Bearing is the key part of mechanical equipment, which can support the rotating machinery running. It is crucial to diagnose bearing fault in time to ensure mechanical equipment working well. Effective feature extraction is an essential step in bearing fault diagnosis. However, the bearing vibration signals collected from the equipment usually contain interference, such as heavy noise. It is difficult to extract effective feature from bearing vibration signals due to the interference. To overcome this issue, a new feature extraction method for rolling bearing faults diagnosis is proposed based on hierarchical improved envelope spectrum entropy (HIESE). First, hierarchical decomposition is used to divide the bearing vibration signal into several hierarchical components. Second, the original feature set is obtained by calculating the improved envelope spectrum entropy (IESE) of each hierarchical component. Then, joint approximate diagonalization of eigenmatrices (JADE) is introduced to fuse the original features into a new one. Finally, support vector machines (SVM) is taken to identify the bearing status. Two cases are used to test the proposed method. In case 1, the bearing vibration signals composed of different fault degrees and different fault types under two speeds are used to verify the proposed method. The recognition rate achieves 100%. In case 2, the bearing vibration signals, which consist of different fault types under three operation conditions, are analyzed. The recognition rate also achieves 100% in this case. Experimental results illustrated that the proposed method has good performance in feature extraction, which can provide a new method for feature extraction of rolling bearing.

Data-driven based fault diagnosis methods play an increasingly important role in rotating machinery. Among these methods, deep learning is widely concerned due to its strong nonlinear feature learning ability, wherein ResNet is a very powerful model. However, the diagnostic performance of the model depends on sufficient labeled data samples, which is extremely difficult to achieve under actual complex working conditions. In this paper, a multidimensional normalized ResNet is proposed for fault diagnosis of cross-working conditions under limited label samples. Firstly, the collected vibration data under different conditions are preprocessed by computed order tracking to reduce the distribution differences. Secondly, batch normalization and group normalization are fused together to enhance the feature extraction ability of ResNet. Moreover, the rectified linear unit is replaced by the gaussian error linear unit to improve the robustness of the trained model. Then, the model is trained by source domain samples, and the obtained parameters will be transferred to the target domain. Finally, the transferred model is fine-tuned through limited samples in the target domain and applied for cross conditions fault diagnosis. Two kinds of datasets are analyzed by the proposed method and existing models to demonstrate its superiority.

A bearing is a key component in rotating machinery. The prompt monitoring of a bearings’ condition is critical for the reduction of mechanical accidents. With the rapid development of artificial intelligence technology in recent years, machine learning-based intelligent fault diagnosis (IFD) methods have achieved remarkable success in the field of bearing condition monitoring. However, most algorithms are developed based on computer platforms that focus on analyzing offline, rather than real-time, signals. In this paper, an edge intelligence diagnosis method called S-AlexNet, which is based on a parameter transplantation convolutional neural network (CNN), is proposed. The method deploys the lightweight IFD method in a low-cost embedded system to monitor the bearing status in real time. Firstly, a lightweight IFD algorithm model is designed for embedded systems. The model is trained on a PC to obtain optimal parameters, such as the model’s weights and bias. Finally, the optimal parameters are transplanted into the embedded system model to identify the bearing status on the edge side. Two datasets were used to validate the performance of the proposed method. The validation using the CWRU dataset shows that the proposed method achieves an average prediction accuracy of 94.4% on the test set. The validation using self-built data shows that the proposed method can identify bearing operating status in embedded systems with an average prediction accuracy of 99.81%. The results indicate that the proposed method has the advantages of high recognition accuracy, low model complexity, low cost, and high portability, which allow for the simple and effective implementation of the edge IFD of bearings in embedded systems.

Train bearing is one of the key parts of a train wheel, and its fault diagnosis is crucial for train safety. This paper proposes an Improved Linear Discriminant Analysis (ILDA) algorithm to diagnose train bearing faults. A new distance function is employed to adjust the between-class divergence matrix of LDA. The adjusted between-class divergence matrix is adopted for selecting the projection direction with the best clustering performance and reducing the ambiguity or overlap among different classes. The sample clustering evaluation index( SI ) is adopted to evaluate the clustering performance of fusion features obtained by ILDA. First, twenty-four original features are extracted from wayside acoustic signals after reducing the Doppler effect. Then, ILDA is adopted to extract fusion features. Finally, the train bearing faults are identified by extreme learning machine(ELM). The experimental results show that the fusion features extracted by the proposed method can be effectively identified, which have good clustering performance and high identification rate. Furthermore, the identification results of two datasets provided by different universities show that the fusion features obtained by ILDA also have good clustering performance and high identification rate, which further verify the validity of the proposed method.