This study develops an event-driven peridynamic framework for modelling rolling contact fatigue (RCF)–wear competition at the wheel–rail interface. An energy-based bond fatigue model and the USFD wear law are treated as modular sub-models and coupled through a unified cycle-advancement strategy. At each iteration, the minimum of the fatigue-related, wear-related, and remaining-cycle increments updates fatigue states, wear depths, and surface geometry on a common cycle scale without simulating every rolling cycle. A particle-size sensitivity analysis shows that a minimum spacing of 0.10 mm provides a practical compromise among result stability, crack-path resolution, and computational cost. The framework is applied to a two-dimensional wheel–rail model to examine the effects of wheel load, creepage, and friction coefficient. Under the selected parameters and mild-wear conditions, increasing wheel load or friction coefficient promotes crack propagation and wear accumulation, with the response tending towards RCF dominance. Creepage produces a non-monotonic crack response: increasing creepage initially promotes crack propagation, whereas further increases after tangential-force saturation mainly enhance sliding and material removal, strengthening the relative influence of wear. The no-wear comparison shows that wear-induced material removal reduces the retained crack length and modifies the crack path. The fatigue-rate sensitivity analysis shows that the quantitative crack response and the occurrence of a complete fatigue-to-wear transition remain parameter-dependent. The framework provides a unified method for investigating coupled fatigue crack evolution and wear-induced surface removal under repeated wheel–rail contact.

Achieving efficient low-frequency sound absorption within limited thickness remains a persistent challenge in acoustic engineering. In this work, an ultra-thin broadband absorber is developed by exploiting the weak coupling interaction among Helmholtz resonant units. Instead of relying on bulky cavity enlargement, the proposed design establishes a gradient distribution of resonant frequencies through deliberate modulation of geometric parameters, enabling broadband performance via spatially compact arrangements. Numerical simulations reveal that the two configurations exhibit high absorption efficiency (sound absorption coefficient > 0.8) over 480–1023 and 300–500 Hz, respectively. Despite their compact geometries, the overall thicknesses are restricted to 32 and 42 mm, corresponding to only 1/10.6 and 1/27 of the minimum operating wavelengths. To verify the predictive model, impedance tube measurements were carried out independently for both structures. The experimental results closely follow the simulated trends, confirming the reliability of the analytical and numerical framework. The proposed designs, therefore, demonstrate effective subwavelength absorption capability across targeted mid-low and low-frequency ranges.

With the continuous increase in train operating speeds, rolling contact fatigue on rail surfaces has become a major concern, especially in welded regions where geometric irregularities and material mismatches intensify stress concentrations and accelerate crack growth. This study integrates ABAQUS and FRANC3D through a fracture-mechanics-based submodeling approach to simulate surface crack propagation in welded rails with realistic 3D weld geometry. The developed model clarifies how crack geometry and weld irregularities jointly affect fatigue behavior under rolling contact loads. The results indicate that surface cracks mainly propagate through a mixed Mode II-III shear mechanism. A 45 degrees inclination angle represents the most critical propagation orientation. Increasing the surface crack length from 2 mm to 6 mm shortens fatigue life by 57.7%, while extending the crack length to 8 mm elevates the peak Mode III stress-intensity factor to 811 MPa, promoting inward growth. In the weld zone, the deeper crack front propagates faster, causing semicircular cracks to evolve into elongated ellipses. A strong interaction is observed between weld wavelength, depth and crack phase position: shorter wavelengths and deeper undulations markedly increase the Mode II stress-intensity factor, with the fastest propagation occurring at 3/8 of the weld wavelength-where fatigue life drops to 44% of that at 7/8 of the wavelength. The findings clarify the shear-dominated crack evolution mechanism and provide theoretical guidance for weld-grinding thresholds and fatigue-life assessment of welded rails in high-speed railway applications.

