In urban rail transit, the rigid catenary systems are widely used for traction power supply, which relying on continuous sliding contact between the pantograph and the catenary. In recent years, many urban rail operators worldwide have reported abnormal pantograph strip wear, which has become a critical issue affecting both maintenance costs and operational safety. This issue is generally believed to be attributed to a combination of environmental conditions, mechanical interactions, and electrical effects. Based on six years of maintenance data from Tianjin Metro Line 6 in last six years, this study investigates the underlying wear mechanisms through an integrated data-driven and experimental approach, which reveals a distinct three-stage wear evolution process, comprising normal wear, mild abnormal wear, and severe abnormal wear. Microscopic observations indicate that, under conditions of low contact force and low relative humidity, the dominant wear mechanism transitions from stable mechanical friction to one governed by arc ablation and material transfer, triggered by the depletion of the carbon film on the contact wire surface. Furthermore, statistical correlation analysis identifies low relative humidity (<35%) and insufficient contact force as the primary contributing factors. These conditions facilitate the initiation of mild electrical arcing, which progressively intensifies through a positive feedback loop involving surface roughening and dynamic contact instability, ultimately resulting in severe arcing activity and a rapid acceleration of the wear rate. Ring-on-block experiments further confirm that reduced humidity and contact pressure markedly aggravate arc discharge behavior and material degradation.

In high-speed railway systems, the contact force is a crucial metric for describing the interaction performance between the pantograph and the catenary. However, measuring the contact force directly is challenging compared to other measurable quantities such as panhead acceleration and uplift. Leveraging advancements in artificial intelligence, this work proposes a deep learning framework to predict the contact force time history based on panhead acceleration and uplift data. Canonical recurrent neural network models, including vanilla RNN, LSTM and GRU are employed to model and predict the contact force. To enhance efficiency by reducing model parameters and extracting sequential feature maps, we propose a hybrid GRU model integrated with a one-dimensional convolutional neural network, called CNN-GRU. The database for this study is built using a validated and reliable finite element model of the pantograph-catenary system, incorporating variations in structural parameters, speed levels and cut-off frequencies. Four training strategies are designed to evaluate prediction performance across these variations. The results demonstrate the potential of using deep learning to predict contact force based on panhead acceleration/uplift, with some limitations in predicting across different cut-off frequencies. While vanilla RNN, LSTM and GRU models achieve acceptable prediction accuracy, the proposed CNN-GRU outperforms them, particularly in terms of prediction consistency across varying train speeds.

The dynamic characteristics of the pantograph-catenary system(PCS)directly affect the current collection quality and operational safety of high-speed trains.While traditional finite element methods(FEM)that utilize nonlinear cable-truss equivalent models accurately characterize the strong nonlinearity and time-varying mechanical behaviors of the PCS,they suffer from prohibitive computational complexity that hinders real-time prediction and digital twin deployment.To address these computational bottlenecks,data-driven surrogate models have emerged.However,standard Fourier neural operators(FNO)rely on fixed-frequency band truncation,which effectively captures low-frequency principal modes but systematically discards critical high-frequency transient details,such as l ocalized contact force mutations and wave reflections.Purely data-driven models also lack explicit physical constraints,leading to severe error accumulation during long-term dynamic simulations.To overcome these multiscale modeling challenges,this paper proposes adaptive Fourier neural operator diffusion model(AFNODM),a novel physics-informed framework that synergistically integrates an adaptive Fourier neural operator(AFNO)with a conditional diffusion model(CDM)to establish a time-frequency collaborative generation paradigm.In the first stage,the AFNO acts as a global physical skeleton generator to capture dominant vibration modes(0-20 Hz).Crucially,we introduce a velocity-based frequency modulation mechanism equipped with deformable convolution kernels,allowing the model to adaptively adjust its spectral receptive field in response to real-time train speeds and neutralize Doppler effects.In the second stage,a CDM-driven post-processing architecture is deployed.Conditioned on the AFNO's output,the diffusion model executes a progressive reverse denoising strategy in the latent space to seamlessly reconstruct the missing high-frequency residual details(above 20 Hz),while kinematic constraint losses are embedded via automatic differentiation to ensure absolute derivative consistency across spatial-temporal fields.Extensive evaluations on a high-fidelity PCS dataset(200-380 km·h-1)demonstrate that AFNODM efficiently solves complex dynamic equations with an unprecedented balance of speed and precision.At 350 km·h-1,the root mean square errors(RMSE)for the displacement,velocity,and acceleration fields are remarkably low at 0.0673,0.1603,and 0.8503,respectively,representing an error reduction of over 50%compared with mainstream baselines such as deep operator network(DeepONet)and physics-informed enhanced Fourier neural operator(PI-EFNO).Frequency-domain analysis confirmed that CDM integration significantly suppressed the high-frequency relative spectral error(RSE)from 16.80%(using pure AFNO)to 6.55%.Cross-line robustness tests across three distinct high-speed railway configurations(Beijing—Shanghai,Guangzhou—Shenzhen,and Beijing—Tianjin)validated the exceptional generalization capabilities of the model under varying structural parameter perturbations.Ultimately,the proposed AFNODM framework provides a highly accurate,resolution-independent,and real-time capable computational engine,paving the way for next-generation digital twins and intelligent predictive maintenance in modern electrified railways.

