A self-supervised learning framework integrating prior knowledge and simulation data for rolling bearing fault diagnosis with few-shot and imbalanced data
- 12 May 2026
- journal article
- Published by IOP Publishing in Measurement Science and Technology
- Vol. 37 (21)
Abstract
The compound problem of few-shot and data imbalance remains a major obstacle in rolling bearing fault diagnosis. To address this issue, this paper proposes IMSSiT-SSL, a self-supervised learning framework that integrates prior knowledge with simulation data-driven approaches. Unlabeled simulation datasets are used for self-supervised pre-training through two physics-informed pretext tasks: a kurtosis-guided masked reconstruction task that selectively masks high-kurtosis segments to force the model to capture intrinsic fault features, and a characteristic indicator-based pseudo-label prediction task that explicitly embeds domain expert knowledge into the representation learning process. This dual-task pre-training strategy alleviates the few-shot challenge while directing the model's attention toward fault-relevant impact segments, improving the physical interpretability of feature extraction. Subsequently, a small amount of labeled data is used for fine-tuning, where an adaptive boundary-regularized dual focal loss (ABR-DFL) is introduced to encourage the model to focus on discriminating minority class samples, thus mitigating the negative effects of imbalanced data. Experimental results demonstrate that the proposed method achieves G-mean improvements of 3.07%–20.62% over competing methods across various imbalance ratios, and maintains a diagnostic accuracy of 98.78% under cross-speed few-shot conditions with only 30 labeled samples per class.Keywords
Funding Information
- Local Capacity Construction Project of Shanghai Municipal Science and Technology Commission (20030501000)
This publication has 29 references indexed in Scilit:
- Prior knowledge-informed multi-task dynamic learning for few-shot machinery fault diagnosisExpert Systems with Applications, 2025
- A prior knowledge-enhanced self-supervised learning framework using time-frequency invariance for machinery intelligent fault diagnosis with small samplesEngineering Applications of Artificial Intelligence, 2024
- TFPred: Learning discriminative representations from unlabeled data for few-label rotating machinery fault diagnosisControl Engineering Practice, 2024
- Extended attention signal transformer with adaptive class imbalance loss for Long-tailed intelligent fault diagnosis of rotating machineryAdvanced Engineering Informatics, 2024
- A self-attention based contrastive learning method for bearing fault diagnosisExpert Systems with Applications, 2023
- Self-supervised Learning: A Succinct ReviewArchives of Computational Methods in Engineering, 2023
- Self-supervised signal representation learning for machinery fault diagnosis under limited annotation dataKnowledge-Based Systems, 2022
- Intelligent fault diagnosis of machines with small & imbalanced data: A state-of-the-art review and possible extensionsISA Transactions, 2022
- Fatigue analysis of the gearbox housing in high-speed trains under wheel polygonization using a multibody dynamics algorithmEngineering Failure Analysis, 2019
- A novel adversarial learning framework in deep convolutional neural network for intelligent diagnosis of mechanical faultsKnowledge-Based Systems, 2019