Tao Guo

With the continuous increase in high-speed train operation speeds, lightweight bogie design has become a key means to enhance dynamic performance, which also increases the risk of structural fatigue. High-frequency wheel–rail excitations are transmitted to the bogie frame and couple with its higher-order modes at around 200 Hz, inducing local high-frequency resonance. This coupling markedly increases the stress amplitude within the affected frequency range and accelerates vibration-induced fatigue damage. This study investigates the vibration fatigue characteristics of a bogie frame with an inner axle box under high-speed operation and wheel polygon wear conditions. Using a high-speed wheel–rail interaction test rig, dynamic stresses and the vibration acceleration of the bogie frame are measured under different speeds and polygon orders. Based on modal analysis and vibration fatigue methods, a high-frequency vibration fatigue assessment method for the bogie is developed. Wheel polygon significantly amplifies mid-to-high-frequency vibration energy, and for the bogie frame with an inner axle box, pronounced modal coupling is observed at around 200 Hz. In particular, under the 11th-order polygon condition, the equivalent stress at critical locations such as the traction motor seat weld seam exceeds the fatigue limit, while the effect of the 20th-order polygon is relatively mitigated. The proposed vibration fatigue assessment method provides a theoretical basis for the safe design and operational maintenance of high-speed trains with bogie frames with inner axle boxes.

The evolution characteristics of multi-source-fatigue-crack propagation in the subsurface of a high-speed wheel’s tread and its influence on tread peeling life are the basis for accurately evaluating wheel service lifetime. This study explores the influence of morphology distribution and the size of cracks in the tread on peeling life. The results show that the crack propagation mode in the wheel is mainly mode II and mode III composite propagation caused by shear stress. A fatigue crack inside the wheel with an angle of 45° represents the most dangerous situation. The maximum value of the von Mises stress inside the wheel increases with the increase in the number of multi-source cracks. The equivalent stress intensity factor (SIF) for multi-source cracks is higher than for a single crack. Also, mode III propagation has higher sensitivity to the number of cracks. The existence of multi-source cracks also increases the initial driving force ΔK_eq of crack propagation. The results are useful for the evaluation of the service life of high-speed wheels.

In the field of modern mechanical engineering, structures often endure multi‐level variable stress loading. The nonlinear fatigue cumulative damage process of these structures is highly complex due to the significant influence of loading sequences and interactions, which makes fatigue life prediction difficult. To accurately describe the impacts of these factors on fatigue damage, this paper proposes a nonlinear fatigue cumulative damage model (EWOA‐MH) based on the enhanced whale optimization algorithm (EWOA) and the Manson–Halford (M‐H) model. This model obtains weight factors through EWOA and incorporates them into the M‐H model. Verified by experimental data of multi‐level variable stress loading and calculated with a weighted method considering different materials' sample numbers, the prediction accuracy is increased by approximately 43%. Its application to the analysis of high‐speed train bogie frames effectively demonstrates the model's effectiveness. The research shows that the EWOA‐MH model performs outstandingly in fatigue life prediction and can effectively solve fatigue damage problems under multi‐level variable stress loading conditions.

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.