To support the safe operation and technological promotion of existing line speed-up projects, this paper presents an assessment method for pantograph–catenary contact performance under the 200 km/h speed conditions, using the Guangzhou–Shenzhen Lines I and II speed-up projects as representative case studies. Based on the ANCF method, a refined pantograph–catenary coupling dynamic model is established to accurately characterize the large deformation and geometric nonlinear behavior of the catenary system. Model validation is achieved using actual measurement data from the CR400AF train. Based on this model, systematic simulation analyses were conducted to evaluate the current collection performance of four mainstream train models—CR300AF, CR400BF, CRH380A, and CRH380B—under both single-unit and double-unit operation conditions. Results indicate that dynamic contact force metrics for pantograph–catenary interactions meet all limit requirements specified in the Technical Specifications for Dynamic Acceptance of High-Speed Railway Projects under all operating conditions. This demonstrates that the pantograph–catenary system on the analyzed Guangzhou–Shenzhen Line exhibits excellent dynamic stability and safety under the targeted speed-up scheme, providing simulation-based justification for implementing the speed enhancement project.

This article presents an experimental analysis of the dynamic characteristics of the overhead contact line (OCL) system at speeds exceeding 400 km/h. Field test data, collected at 350, 400, and 420 km/h on a real-world high-speed railway, are used to investigate key dynamic behaviors, including steady arm uplift and contact force. Among them, the steady arm uplift was captured via a vision-based measurement device, and the contact forces were reported through an instrumented pantograph. The pantograph was tested while operating in both the forward and reverse directions. The experimental results highlight significant performance differences at 400 km/h compared to 350 km/h: 1) the OCL uplift peak increases by 105.19% as the speed rises from 350 to 400 km/h; 2) the OCL vibration mode transitions from a single-mode dominant behavior at 350 km/h to a multimode vibration at speeds above 400 km/h; 3) the mean contact force increases by 45.7% as the speed rises from 350 to 400 km/h; 4) the aerodynamic effects of the pantograph exhibit a notable difference between the forward and reverse directions, leading to a 20% variation in the mean contact force at 400 km/h as compared to 350 km/h; and 5) the fluctuation in the contact force’s standard deviation increases, and a higher number of outliers are observed, indicating a decrease in system stability and a deviation from a Gaussian distribution as the speed surpasses 400 km/h.

