The scarcity of fault samples degrades the accuracy of data-driven intelligent fault diagnosis (IFD). An auxiliary classifier generative adversarial network (GAN) has therefore emerged as a dominant paradigm for generating multiclass data to mitigate this issue; however, this framework suffers from low intraclass diversity and gradient instability. Specifically, the classifier's strong class-label separability compresses the category's support space, reducing diversity and also causing optimization conflicts. To this end, this article proposes a discriminative condition-guided generative model (DCGM) to synthesize high-fidelity data across categories for induction motor fault diagnosis under small-sample conditions. First, a discriminative classifier is integrated into the conditional GAN to replace the auxiliary classifier, theoretically alleviating diversity constraints and instability. After that, an adaptive feature matching loss based on supervised contrastive learning is proposed to enhance the synthetic quality of the class-label data. Then, three novel evaluation metrics are developed to quantitatively assess the generated data quality. Extensive experiments on induction motor datasets demonstrate that DCGM achieves state-of-the-art evaluation scores compared to diffusion-, transformer-, and GANs-based models. Finally, small-sample fault diagnosis further validates the superiority of the proposed approach, highlighting its potential in engineering applications. To the best of our knowledge, this is the first to introduce the novel discriminative generative framework for conditional data generation in the IFD field.

Low-friction surface conditions significantly contribute to the reduction of the wheel/rail adhesion capability and the occurrence of wheel/rail slipping behaviors, which may lead to the degradation of mechanical properties and frictional wear damage at the wheel/rail interface. To mitigate these undesirable consequences, modern railway locomotives are equipped with on-board anti-slip control systems. In this study, three different anti-slip controller models, comprising the traditional re-adhesion anti-slip controller and PID-based anti-slip controller with fixed threshold and with optimal threshold, are established. The wheel/rail rolling-slipping performances subjected to different anti-slip control algorithms under changing wheel/rail friction conditions are compared based on train-track interaction simulations. The results demonstrate that the PID-based anti-slip controller with an optimal threshold achieves the maximum utilization of wheel/rail adhesion in the presence of low-friction conditions, outperforming the other two types of anti-slip controllers. Additionally, the adoption of an anti-slip controller with a lower control threshold can effectively reduce the tread wear of locomotive wheels. This research can provide a deep going understanding of optimization design of anti-slip controller on railway vehicles.

The integration of lift airfoils (LA) into high-speed trains (HSTs) offers significant advantages, such as improved transport efficiency, reduced wheel-rail forces, and lower energy consumption. However, these benefits are accompanied by challenges to ride comfort and operational safety due to the varying aerodynamic loads experienced at high speeds. Existing passive suspension systems are insufficient in addressing these issues effectively. Therefore, it is crucial to develop an active suspension system that can mitigate the impact of fluctuating aerodynamic forces on LA-HSTs. An adaptive robust displacement inequality constraint-following control (AICFC) active suspension system is proposed, leveraging constraint-following control theory to enhance HST stability under aerodynamic load disturbances. Additionally, a linear eddy current brake (ECB) is integrated with the AICFC active suspension to further improve operational safety in complex environments. Simulation results demonstrate that the proposed AICFC-based active suspension system significantly reduces car-body vibrations and displacements, thereby enhancing ride comfort. Moreover, the combination of the active suspension and ECB enhances operational safety beyond what the AICFC active suspension alone can achieve. Finally, this study incorporates the delay effects of both the actuator and ECB to evaluate the performance of the active suspension system. The findings indicate that these delays can degrade the enhancement of dynamic performance in HSTs-LA achieved by the active suspension. However, the active suspension retains a significant and unparalleled advantage over passive suspension systems.

Higher running speed and more complex operating circumstances are the development trends for future high-speed trains. Due to these two key features, robust anti-slip control system and powerful traction/braking capability are necessary. This paper presents a new methodology to achieve maximum wheel-rail adhesion and reinforce traction/braking capacity based on wheel-rail contact condition recognition, sliding mode control, and the assistance of electromagnetic actuators. The proposed anti-slip strategy is implemented in a three-dimensional high-speed train-track coupled dynamics model. Effects of the control strategy on the traction/braking performance and the wheel-rail dynamic interactions of high-speed trains under complex contact conditions are simulated and discussed. The influence of electromagnetic actuators on wheel surface wear and rail deformation is also investigated. It is shown that the combination of electromagnetic actuator and sliding mode algorithm derived anti-slip control have a significant synergy effect on wheel-rail interactions and adhesion performance. Compared with the traditional anti-slip control methods, the proposed control strategy can effectively improve the traction/braking performance of high-speed trains with acceptable side effects on wheel surface wear and rail deformation.

A robust constraint-following control (CFC) scheme is proposed for active suspension system to improve ride comfort of high-speed trains (HSTs) by employing nonlinear stiffness of bio-inspired structure (BIS). The control problem of driving active suspension system to emulate nonlinear dynamics of BIS is formulated as servo constraint-following. The disturbance observer is theoretically integrated into CFC framework to address critical practical issues, including parameter uncertainty, suspension nonlinearity and aerodynamic disturbance. The complex uncertainty decomposition and uncertainty bound function design can be avoided in the control design compared with past adaptive CFC. Lyapunov stability theory is applied to prove the uniform ultimate boundedness of proposed robust CFC, enabling the constraint-following error to converge to a very small range. The simulation results demonstrate that the active suspension system could exactly follow the servo constraints of reference BIS with high constraint-following accuracy regardless of parameter uncertainty and external disturbance. And the active suspension system can effectively suppress the vibration and improve the ride comfort steadily with random structure parameters and strong aerodynamic loads. This study also provides a new insight and an alternative approach for designing robust CFC in the active suspension of HSTs.

Based on the information loss analysis of the blur accumulation model, a novel single-image deblurring method is proposed. We apply the recurrent neural network architecture to capture the attention perception map and the generative adversarial network (GAN) architecture to yield the deblurring image. Considering that the attention mechanism has to make hard decisions about specific parts of the input image to be focused on since blurry regions are not given, we propose a new adaptive attention disentanglement model based on the variation blind source separation, which provides the global geometric restraint to reduce the large solution space, so that the generator can realistically restore details on blurry regions, and the discriminator can accurately assess the content consistency of the restored regions. Since we combine blind source separation, attention geometric restraint with GANs, we name the proposed method BAGdeblur. Extensive evaluations on quantitative and qualitative experiments show that the proposed method achieves the state-of-the-art performance on both synthetic datasets and real-world blurry images.