Skip to content
Open access

Physics-guided conditional diffusion model for small-sample and class-imbalanced bearing fault diagnosis

Oct 2026 · Journal of Vibration and Control · 0 citations · 33 references

TL;DR

A physics-guided conditional diffusion model (PG-CDM) is proposed to generate high-quality fault data under small-sample conditions and higher fault-related feature fidelity is achieved compared with mainstream generative models.

Abstract

Deep learning has become a mainstream approach for fault diagnosis. However, its application in industrial settings remains constrained by data scarcity. Although generative models have been introduced for data augmentation, challenges such as mode collapse and fault-frequency-band drift are still observed in existing methods. To address these issues, a physics-guided conditional diffusion model (PG-CDM) is proposed to generate high-quality fault data under small-sample conditions. Specifically, a dual-path physics guidance (DPPG) strategy is introduced at the input stage of the denoising U-Net. A theory-derived fault-band mask is concatenated with the noisy time-frequency map. In this way, the theoretical fault-frequency regions are anchored on the time-frequency plane, thereby improving consistency with bearing kinematics. In addition, a physics-constrained adaptive group normalization (PC-AdaGN) module is introduced during feature propagation. Physical parameters, such as the theoretical fundamental frequency, are mapped into a prior vector. The prior vector is fused with latent features to modulate affine transformations in residual blocks. This design aligns the energy distribution of weak periodic impulsive components with the expected bearing kinematics. Furthermore, a physics-aware coordinate attention (PA-CA) mechanism is integrated into the multi-scale encoding-decoding pathway. The model is guided to focus on critical fault-frequency regions, and weak periodic impulsive signatures are reconstructed. Experiments are conducted on the laboratory dataset, and the effectiveness of PG-CDM is validated. Fault-frequency-band drift in generated samples is effectively suppressed by PG-CDM. Higher fault-related feature fidelity is achieved compared with mainstream generative models. Under extreme class-imbalance ratios, classification accuracy above 90% is achieved. Overall, this study provides a physics-guided data augmentation framework for industrial fault diagnosis under small-sample and class-imbalanced conditions.

Read PDF

Similar papers

#reinforcement learning Open access Sep 2026

Multi-scale physics-informed reinforcement learning method for bearing fault diagnosis under data scarcity

A novel multi-scale physics-informed neural network with enhanced reinforcement learning (MPINN-ERL) for bearing fault diagnosis under data scarcity and achieves competitive diagnostic accuracy.

Xin Xia, Xi-Hong Fei, Kang Wang et al. · 0 citations
Open access Oct 2026

Physics-Frequency-Guided Bidirectional Mamba Network for Cross-Domain Few-Shot Bearing Fault Diagnosis

A multi-scale physics-frequency bidirectional Mamba network for cross-sampling-rate, cross-device, and few-shot bearing fault diagnosis, which ranks first among the compared methods on the three-class cross-device few-shot task and remains competitive on the remaining tasks, while requiring only 3.71M parameters.

Su-Hui Feng, Lin Li, Qian Lv · 0 citations
Open access Oct 2026

Physics-Guided Representation Learning and Selective Domain Adaptation for Cross-Condition Bearing Fault Diagnosis

Cross-condition bearing fault diagnosis remains challenging because changes in operating conditions can cause distribution shifts that degrade diagnostic performance. Moreover, domain adaptation is not always beneficial and may introduce negative transfer and unnecessary computational cost. This paper proposes a Physic...

Syed Sajjad Haider Zaidi, Alex Shenfield, Hong-Wei Zhang et al. · 0 citations
#diffusion models Conference Oct 2026

A ResUNet diffusion adversarial generative network for small-sample bearing fault diagnosis

Rolling bearing fault diagnosis under limited-sample conditions often suffers from insufficient training data. To address this issue, a ResUNet Diffusion Adversarial Generative Network (RUDAGN) is proposed. First, vibration signals are transformed into time–frequency representations using Continuous Wavelet Transform (...

Hao-Tian Geng, Zhi-Gang Wei · 0 citations
Open access Oct 2026

Physics driven incremental meta-learning rolling bearing fault diagnosis under few-shot emerging conditions

Traditional deep learning fault diagnosis methods rely on sufficiently labeled training data, which limits their ability to adapt to the continual emergence of new fault data. Although continual learning methods can alleviate catastrophic forgetting to some extent, the model’s diagnostic performance is still significan...

Dong-Ming Xiao, Heng-Shu Chen, Jia-Kai Ding et al. · 0 citations
Oct 2026

Physics-informed hierarchal graph neural network with multi-scale feature extraction for bearing fault diagnosis

In modern industry, bearing fault diagnosis is of crucial significance for ensuring the reliability and safety of rotating machinery. With the widespread deployment of the Industrial Internet of Things (IIoT), massive amounts of sensor data have become available, enabling the extensive application of end-to-end deep...

Yu-Huang Jia, Xin-Hui Shao, Jiang-Hua Liu et al. · 0 citations

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.