Unexpected breakdowns in rotating machinery can result in substantial economic losses and safety hazards, underscoring the critical need for reliable bearing fault diagnosis under variable operating conditions. Although deep learning has demonstrated a strong diagnostic capability, its practical deployment is often hindered by domain shifts caused by variations in speed or load, along with the difficulty of obtaining labeled data for all possible operating conditions. To address this challenge, a physics-informed single-source domain generalization (PI-SSDG) framework is proposed that learns from vibration signals collected from a single-source domain corresponding to a specific operating condition and effectively generalizes to unseen conditions without requiring target-domain data. The proposed method introduces a log-mean-removed, low-pass liftered cepstrum that explicitly suppresses global amplitude scaling induced by operating conditions and excitation-related periodicities while preserving fault-discriminative transfer-function-related features. In addition, a dual-branch architecture is designed that jointly exploits the preprocessed cepstrum and raw vibration signals, together with a joint training strategy using branch-specific losses that promotes complementary feature representations and mitigates residual domain variability. Extensive experiments conducted on three public bearing datasets demonstrate that the proposed method consistently achieves superior cross-domain accuracy under speed and load variations and outperforms recent domain-generalization baselines. Comprehensive ablation studies and feature visualization further confirm the effectiveness of the proposed components in improving diagnostic robustness.
Fault diagnosis of rolling bearings is crucial for operational safety. However, the scarcity of labeled data and significant domain shifts are two key challenges. Existing studies neglect robustness to physical disturbances and the interpretability of diagnostic decisions. To address these issues, this paper proposes a physics-informed cross-domain fault diagnosis framework. First, based on bearing fault modulation mechanisms, multi-domain features are extracted. A physics-regularized feature selection strategy combining Sparse Group Lasso with stability selection and Random Forest is used to retain fault-relevant features. Recognizing that velocity variations are a major source of domain shifts, order analysis is introduced as a physics-prior alignment method. Deep CORAL and pseudo-label self-training are combined for unsupervised knowledge transfer. Finally, multi-level interpretability analysis is embedded to quantify domain shift, track adaptation geometry and explain final decisions. Two case studies demonstrate its robustness, trustworthiness and engineering applicability.
Yi Xie, Ruyang Zheng, Xuepeng Guo et al.· Eksploatacja I Niezawodnosc-...· 0 citations
While conventional fault diagnosis excels in steady-state scenarios, reliable detection under variable operating conditions remains a critical challenge in modern industrial drive systems. Although domain adaptation techniques address cross-domain distribution shifts, their deployment is often hindered by data privacy concerns and the impracticality of labeling massive target data in real-world factories. Furthermore, standard methods typically require concurrent access to source and target data, which violates the data isolation protocols common in industrial scenarios. To overcome these barriers, this paper proposes a novel and robust Source-Free Domain Adaptation (SFDA) framework equipped with post-hoc explainability for bearing fault diagnosis. First, to combat complex environmental interference and uneven sample distributions, we introduce a Gaussian noise-based stochastic perturbation strategy, forcing the model to learn robust decision boundaries. Second, high-confidence pseudo-labels are mined as reliable supervision signals. To fundamentally optimize the feature manifold, we introduce a contrastive learning-based representation strategy. Specifically, a mixed loss combining Supervised Contrastive loss and dot product loss is proposed to explicitly enhance feature discriminability while preventing feature collapse, achieving a delicate balance between intra-class compactness and overall diversity. Finally, a post-hoc visualization mechanism is incorporated to explain the model's decision-making process, thereby enhancing the transparency and credibility of the diagnosis for practical industrial deployment. Extensive experiments on two publicly available bearing fault datasets demonstrate that the proposed method significantly outperforms existing SFDA-based baselines. On the PU dataset, our approach achieves an average diagnostic accuracy of 99.11%, yielding a performance gain of 1.86% over the state-of-the-art SDALR method. Similarly, on the JNU dataset, the proposed framework attains an average accuracy of 98.36%, demonstrating superior robustness and trustworthiness in real-world fault diagnosis.
chenghao yan, Dongsheng Liu, Tong Wu et al.· Measurement science and tech...· 0 citations
Domain generalization (DG)-based fault diagnosis can effectively suppress domain drift interference induced by varying operating conditions of rolling bearings, enabling the model to maintain high accuracy and stable diagnostic performance under unseen working scenarios. However, most existing DG methods passively extract domain-invariant features via domain alignment, feature decoupling or data augmentation. Due to the absence of explicit constraints, their performance is restricted and unstable. Furthermore, they fail to take into account the suppression of spurious features triggering “shortcut learning,” which further degrades the generalization capability of the model. To address these limitations, this paper proposes a novel joint hierarchical suppression framework. Firstly, a cross-dimensional suppression algorithm is developed, which collaboratively operates on channel and spatial dimensions under domain label supervision to identify and remove domain-specific features in shallow network layers. Secondly, a global feature suppression mechanism is embedded into fully connected layers to drive the model to learn residual domain-invariant features, thus greatly boosting generalization performance. Finally, the above modules are integrated into an end-to-end trainable architecture to systematically extract robust domain-invariant representations. The proposed approach is verified on three rolling bearing datasets from the University of Ottawa, Huazhong University of Science and Technology, and Lanzhou University of Technology under unseen working conditions, achieving average diagnostic accuracies of 98.73, 96.32, and 99.22%, respectively. Experimental results verify the effectiveness and superiority of the presented method in identifying and eliminating domain-specific features.