As an essential subsystem of electrified railway operation and maintenance, intelligent detection of catenary support components still faces several critical challenges: (1) the number of abnormal (negative) samples for components is severely limited; (2) component anomalies are highly diverse and exhibit heterogeneous visual characteristics; and (3) existing models generally show unsatisfactory detection performance when confronted with previously unseen anomaly types. To address these issues, this paper proposes a novel few-shot anomaly detection model for catenary components, termed BCLIP-ADer, built upon a Bayesian prompt contrastive vision–language pretraining framework. Specifically, a Bayesian prompt flow module (PFM) is designed to regularize the text prompt space via the jointly learned image-specific feature distribution (ISFD) and image-agnostic feature distribution (IAFD), thereby mitigating the degradation in detection performance on unseen component anomalies. Monte Carlo sampling over these learned distributions is further employed to generate diverse text prompts, leading to more comprehensive coverage of the prompt space. In addition, a cross-modal feature refinement module (CFRM) is designed to more effectively align dynamic text embeddings with fine-grained image features, thus enhancing anomaly detection at the component level. Finally, extensive experiments conducted on a UAV-based catenary dataset (CSCUD) demonstrate the effectiveness and superiority of the proposed approach. Specifically, the proposed method achieves I-AUROC/I-AP/I-F1_max scores of 94.2/93.2/93.1 under few-shot conditions.

With the rapid development of high-speed railways, the dynamic performance of the pantograph-catenary system plays a crucial role in ensuring the safe and stable operation of trains. This study investigates the effect of the structural parameters of the pantograph-catenary system to achieve good dynamic interaction performance under high-speed conditions. A finite element model of the catenary system, incorporating nonlinear cable and truss elements, and a lumped mass model of the pantograph are developed. The penalty function method is employed to simulate the pantograph-catenary interaction. A total of 2187 dynamic simulations are performed, with seven variables-pantograph parameters, span length, contact wire tension, messenger wire tension, number of droppers, stitch wire length, and stitch wire tension. The comprehensive effect of these parameters is evaluated based on dynamic performance indicators, such as pantograph-catenary contact force, pantograph head lift, and support point lift. The results indicate that increasing the number of droppers, contact wire tension, and messenger wire tension enhances dynamic performance, while an increase in span length negatively affects performance. Stitch wire tension has little to no effect.

The overhead contact line (OCL), installed along railway tracks, serves as the only source of power for electric trains. Extreme weather-induced ice on OCL causes aerodynamic instability, leading to large-amplitude self-excited galloping, endangering rail safety. To mitigate this issue, this study focuses on the messenger wire (MW)—with larger diameter, greater rigidity, and no direct pantograph contact —critical for OCL aerodynamic stability. A novel anti-galloping spoiler, designed as a helical attachment to the MW, is proposed to regulate the surrounding flow field and suppress galloping at its source. Computational fluid dynamics (CFD) simulations are conducted to analyse ice accretion patterns and the aerodynamic behaviour of iced cross-sections with the spoiler installed. Subsequently, a nonlinear finite element (FE) model using the Absolute Nodal Coordinate Formulation (ANCF) beam element is developed to numerically evaluate the spoiler’s effectiveness in mitigating galloping. Various installation schemes are systematically assessed, and simulation results reveal that when the spoiler covers 57.6 % of the MW span, the vertical galloping amplitude is reduced by 95.2 %, demonstrating the proposed design’s strong suppression capability without disturbing the pantographs’ operation.

Automated anomaly detection of catenary support components is pivotal for ensuring the operational safety of electrified railways. However, deploying high-performance deep learning models on resource-constrained Unmanned Aerial Vehicles (UAVs) remains challenging due to prohibitive computational costs and the complexity of unstructured inspection environments. While Spiking Neural Networks (SNNs) offer a promising energy-efficient alternative, they typically suffer from feature degradation caused by binary quantization and sensitivity to high-frequency background noise. To address these limitations, a novel energy-efficient framework named Hybrid Spiking CLIP (HS-CLIP) is proposed. Specifically, a SAM-based Foreground Component Segmentation and Cleaning (SAM-FCSC) strategy is first designed to explicitly filter out environmental interference (e.g., ballast and vegetation) before encoding. To compensate for information entropy loss induced by spike quantization, Multi-level Visual Feature Extraction (MVFE) and Learnable Textual Prompts (LTP) are introduced, ensuring robust alignment between sparse visual signals and rich semantic descriptions. Furthermore, a Sparsity Constraint (SC) mechanism is devised to mitigate neuronal over-activation, effectively suppressing task-irrelevant noise. Extensive experiments on a self-constructed catenary defect dataset (CSCUAV) and the public MVTec AD benchmark demonstrates that HS-CLIP achieves state-of-the-art detection accuracy with superior generalization capabilities. Crucially, the proposed method significantly reduces energy consumption compared to traditional ANN-based approaches, providing a viable solution for real-time, green intelligent inspection on edge platforms.

