The data-driven intelligent fault diagnosis method of rolling bearings has strict requirements regarding the number and balance of fault samples. However, in practical engineering application scenarios, mechanical equipment is usually in a normal state, and small and imbalanced (S & I) fault samples are common, which seriously reduces the accuracy and stability of the fault diagnosis model. To solve this problem, an auxiliary classifier generative adversarial network with spectral normalization (ACGAN-SN) is proposed in this paper. First, a generation module based on a deconvolution layer is built to generate false data from Gaussian noise. Second, to enhance the training stability of the model, the data label information is used to make label constraints on the generated fake data under the basic GAN framework. Spectral normalization constraints are imposed on the output of each layer of the neural network of the discriminator to realize the Lipschitz continuity condition so as to avoid vanishing or exploding gradients. Finally, based on the generated data and the original S & I dataset, seven kinds of bearing fault datasets are made, and the prediction results of the Bi-directional Long Short-Term Memory (BiLSTM) model is verified. The results show that the data generated by ACGAN-SN can significantly promote the performance of the fault diagnosis model under the S & I fault samples.

The transient thermal performance of rolling bearings affects the mechanical performance and system safety of traction motors. Most of the traditional empirical formulas used in temperature analysis have been simplified and cannot be completely applied to the calculation of heat generation and convection heat transfer coefficients. Based on the comparative analysis of finite element transient temperature and experimental data, this paper proposes a correction method of mathematical model and derives an accurate calculation formula for the heat generation and lubricant convection heat transfer coefficient of ball bearings applicable for the non-driving end in the traction motor of a high-speed EMU (Electric Multiple Unit). The accuracy of the results has been verified by durability experiment data. In addition, with changes in speed, radial load and other factors taken into account, we have analyzed the influence of these time-varying factors on ball bearing temperature, as well as the temperature distribution law of each component in a grease-lubricated bearing, in a bid to lay a foundation for follow-up research on the heat transfer laws of traction motors and rolling bearings.

The fault diagnosis of rolling element bearing is of great significance to avoid serious accidents and huge economic losses. However, the characteristics of the nonlinear, non-stationary vibration signals make the fault feature extraction of signal become a challenging work. This paper proposes an improved variational mode decomposition (IVMD) algorithm for the fault feature extraction of rolling bearing, which has the advantages of extracting the optimal fault feature from the decomposed mode and overcoming the noise interference. The Shuffled Frog Leap Algorithm (SFLA) is employed in the optimal adaptive selection of mode number K and bandwidth control parameter α. A multi-objective evaluation function, which is based on the envelope entropy, kurtosis and correlation coefficients, is constructed to select the optimal mode component. The efficiency coefficient method (ECM) is utilized to transform the multi-objective optimization problem into a single-objective optimization problem. The envelope spectrum technique is used to analyze the signals reconstructed by the optimal mode components. The proposed IVMD method is evaluated by simulation and practical bearing vibration signals under different conditions. The results show that the proposed method can improve the decomposition accuracy of the signal and the adaptability of the influence parameters and realize the effective extraction of the bearing vibration signal.