A Physics-Constrained Spectral Generative Adversarial Network for Rolling Bearing Fault Diagnosis
Eksploatacja i Niezawodność – Maintenance and Reliability · 10.07.2026
DOI: 10.17531/ein/225075 ↗Kurz gefasst
Fachartikel in der Zeitschrift Eksploatacja i Niezawodność – Maintenance and Reliability, online erschienen am 10.07.2026. Verfasst von Yitong Liu und 3 weiteren Autor:innen. Die Arbeit nennt 41 Literaturangaben.
Abstract (Originalsprache)
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.
Zitieren
APA
Liu, Y., Liu, X., Wang, G., & Zhang, X. (2026). A Physics-Constrained Spectral Generative Adversarial Network for Rolling Bearing Fault Diagnosis. Eksploatacja i Niezawodność – Maintenance and Reliability. https://doi.org/10.17531/ein/225075
Vancouver
Liu Y, Liu X, Wang G, Zhang X. A Physics-Constrained Spectral Generative Adversarial Network for Rolling Bearing Fault Diagnosis. Eksploatacja i Niezawodność – Maintenance and Reliability. 2026. doi:10.17531/ein/225075
BibTeX
@article{Liu2026,
author = {Liu, Yitong and Liu, Xintian and Wang, Guoqiang and Zhang, Xia},
title = {A Physics-Constrained Spectral Generative Adversarial Network for Rolling Bearing Fault Diagnosis},
journal = {Eksploatacja i Niezawodność – Maintenance and Reliability},
year = {2026},
doi = {10.17531/ein/225075}
}
EUVIMED