161. https://doi.org/10.1016/j.eswa....
19.
Ding P, Xia J, Zhao X, Jia M. Graph structure few-shot prognostics for machinery remaining useful life prediction under variable operating conditions. Advanced Engineering Informatics 2024; 60: 102360. https://doi.org/10.1016/j.aei.....
20.
Chang L, Lin YH. Few-shot remaining useful life prediction based on Bayesian meta-learning with predictive uncertainty calibration. Engineering Applications of Artificial Intelligence 2025; 142: 109980. https://doi.org/10.1016/j.enga....
21.
Wang Z, Jiang P, Chen Z, Li Y, Ren W, Dong L, et al. Remaining useful life prediction method based on two-phase adaptive drift Wiener process. Reliability Engineering & System Safety 2025; 258: 110908. https://doi.org/10.1016/j.ress....
22.
Al-Atroush ME. A deep learning physics-informed neural network (PINN) for predicting drilled shaft axial capacity. Applied Computing and Geosciences 2025; 26: 100246. https://doi.org/10.1016/j.acag....
23.
Mo Y, Li L, Huang B, Li X. Few-shot RUL estimation based on model-agnostic meta-learning. Journal of Intelligent Manufacturing 2022; 34(5): 2359–2372. https://doi.org/10.1007/s10845....
24.
Sun J, Li C, Song Y. Network security situation prediction approach based on MAML and BiGRU. Journal of Intelligent & Fuzzy Systems 2024; 47(3–4): 307–319. https://doi.org/10.3233/JIFS-2....
25.
Nichol A, Achiam J, Schulman J. On first-order meta-learning algorithms. arXiv 2018. https://arxiv.org/abs/1803.029....
26.
Saxena A, Goebel K, Simon D, Eklund N. Damage propagation modeling for aircraft engine run-to-failure simulation. Proceedings of PHM 2008: 1–9. https://doi.org/10.1109/PHM.20....
27.
Costa N, Sánchez L. Variational encoding approach for interpretable assessment of remaining useful life estimation. Reliability Engineering & System Safety 2022; 222: 108353. https://doi.org/10.1016/j.ress....
28.
Tao Z, Zhang C, Xiong J, et al. Evolutionary GRU-CNN for performance degradation prediction of PEMFC. Applied Energy 2023; 336: 120821. https://doi.org/10.1016/j.apen....
29.
Babu GS, Zhao P, Li XL. Deep convolutional neural network based regression approach for estimation of remaining useful life. DASFAA 2016: 214–228. https://doi.org/10.1007/978-3-....
30.
Zheng S, Ristovski K, Farahat A, Gupta C. LSTM network for remaining useful life estimation. PHM 2017: 88–95. https://doi.org/10.1109/ICPHM.....
31.
Zhang Y, Zhang C, Wang S, Dui H, Chen R. Health indicators for remaining useful life prediction based on LSTM and particle filter. Reliability Engineering & System Safety 2024; 241: 109666. https://doi.org/10.1016/j.ress....
32.
Liao X, Chen S, Wen P, Zhao S. Remaining useful life with self-attention assisted physics-informed neural network. Advanced Engineering Informatics 2023; 58: 102195. https://doi.org/10.1016/j.aei.....
33.
Liu Z, Zheng X, Xue A, Ge M. Multi-scale temporal-spatial feature-based hybrid deep neural network for aero-engine RUL prediction. ACS Omega 2024; 9(48): 47410–47427. https://doi.org/10.1021/acsome....
34.
Kim G, Choi JG, Lim S. Transformer-based RUL estimation for turbofan engines. Engineering Applications of Artificial Intelligence 2024; 133: 108475. https://doi.org/10.1016/j.enga....
35.
Wang Y, Liu S, Lv S, Liu G. Few-shot rotating machinery RUL prediction based on Reptile framework and transformer. PHM Beijing 2024: 1–5. https://doi.org/10.1109/PHM-Be....
36.
Liu S, Lv C, Song F, Liu X, Chen D. Remaining useful life prediction integrating working conditions and uncertainty quantification based on multilayer graph neural networks. Journal of the Brazilian Society of Mechanical Sciences and Engineering 2025; 47(2). https://doi.org/10.1007/s40430....
37.
Wang F, Liu A, Qu C, Xiong R, Chen L. Deep-learning method for remaining useful life prediction via dual-attention mechanism. Sensors 2025; 25(2): 497. https://doi.org/10.3390/s25020....
38.
Nectoux P, Gouriveau R, Medjaher K, et al. PRONOSTIA: An experimental platform for bearings accelerated degradation tests. PHM Conference 2012: 1–8. https://doi.org/10.1109/PHM.20....
39.
Zhan X, Liu Z, Yan H, Wu Z, Guo C, Jia X. Health indicator construction and RUL prediction based on deep learning. Eksploatacja i Niezawodnosc 2023; 25(4). https://doi.org/10.17531/ein/1....
40.
Xu J, Sun X, Zhang Z, Zhao G, Lin J. Understanding and improving layer normalization. NeurIPS 2019. https://proceedings.neurips.cc....
| eISSN: | 2956-3860 |
| ISSN: | 1507-2711 |
We process personal data collected when visiting the website. The function of obtaining information about users and their behavior is carried out by voluntarily entered information in forms and saving cookies in end devices. Data, including cookies, are used to provide services, improve the user experience and to analyze the traffic in accordance with the Privacy policy. Data are also collected and processed by Google Analytics tool (more).
You can change cookies settings in your browser. Restricted use of cookies in the browser configuration may affect some functionalities of the website.
You can change cookies settings in your browser. Restricted use of cookies in the browser configuration may affect some functionalities of the website.