Stable sliding contact between the pantograph and catenary is essential for reliable current collection in high-speed rail. As operating speeds increase beyond 350 km/h, the pantograph-catenary system exhibits unprecedented high-frequency dynamic behaviors caused by the excitation of flexible structural modes. Traditional lumped mass pantograph models are unable to represent the structural flexibility, while fully flexible finite element models, though accurate, are computationally constrained for large-scale simulations. To fill this gap, a Koopman Neural Operator-based surrogate modeling framework is proposed to integrate dynamic theory into a data-driven learning scheme, enabling efficient and accurate representation of the high-frequency pantograph dynamics. The surrogate model integrates a deep autoencoder for the observable and inverse-observable functions with a modified Transformer as the evolution operator. Training data are generated from an accurate fully flexible pantograph model, which has been validated via experimental tests, under broadband random excitations. The training process is supervised by a combined loss of multi-step prediction and physical consistency. Inference tests demonstrate that the surrogate model maintains finite element-level accuracy while reducing computational time by three orders of magnitude when predicting 1000-step dynamic responses. Comparative analyses confirm its superior balance between physical interpretability and forecasting performance, highlighting its potential as a powerful tool for real-time simulation, efficient optimization, and comprehensive evaluation of the high-frequency behaviors of a high-speed pantograph-catenary system.

In the published publication [...]

In electrified rail systems, the contact wire of catenary system is subjected to intense mechanical and long-term transient arcing, making it susceptible to performance degradation and even fracture failure. This seriously threatens both the safety and service life of high-speed rail power-supply systems. Unlike previous studies that primarily focused on the intrinsic physical behaviors of the arc, this work quantitatively clarified the dominant influence of contact wire material parameters in arc-energy transfer and ablation evolution. First, the coupled interactions among the electromagnetic, airflow, and thermal fields during offline arcing are systematically described, and a multi-physics coupled arc model with strong engineering applicability is developed. Then, an experimental setup is designed to reproduce high-speed sliding contact and transient arcing, and the model is confirmed to accurately capture the dynamic evolution of the transient arcing. Furthermore, by incorporating the energy-dissipation mechanisms associated with material evaporation, a quantitative formulation for calculating contact wire ablation damage is established. Finally, based on the developed numerical platform for PC transient arcing, the influences of representative material parameters on the arc-induced ablation behavior are comprehensively assessed. Results show that increasing the thermal conductivity from 100 W/(m·K) to 380 W/(m·K) reduces the ablation area from 0.28 mm² to nearly zero, demonstrating that high-conductivity materials effectively enhance heat dissipation and suppress local overheating. Moreover, increases in specific heat capacity and density mitigate transient heat accumulation, thereby significantly reducing contact-wire melting and vaporization.

