Domain generalization in fault diagnosis (FD) faces significant challenges, primarily due to its inability to fully leverage multi-scale feature information. This study proposes the multi-scale feature alignment-based domain generalization network for bearing FD, which overcomes the limited diagnostic performance of previous networks that focus solely on single-scale feature information. First, the generalized s-transform is employed to convert one-dimensional vibration signals into two-dimensional time-frequency representations. Next, a multi-scale feature extractor is designed by integrating the coordinate attention mechanism, the gate recurrent unit module, and the transformer encoder module, enhancing the model’s capacity to extract features at multiple scales. Furthermore, the training process is enhanced by combining the maximum mean discrepancy and correlation alignment (Coral) loss functions. Finally, adversarial training strategies are employed to boost the diagnostic accuracy of the discriminator, thereby strengthening the model’s adversarial diagnostic capabilities. The proposed method was experimentally validated using a self-constructed bearing fault dataset and a publicly available bearing fault dataset from Jiangnan University (JNU). The results indicate that, on the self-constructed dataset, the proposed method achieves an accuracy of 98.54% in the multi-source domain diagnosis task, marking a 6.74% improvement over the inadequate source domain task. Additionally, for the insufficient source domain diagnosis task on the JNU dataset, an average accuracy of 92.52% was attained. These results thoroughly demonstrate the reliability and effectiveness of the proposed method.

In rolling bearing intelligent fault diagnosis (FD), lightweight models are constrained by issues such as noise interference and the scarcity of fault data, making it challenging to achieve real-time, high-accuracy diagnosis on resource-limited devices. To address these challenges, this study proposes a lightweight model that combines the hierarchical fine-grained decision fusion (HFDF) strategy with an improved EfficientNetV2 architecture (HFDF-EffNetV2). The model optimizes depth and width multiplicity factors to enhance parameter utilization efficiency. It uses pyramidal convolution (PyConv) combined with Fused-MBConv (Fused-MBPyConv) to obtain multi-scale time-frequency information. Additionally, an enhanced MBConv, termed BSMB-Conv-MLCA, integrates subspace blueprint separable convolution (BSConv-S) with mixed local channel attention (MLCA) extract deep-grained fault features. The HFDF strategy outputs confidence in stages and updates weights to prevent the model from falling into local overfitting when handling confusable samples. Experimental results on Case 1 and Case 2 show that HFDF-EffNetV2 achieved 100% accuracy with diagnostic times of 18.67 millisecond (ms) and 17.56 ms, respectively, and 1.85 million (M) parameters. Under noisy conditions, average accuracies reached 98.19% and 85.68%, respectively. Additionally, the model performed well with small samples, yielding accuracies of 98.69% and 97.51%. These results highlight its superior robustness to noise and lightweight performance compared with other advanced models.

Remaining useful life (RUL) prediction of rolling bearings is of significance for improving the reliability and durability of rotating machinery. Aiming at the problem of suboptimal RUL prediction precision under cross-working conditions due to distribution discrepancies between training and testing data, enhanced cross-working condition RUL prediction for rolling bearings via an initial degradation detection-enabled joint transfer metric network is proposed. Specifically, the health indicator, called reconstruction along projection pathway (RAPP), is calculated for initial degradation detection (IDD), in which RAPP is obtained from a novel deep adversarial convolution autoencoder network (DACAEN) and compares discrepancies between the input and the reconstruction by DACAEN, not only in the input space, but also in the hidden spaces, and then RUL prediction is triggered after IDD via RAPP. After that, a joint transfer metric network is proposed for cross-working condition RUL prediction. Joint domain adaptation loss, which combines representation subspace distance and variance discrepancy representation, is designed to act on the final layer of the mapping regression network to decrease data distribution discrepancies and ultimately obtain cross-domain invariant features. The experimental results from the PHM2012 dataset show that the proposed method has higher prediction accuracy and better generalization ability than typical and advanced transfer RUL prediction methods under cross-working conditions, with improvements of 0.047, 0.053, and 0.058 in the MSE, RMSE, and Score.

To address the issue that fault features of rolling bearings are difficult to extract effectively under strong background noise interference, leading to challenges in fault diagnosis, an adaptive diagnostic method optimized by an Improved Weighted Averaging Algorithm (IWAA) is proposed. The method incorporates three core techniques: smooth logistic chaotic map initialization, dynamic adaptive position updating, and thinking innovation strategy. These significantly enhance the global optimization capability of the algorithm, with a 59.4% improvement in convergence speed and a 72.4% increase in stability compared to WAA. A joint denoising framework is constructed by integrating Sparse Maximum Harmonic-to-Noise Ratio Deconvolution (SMHD) and Successive Variational Mode Decomposition (SVMD); in this framework, the filter length of SMHD and the penalty factor of SVMD are simultaneously optimized using IWAA, eliminating reliance on manual parameter tuning. An innovative adaptive fluctuation index is introduced, incorporating a dynamic weighting mechanism combining kurtosis and energy entropy to enable intelligent selection of modal components. Verified using public datasets and a custom experimental platform, the method successfully extracts fundamental to 9th and 11th harmonic features of outer race bearing faults under strong noise interference; it significantly enhances impulse periodicity and outperforms mainstream methods in high-order feature recognition, providing a robust diagnostic solution for complex industrial scenarios.