Analyzing the evolution trend of rail corrugation using signal processing and deep learning is critical for railway safety, as current traditional methods struggle to capture the complex evolution of corrugation. This present study addresses the challenge of accurately capturing this trend, which relies significantly on expert judgment, by proposing an intelligent prediction method based on self-attention (SA), a bidirectional temporal convolutional network (TCN), and a bidirectional gated recurrent unit (GRU). First, multidomain feature extraction and adaptive feature screening were used to obtain the optimal feature set. These features were then combined with principal component analysis (PCA) and the Mahalanobis distance (MD) method to construct a comprehensive health indicator (CHI) that reflects the evolution of rail corrugation. A bidirectional fusion model architecture was employed to capture the temporal correlations between forward and backward information during corrugation evolution, with SA embedded in the model to enhance the focus on key information. The outcome was a rail corrugation trend prediction network that combined a bidirectional TCN, bidirectional GRU, and SA. Subsequently, a multi-strategy improved crested porcupine optimizer (CPO) algorithm was constructed to automatically obtain the optimal network hyperparameters. The proposed method was validated with on-site rail corrugation data, demonstrating superior predictive performance compared to other advanced methods. In summary, the proposed method can accurately predict the evolution trend of rail corrugation, offering a valuable tool for on-site railway maintenance.

In the railway industry, the rail is the basic load-bearing structure of railway tracks. The prediction of the remaining useful life (RUL) for rails is important to avoid unexpected system failures and reduce the cost of maintaining the system. However, the existing detection of rail flaws is difficult, the rail deterioration mechanisms are diverse, and the traditional data-driven methods have insufficient feature extraction. This causes low prediction accuracy. With objectives set in relation to the problems outlined above, a rail RUL prediction approach based on a convolutional bidirectional long- and short-term memory neural network with a residual self-attention (CNNBiLSTM-RSA) mechanism is designed. Firstly, the pre-processed vibration data are taken as the input for the convolutional bi-directional long- and short-term memory neural network (CNNBiLSTM) to extract the forward and backward dependencies and features of the rail data. Secondly, the RSA mechanism is introduced in order to obtain the contributions of the features at different moments during the degradation process of the rail. Finally, an end-to-end RUL prediction implementation based on the convolutional bi-directional long- and short-term memory neural network with the residual self-attention mechanism is established. The experiments were carried out using the full life-cycle data of rails collected at the railway site. The results show that the method achieves a higher accuracy in the RUL prediction of rails.

Due to the long-term service through wheel-rail rolling contact, the train wheelset tread will inevitably suffer from different types of defects, such as wear, cracks, and scratches. The effective detection of wheelset tread defects can provide critical support for the operation and maintenance of trains. In this paper, a new method based on a local inference constraint network is proposed to detect wheelset tread defects, and the main purpose is to address the issue of insufficient feature spaces caused by small samples. First, a generative adversarial network is applied to generate diverse samples with semantic consistency. An attention mechanism module is introduced into the feature extraction network to increase the importance of defect features. Then, the residual spine network for local input decisions is constructed to establish an association between sample features and defect types. Furthermore, the network’s activation function is improved to obtain higher learning speed and accuracy with fewer parameters. Finally, the validity and feasibility of the proposed method are verified using experimental data.

Prolonged rolling contact fatigue between wheels and rails results in the formation of surface cracks on the rail and accurately analyzing the crack expansion behavior is essential to ensuring the safe operation of the train. Drawing upon the principles of fracture mechanics and finite element theory, this study establishes a finite element model of wheel–rail rolling contact that incorporates the presence of cracks. The method utilizes an interaction integral to calculate the stress intensity factors at the leading edge of the crack; then, the Paris formula is used to solve the crack spreading rate. It systematically examines the effects of the initial crack angle, the coefficient of friction of wheels to rails, and crack size on the behavior of fatigue crack propagation. The results indicate that the cracks primarily extend in the depth direction of the rail, transforming the semi-circular surface cracks into elliptical cracks with the major axis oriented along the rail’s width. Crack propagation is primarily driven by model II and III composite crack propagation, with their expansion rates influenced by operating conditions. In contrast, mode-I expansion is less sensitive to these conditions. Under single-variable loading conditions, a smaller initial crack angle results in a faster crack growth rate. Increasing crack length accelerates crack growth, while a higher friction coefficient inhibits it.

To address the challenges in existing multi-sensor data fusion methods for rail surface defect diagnosis, particularly their limitations in fully exploiting potential synergistic information among multimodal data and effectively bridging the semantic gap between heterogeneous multi-source data, this paper proposes a diagnostic approach based on a Progressive Joint Representation Graph Attention Fusion Network (PJR-GAFN). The methodology comprises five principal phases: Firstly, shared and specific autoencoders are used to extract joint representations of multimodal features through shared and modality-specific representations. Secondly, a squeeze-and-excitation module is implemented to amplify defect-related features while suppressing non-essential characteristics. Thirdly, a progressive fusion module is introduced to comprehensively utilize cross-modal synergistic information during feature extraction. Fourthly, a source domain classifier and domain discriminator are employed to capture modality-invariant features across different modalities. Finally, the spatial attention aggregation properties of graph attention networks are leveraged to fuse multimodal features, thereby fully exploiting contextual semantic information. Experimental results on real-world rail surface defect datasets from domestic railway lines demonstrate that the proposed method achieves 95% diagnostic accuracy, confirming its effectiveness in rail surface defect detection.

Under non-independent and identically distributed (Non-IID) conditions, significant variations exist in local model updates across clients and training phases during the collaborative modeling process of differential privacy federated learning (DP-FL). Fixed clipping thresholds and noise scales struggle to accommodate these diverse update differences, leading to mismatches between local update intensity and noise perturbations. This imbalance results in data privacy leaks and suboptimal model accuracy. To address this, we propose a differential privacy federated learning method based on adaptive clipping thresholds. During each communication round, the server adaptively estimates the global clipping threshold for that round using a quantile strategy based on the statistical distribution of client update norms. Simultaneously, clients adaptively adjust their noise scales according to the clipping threshold magnitude, enabling dynamic matching of clipping intensity and noise perturbation across training phases and clients. The novelty of this work lies in a quantile-driven, round-wise global clipping adaptation that synchronizes sensitivity bounding and noise calibration across heterogeneous clients, enabling improved privacy–utility behavior under a fixed privacy accountant. Using experimental results on the rail damage datasets, our proposed method slightly reduces the attacker’s MIA ROC-AUC by 0.0033 and 0.0080 compared with Fed-DPA and DP-FedAvg, respectively, indicating stronger privacy protection, while improving average accuracy by 1.55% and 3.35% and achieving faster, more stable convergence. We further validate its effectiveness on CIFAR-10 under non-IID partitions.