Under variable working conditions, a problem arises, which is that it is difficult to obtain enough labeled data; to address this problem, an adaptive transfer autoencoder (ATAE) is established to diagnose faults in rotating machinery. First, a data adaptation module, which calculates the maximum mean discrepancy for the network hidden-layer data in reproducing kernel Hilbert space, is introduced to the autoencoder network, thus making the classification model operate under variable working conditions. Variation particle-swarm optimization is then invoked to optimize the data adaptation parameters. Finally, the k-nearest neighbors algorithm, as the classification layer of the network, identifies the state of health of the rotating machinery. The capabilities of the intelligent fault-diagnosis network are verified using vibration signals from a bearing test rig and a gearbox test rig. The experimental results suggest that, compared with state-of-the-art diagnosis methods, the proposed ATAE network can significantly boost diagnostic performance in the absence of target vibration signal labels.

Axle-box bearings are one of the most critical mechanical components of the high-speed train. Vibration signals collected from axle-box bearings are usually nonlinear and nonstationary, caused by the complicated operating conditions. Due to the high reliability and real-time requirement of axle-box bearing fault diagnosis for high-speed trains, the accuracy and efficiency of the bearing fault diagnosis method based on deep learning needs to be enhanced. To identify the axle-box bearing fault accurately and quickly, a novel approach is proposed in this paper using a simplified shallow information fusion-convolutional neural network (SSIF-CNN). Firstly, the time domain and frequency domain features were extracted from the training samples and testing samples before been inputted into the SSIF-CNN model. Secondly, the feature maps obtained from each hidden layer were transformed into a corresponding feature sequence by the global convolution operation. Finally, those feature sequences obtained from different layers were concatenated into one-dimensional as the fully connected layer to achieve the fault identification task. The experimental results showed that the SSIF-CNN effectively compressed the training time and improved the fault diagnosis accuracy compared with a general CNN.

An unforeseen fault of the key bearing of production system due to different reasons has the potential to cause an interruption in the entire production line, resulting in economic and production losses. To improve the reliability of industry production, this paper presents an intelligent diagnosis method for element rolling bearing based on the integrated interaction relationship of vibration nonlinear features. The nonlinear features of vibration signals are extracted using recurrence quantification analysis (RQA) and regrouped into different subsets of non-redundant features with the same level of discrimination ability through the technology of ReliefF-affinity propagation clustering. The weighted voting variable predictive model class discrimination (WV-VPMCD) is proposed to fully utilize the interaction of RQA features to do intelligent diagnostics for bearing faults. The experimental results showed that the WV-VPMCD outperformed the conventional intelligent diagnosis methods in terms of accuracy, consistency, stability and robustness, especially in the case of small number of samples.

Numerous unsupervised domain adaptation methods for bearing fault diagnosis rely on extracting domain-invariant features to mitigate the impact of domain shift interference. However, the lack of evaluation criteria results in limited interpretability of domain-invariant features. Additionally, current pseudo-labels prediction methods heavily rely on label information or computational resources, and the traditional Softmax function fails to capture valuable information. To address these problems, this paper proposes an unsupervised domain adaptation method based on domain-invariant features evaluation and knowledge distillation for bearing fault diagnosis. Firstly, mutual information and soft attention mechanism are integrated into the extraction of multivariate features to access the quality of domain-invariant features and enhance interpretability. Then, the concept of knowledge distillation is introduced to predict pseudo-labels in the target domain without relying on label information or computational resources. Furthermore, an asynchronous feature metric adaptive strategy is developed to adjust the feature alignment metric, considering the maturity and precision of pseudo-labels. The effectiveness and superiority of the proposed method are demonstrated through comparative experiments and ablation studies on two bearing datasets.

Spectral kurtosis (SK) has been proved to be a powerful tool to help extract impulse-like fault characteristics from original non-stationary vibration signals buried in strong masking noise. However, the SK coefficient based on the fourth sample moments suffers from a non-robustness problem, which may affect the accuracy of the SK-based kurtogram. Therefore, robust spectral kurtosis coefficients are defined based on quantiles in order to eliminate the influence of the outliers in original signals. Robust kurtograms are firstly utilized to analyze the fault signals. Subsequently, intrinsic time-scale decomposition followed by envelope demodulation is introduced to decompose the signal filtered by the kurtogram for fault detection. Compared with the conventional kurtogram and empirical mode decomposition, the results demonstrate that the improved method is able to facilitate the ability enhancement of the fault diagnosis of rolling element bearings.