The frequent occurrence of transient arcing in pantograph-catenary (PC) electrical section overlaps (ESO) poses a serious threat to the operational safety of electric locomotives. Extending previous works that oversimplified the mechanical configuration or considered only limited operating conditions, this study incorporates the specialized ESO structure and systematically analyzes the PC arc temperature distribution under varying conditions. First, by incorporating the specialized parabolic structure of the ESO, a mathematical model was established to describe the dynamic separation trajectory between the slide and the non-working branch of the contact wire. Subsequently, a multiphysics-coupled arc model for the ESO was developed and validated through comparative analysis based on magnetohydrodynamic theory. Based on this, the morphological evolution of the ESO arc during the first arcing cycle under alternating voltage difference was elucidated. Finally, the temperature field distribution of transient arcing under various conditions was numerically analyzed using dynamic mesh techniques. The results indicate that the average temperature of the ESO arc increases from 7840 K to 101,386 K with increasing voltage difference, while showing a significant negative correlation with both train speed and crosswind velocity. This suggests that higher train speeds or lower voltage differences promote arc extinction. Moreover, the average arc temperature exhibits an oscillatory decay over time and stabilizes at approximately 8000 K within the first arcing cycle. Notably, under low voltage conditions, the minimum arc temperature is delayed by 1.8 ms, indicating a pronounced lag effect.

The rational matching of vehicle, control, and track parameters is critical for enhancing the ride comfort and stability of a 600 km/h electromagnetic suspension (EMS) high-speed Maglev system and facilitating its engineering application. To address this issue, this study investigates the sensitivity of key parameters influencing the system dynamic responses and conducts multi-objective parameter matching optimization based on the quantified sensitivity results. First, the stability conditions of the minimum suspension unit are derived, and resonance avoidance analysis is performed to exclude unstable parameter combinations. Second, a data-driven surrogate model integrating deep neural network (DNN) and Gaussian process (GP) methods is trained to enable rapid large-scale computations. Then, global sensitivity analysis is conducted using the Sobol method. Finally, the quantified sensitivity results are used to guide a multi-objective optimization process, yielding an optimized parameter set. The computational results provide quantitative sensitivity coefficients for the effects of individual parameters on the system responses. Under the optimized parameters, compared with the initial parameters, the standard deviation of the suspension gap is reduced by 15.45%, and the ride comfort index is reduced by 10.93%. The findings of this study provide theoretical support and strategic guidance for parameter matching in 600 km/h high-speed Maglev systems.

In electric railway systems, the interaction performance of pantograph-catenary systems (PCS) is essential for ensuring a stable electrical supply. Establishing high-fidelity numerical models using the finite element method is generally desirable, yet it involves considerable computational complexity and time demands. In this paper, we propose a novel dynamic modelling method that integrates physical information and data-driven approaches to solve the pantograph-catenary interaction, called Physics-Informed Enhanced Fourier Neural Operator (PI-EFNO). Firstly, the enhanced Fourier Neural Operator (EFNO) based on global perceptron is developed to capture the nonlinear mapping between parameter space and output solution space by approximating solutions to dynamic equations in the frequency domain. Then, we integrate multiple physics-informed loss terms into the EFNO architecture to handle implicit constraints within coupled equations, which leverages physical principles to guide learning while reducing the need for labelled training data. Additionally, a dynamic weighting mechanism adaptively balances the contributions of various terms in the physics-based loss function. Experimental results validate the effectiveness and advantages of PI-EFNO in modelling pantograph-catenary dynamics. It demonstrates exceptional accuracy, computational efficiency, and generalizability across diverse pantograph models and catenary conditions, exhibiting strong capability in learning physically coupled equation systems.

The overhead conductor rail (OCR) is widely used in urban rail transit systems to supply electric power to trains through sliding contact with the pantograph. However, contact wire height irregularity (CWHI) caused by installation and maintenance errors can significantly disturb the dynamic performance of the pantograph-OCR (POCR). Based on measured CWHI from Shenzhen Metro Line 10, this study uses the power spectral density (PSD) function to characterize the stochastic features of CWHI and generates representative samples using the Monte Carlo method. Then a procedure is proposed to incorporate the CWHI into the OCR and compute the OCR's initial configuration. The model's reliability is ensured via the validation against experimental data. Finally, numerical simulations are conducted at different train speeds to analyze the stochastic CWHI's effect on the contact force. The results indicate that CWHI tends to increase the dispersion of the contact force. The average (AVE) standard deviation (STD) at all speeds exceeds that of the ideal condition, with an exceedance probability of 100%. At a speed of 200 km/h, the probability of the maximum (MAX) contact force exceeding the safety threshold is 29.2%, corresponding to a reliability level of only 0.708.

