As the main driving equipment of modern industrial production activities, if a motor fails, it causes serious consequences. Bearings are the component with the highest motor failure frequency. It is of practical engineering significance to establish a high-precision algorithm diagnostic model for motor bearings. At present, in data-driven motor bearing fault diagnosis methods, the method of manually adjusting hyperparameters is usually adopted in complex network structure models with many hyperparameters. To realize the automatic optimization selection of hyperparameters, in this paper, a motor bearing fault diagnosis algorithm based on a convolutional long short-term memory network of Bayesian optimization (BO-CLSTM) is proposed. The algorithm combines the Bayesian optimization algorithm (BO), a long short-term memory network (LSTM) and the convolutional layer of a convolutional neural network (CNN). It saves the considerable workload of manually adjusting the hyperparameters, has good noise resistance, and realizes the true end-to-end motor bearing fault diagnosis. The proposed method is trained based on the original vibration signal of the bearing, and the accuracy of the final model reaches 100%. In addition, compared with other advanced fault diagnosis methods based on deep learning, the performance of the proposed method is significantly improved.

In many industrial systems, symmetry is the key to ensuring efficiency and reliability. For example, in electric vehicles, the driving system often requires high symmetry. As widely used motors, permanent magnet synchronous motors (PMSMs) are often used in highly symmetrical structures as the driving devices. Consequently, maintaining the symmetry of the system relies on the normal and stable operation of the PMSM, and it is necessary to diagnose faults in the PMSM in a timely manner. In PMSM fault diagnosis methods, frequency domain features of the stator current are extensively used. However, these features change with the switching of motor operating conditions, leading to difficulty of diagnosis in multiple operating conditions. Therefore, a fault diagnosis method based on a convolutional neural network (CNN) phase tracker is proposed in this paper. Through phase tracking and angular domain resampling, the fundamental frequency of stator currents in different operating conditions are aligned, so as to fix the distribution of frequency domain features and solve the problem of features changing with operating conditions. Experimental results show that the proposed method can resample the stator current signals with a small error, detect faults in a relatively short time with high accuracy, and diagnose fault type and severity level under multiple operating conditions.

The grey wolf optimization (GWO) algorithm is widely utilized in many global optimization applications. In this paper, a dynamic opposite learning-assisted grey wolf optimizer (DOLGWO) was proposed to improve the search ability. Herein, a dynamic opposite learning (DOL) strategy is adopted, which has an asymmetric search space and can adjust with a random opposite point to enhance the exploitation and exploration capabilities. To validate the performance of DOLGWO algorithm, 23 benchmark functions from CEC2014 were adopted in the numerical experiments. A total of 10 popular algorithms, including GWO, TLBO, PIO, Jaya, CFPSO, CFWPSO, ETLBO, CTLBO, NTLBO and DOLJaya were used to make comparisons with DOLGWO algorithm. Results indicate that the new model has strong robustness and adaptability, and has the significant advantage of converging to the global optimum, which demonstrates that the DOL strategy greatly improves the performance of original GWO algorithm.

Abnormal detection of wind turbine converter (WT) is one of the key technologies to ensure long-term stable operation and safe power generation of WT. The number of normal samples in the SCADA data of WT converter operation is much larger than the number of abnormal samples. In order to solve the problem of low abnormal data and low recognition rate of WTs, we propose a sample enhancement method for WT abnormality detection based on an improved conditional Wasserstein generative adversarial network. Since the anomaly samples of WT converters are few and difficult to obtain, the CWGANGP oversampling method is constructed to increase the anomaly samples in the WT converter dataset. The method adds additional category labels to the inputs of the generative and discriminative models of the generative adversarial network, constrains the generative model to generate few types of anomalous samples, and enhances the generative model’s ability to generate few types of anomalous samples, enabling data generation in a prescribed direction. The smooth continuous Wasserstein distance is used instead of JS divergence as a distance metric to measure the probability distribution of real and generated data in the conditional generative response network and reduce pattern collapse. The gradient constraint is added to the CWGANGP model to enhance the convergence of the WGAN model, so that the generative model can synthesize minority class anomalous samples more effectively and accurately under the condition of unbalanced sample data categories. The quality of anomalous sample generation is also improved. Finally, the anomaly detection is made on the actual operating variator dataset for the unbalanced dataset and the dataset after reaching Nash equilibrium. The experimental results show that the method used in this paper has lower MAR and FAR in WT converter anomaly detection compared with other oversampling data balance optimization methods such as SMOTE, RandomOverSampler, GAN, etc. The method can be well implemented for anomaly detection of large wind turbines and can be better applied in WT intelligent systems.