Tianlong Huo, Rongzhen Zhao, Jun Gong et al.· Structural Health Monitoring· 0 citations
Variations in mechanical load constitute a dominant source of domain shift in data-driven fault diagnosis of rotating machinery, causing models trained under one operating condition to degrade sharply when deployed under another. This work addresses the problem through unsupervised domain adaptation (UDA)—which transfers diagnostic knowledge from a labeled source condition to an unlabeled target condition by aligning their feature distributions—and introduces Padé Approximant Neural Networks (PadéNets) as compact yet highly expressive feature extractors. One-dimensional PadéNet encoders are embedded into three established adaptation frameworks—Deep CORAL, Domain-Adversarial Neural Networks (DANNs), and Conditional Domain-Adversarial Networks (CDANs)—to learn load-invariant representations without any labeled target data. On the Case Western Reserve University benchmark, where the models operate directly on raw time-domain vibration signals, replacing conventional convolutional encoders with PadéNets consistently improves cross-load diagnostic accuracy, reaching up to 99.28% average target-domain accuracy at a low parameter count. To assess generalization to a more demanding setting, the CDAN–PadéNet configuration is further evaluated on frequency-domain representations of the Paderborn University dataset, where domain shift arises from simultaneous variation of load torque and radial force on bearings with real accelerated-lifetime damage, attaining 99.84% average accuracy across six cross-condition transfer tasks while requiring fewer parameters than competing methods. These results establish PadéNet-enhanced UDA as an accurate, broadly applicable approach for robust bearing fault diagnosis under varying operating conditions, with a reduced parameter count suited to resource-constrained embedded platforms.
Sertac Kilickaya, Cansu Celebioglu, Murat Askar et al.· Machines· 0 citations
In practical industrial applications, bearing fault diagnosis techniques frequently suffer from poor accuracy and limited generalization due to variable operating conditions, severe background noise, and weak fault vibration signatures. To address these critical challenges, a novel intelligent fault diagnosis framework, termed PGDS-CLNet, is proposed in this study. Different from traditional methods relying on handcrafted diagnostic features, the proposed PGDS-CLNet is an end-to-end trainable diagnostic network after standard signal normalization and segmentation. It directly learns discriminative representations from one-dimensional vibration segments without manually designing time-domain, frequency-domain, or time-frequency features. Elegantly extracting robust feature representations through a Prior-Guided Feature Extraction Module (PGFEM) and a Dual-Skip Separable Convolution Unit (DSCU), ultimately achieving precise multi-class fault identification via a Softmax layer. To effectively mitigate overfitting and minimize generalization errors, a comprehensive optimization strategy incorporating a learning-rate adaptive Adam optimizer, dropout, mini-batch training, L2 regularization, and Batch Normalization (BN) is seamlessly integrated into the model training process. Extensive experimental results demonstrate that the proposed PGDS-CLNet exhibits exceptional generalization capabilities under varying load conditions and intense noise environments. Furthermore, experimental results show that PGDS-CLNet achieves testing accuracies of 100.00%, 98.34%, and 98.66% under 0, 300, and 600 N load conditions, respectively, with a mean testing accuracy of 99.00%. Under strong Gaussian noise of −8 dB, the proposed model still achieves 87.24% accuracy, outperforming LSTM-CNN, and WDCNN.
Fanlong Zhu, Junyu Lai, Peiwen Lu et al.· Advances in Mechanical Engin...· 0 citations
As critical components in rotating machinery, the operational status of rolling bearings directly affects the safety and reliability of equipment. However, in real-world industrial environments, factors such as strong background noise, multi-source coupled vibrations, and complex transmission paths often mask the early failure characteristics of bearings, making it difficult for traditional methods to achieve accurate diagnosis. While existing signal-processing-based diagnostic methods possess clear physical significance, they exhibit poor adaptability under non-stationary, high-noise conditions; conversely, purely data-driven deep learning methods, despite their powerful feature extraction capabilities, face issues such as poor interpretability, heavy reliance on labeled data, and limited generalization ability. To address this, this paper proposes a fault diagnosis network that integrates physical feature enhancement and attention mechanisms-PFA-Net. This network consists of a cascaded Feature Decoupling Enhancement Module (FDEM) and a Feature Aggregation Module (FAM). The FDEM comprises a Quadratic Convolution Feature Enhancement Submodule (QCFEM) and a Physical Feature Decoupling Submodule (PFDM). the former enhances higher-order nonlinear features in time-frequency images through learnable quadratic convolution operations, while the latter introduces physical operators such as the Fourier transform and the Hilbert transform to decouple fault pulses and envelope information from the enhanced features into interpretable physical metrics, thereby effectively highlighting fault modes in noisy environments. FAM, on the other hand, integrates an improved residual block with a Swin Transformer. It enhances local key features through a convolutional attention module (CBAM) and captures global dependencies using a shift-window self-attention mechanism, achieving efficient synergy between local and global features. Through optimization via a dynamic joint loss function, PFA-Net significantly improves fault diagnosis accuracy in high-noise environments while ensuring physical interpretability. Experimental results demonstrate that this method exhibits excellent robustness and generalization capabilities under complex operating conditions, providing a structurally clear and high-performance solution for intelligent fault diagnosis.
Chengcheng Wang, Yunge Li, Rui Li et al.· Scientific Reports· 0 citations