Slip ring–brush assemblies are widely used in satellite mechanisms to transmit power and signals across rotating interfaces. Under authentic space environments—vacuum, radiation-dominated thermal exchange, and long-duration operation—the coupled effects of mechanical contact dynamics, electrical conduction, intermittent separation, and arcing can accelerate wear and degrade reliability. This paper presents a surface-resolved multiphysics model for multi-track slip rings with staggered brushes. The ring surface is discretized on a circumferential–axial grid and endowed with correlated 3D roughness, enabling interference-based asperity contact. Brush normal dynamics (mass–spring–damper) convert runout and micro-vibration into normal-force ripple and separation events. Electrical conduction is modeled by a parallel admittance network combining pressure-dependent micro-contact conduction and an event-based arc channel activated by separation, opening velocity, and current density with stochastic ignition. A 2D thermal model with ADI integration accounts for Joule/friction heating, radiative cooling, and optional hub conduction. Wear evolves via an Archard-type mechanical term and an arc-energy-driven erosive term. A FAST–MACRO multiscale scheme (20 s FAST, 100 h MACRO with periodic recalibration) enables tractable long-horizon wear prediction while preserving arc statistics. Baseline simulations for a 28 V bus demonstrate rare but nonzero arc activity and predict spatially non-uniform wear at the micrometer scale after 100 h.

The process of pantograph lowering in electric locomotives generates intense transient arcing, which significantly influences the service life of the pantograph-catenary (PC) contact pair. This article investigates slide ablation during pantograph lowering by incorporating heat dissipation from material evaporation, a factor not addressed in prior studies. First, a high-precision arc voltage acquisition system was developed and implemented under real railway conditions. This system served as the basis for validating the proposed multiphysics-coupled transient arcing model. Subsequently, an ablation model was established to account for the effect of material evaporation on surface temperature evolution under transient arcing. Then, numerical simulations were conducted to reveal the temperature field distribution on the slide surface. Finally, the effects of pantograph-lowering voltages, pantograph-lowering speeds, and ambient wind speeds on the slide ablation were systematically analyzed. The results indicate that reducing the pantograph-lowering voltage and increasing the pantograph-lowering speed can effectively mitigate slide surface ablation. Furthermore, the transient arcing produces an arc-shaped ablation pit on the slide surface, with the peak temperature stabilizing at approximately 4000 K, close to the carbon sublimation point.

As a critical subsystem in the operation and maintenance of electrified railways, the intelligent structural health monitoring (SHM) of catenary support components faces several key engineering challenges: (1) randomness and diversity of structural damage modes make it impractical to construct comprehensive supervised datasets; (2) the complex geometric topology of catenary support assemblies leads to information loss of subtle features during deep network extraction; (3) current reconstruction models, lacking normative semantic constraints, tend to replicate input defect features, thus leading to ineffective structural repair of abnormal catenary regions. To address these issues, we propose a Global State-aware Clustering-guided Masked Autoencoder (GSC-MAE) for robust unsupervised anomaly detection of catenary support structures. First, a Global State Aggregation (GSA) backbone leverages State Space Modeling (SSM) to enhance global topological perception, mitigating micro-structural feature loss in complex catenary components. Second, clustering-driven adaptive masking targets anomalous catenary regions, enforcing restoration from normal structural contexts and blocking defect propagation. Additionally, a SCGFormer decoder utilizes strided memory flow to prevent information forgetting, enhancing reconstruction fidelity of intricate catenary topologies. Extensive UAV-based experiments yield 99.4% I-AUROC and 98.5% I-AP, surpassing state-of-the-art methods. This framework provides an efficient solution for automated catenary structural integrity assessment, promoting intelligent infrastructure maintenance.