Minor defects are easy to appear at the rail web bolt holes, which seriously affect the safety of trains when passing through, and accurately identifying the bolt hole defects is crucial. Leveraging the symmetry of the rails, this study proposes an innovative guided wave signal processing workflow. This enables the orientation determination and relative size identification of millimetre-scale defects in rail bolt holes. In this paper, a comb excitation method based on the group velocity principle is employed to obtain a single guided wave mode for the detection of bolt holes in the rail web. Then, millimetre-scale defects were fabricated by a hybrid division of hexahedral and tetrahedral elements. The model utilises a circular dot matrix on both sides of the rail web as the signal receiving points. The symmetry of rails is used to superimpose the signals and apply a band-pass filtering process. This allows for the successful localisation of defects in eight positions of bolt holes using cumulative amplitude intensity and normalised dynamic time warping. The model also investigates the relative size recognition of bolt hole defects in terms of length and depth. Furthermore, the feasibility of detecting bolt hole defects was experimentally validated.

High-ductility, compositionally clean austenitic steels deform almost homogeneously at the macroscopic scale, making crack nucleation and crack-path selection under near-ultimate tension highly sensitive to microstructural heterogeneity. Here we combine in-situ scanning electron microscopy/electron backscatter diffraction (SEM/EBSD) tensile tests on a high-Mn austenitic steel with a strongly coupled crystal-plasticity-phase-field framework to clarify how crystallographic orientation, texture and grain-boundary misorientation govern plasticity-controlled fracture. Experiments on pre-notched specimens reveal that damage initiates within low-Taylor-factor/high-Schmid-factor "soft corridors" decorated by dense {111}< 110 > slip bands. Slip-band intersections and grain boundaries with a low slip-transfer coefficient act as sites of kernel average misorientation accumulation, multi-site micro-crack nucleation and subsequent crack deflection. The proposed crystal plasticity coupled phase-field model uses slip-system-resolved plastic work, together with a tension-compression split of the elastic energy, as the crack-driving field and degrades only the tensile contribution. With parameters calibrated by the in-situ tests, the model accurately reproduces the macroscopic stress-strain response, the evolution of plastic-damage bands and the experimentally observed crack trajectory. Systematic simulations show that grain orientation effects cannot be reduced to a single Schmid factor: orientations with higher Schmid factor may generate shorter cracks by promoting crack-tip blunting through more diffuse plastic zones, whereas low-Schmid-factor orientations confine plasticity to narrow soft corridors and favour long cracks. Texture engineering tunes soft-corridor connectivity and thus the strength-ductility trade-off and crack tortuosity, while bicrystal analyses demonstrate that intermediate grain-boundary misorientations maximise cracking strain and minimise crack length by balancing slip transfer and plastic incompatibility. These findings provide a microstructure-informed basis for designing fracture-resistant high-ductility austenitic steels.

Reliable detection and quantitative evaluation of rail surface spalling are challenged by strong reflections, worn textures, and irregular defect morphology. This paper presents an integrated framework covering image-based observation and three-dimensional damage assessment. A non-contact region extraction method identifies the geometrically preserved rail side using only grayscale statistics, providing a stable anchor for downstream processing. This region guides point cloud alignment and allows multi-segment scans to be rectified into a unified coordinate system with improved geometric consistency. On the rectified geometry, the proposed spalling segmentation network achieves an F-measure of 88.62% and an IoU of 85.01% on field-acquired data. The segmented regions are then mapped to the aligned point cloud to reconstruct their three-dimensional morphology and compute depth-, width-, and area-related indicators. The results demonstrate reliable defect geometry and stable pose alignment, offering an effective solution for automated rail condition assessment.

Weak magnetic sensing on mobile or mechanically disturbed platforms is often limited by vibration-induced common-mode interference that masks the target magnetic response. To address this problem, we propose an eigenmode-decoupled multilayer differential magnetoelectric architecture in which magnetic-field-induced and disturbance-dominant responses are separated into differential and common-mode channels, respectively, through anti-parallel poling and differential processing. A mode-matching factor k is further introduced as a directly measurable indicator to quantify the relative dominance of common-mode and differential-mode components. The proposed approach is validated under magnetic excitation, base-vibration loading, line-frequency interference, mixed magnetic-vibration conditions, and UAV-borne operation. Experiments show that low-frequency vibration components are suppressed by more than −15 dB and 50 Hz common-mode interference is attenuated by up to −47.13 dB, while differential processing reduces the equivalent magnetic noise density to 21.45 pT/Hz1/2 at 1 Hz. At 20.19 kHz, the differential-mode magnetoelectric coefficient reaches 8689 V·cm−1·Oe−1, with an equivalent magnetic noise density of 15.77 fT/Hz1/2. These results suggest that the proposed structural design, combined with physics-informed differential processing, provides a compact and robust solution for weak magnetic signal extraction in vibration-contaminated environments.

Significant interactions exist among multiple surface cracks in railheads. This study investigates the competitive propagation behavior of rail surface cracks using the compact tension (CT) tests and a peridynamic (PD) model. Four sets of CT tests for multi-crack propagation were designed with U75V railhead material, and corresponding PD fatigue models were established. Significant shielding effects were observed among cracks during propagation, with the PD model accurately replicating crack propagation paths and fatigue lives from CT tests. A PD model was constructed to simulate the dynamic crack propagation on rail surfaces under rolling wheel loading, revealing significant promotion and suppression effects among cracks dominantly influenced by crack number and spacing. PD-predicted crack branching and coalescence align with field rail damage patterns.

The vehicle-turnout system is a complex wheel-rail interaction system affected by multiple sources of uncertainty. These uncertainties include stochastic excitation from track irregularities, random variables, and spatially non-uniform stiffness degradation of fastener. These uncertainties introduce high-dimensional random inputs and nonlinear dynamic responses, making it difficult to capture system response characteristics with limited samples, reduce sampling-induced variability, and extract probabilistic information from response data. Existing reliability methods often depend on large-scale simulations, which limits their efficiency for degraded turnouts under limited sample conditions. To address this issue, this paper proposes a dynamic reliability evaluation method considering spatial stiffness degradation of fastener. A deterministic representation based on the good lattice point method with adaptive weighted dimension-by-dimension (AW-DD-GLP) is incorporated into the random-phase trigonometric series method to generate track irregularity samples consistent with the target spectrum. A vehicle-turnout coupled dynamic model is then established, where fastener stiffness is treated as a spatially non-uniform degradation parameter. The probability density evolution method (PDEM), combined with the path integral solution (PIS) strategy, is used to obtain probability density functions of extreme dynamic responses at different degradation levels. A standardized multi-indicator system reliability model is further developed by integrating safety and ride-comfort indicators. Results show that the proposed method accurately describes probability distributions of random dynamic responses with limited samples and agrees well with Monte Carlo simulations. As fastener stiffness degradation intensifies, dynamic responses deteriorate and system reliability decreases. The wheel unloading rate and Sperling index are identified as the dominant failure-controlling indicators. This method provides an efficient probabilistic tool for degradation assessment and maintenance decision-making of turnout fastener systems.

Accurate and stable contact modeling in complex three-dimensional (3D) peridynamic (PD) simulations across diverse contact scenarios remains challenging. This study proposes a GPU-accelerated 3D element-based PD contact modeling framework (3D-EBCM). The framework reconstructs finite local micro contact surfaces in the current configuration and determines normal and tangential contact stiffnesses from the local incremental PD responses of contacting particles. A critical time-step estimate accounting for contact stiffness is then established. The 3D-EBCM contact routine is implemented on GPUs using bond- and particle-mapping strategies combined with background-grid-assisted candidate search. The results of the two-cylinder benchmarks agree well with the Hertz solution over the examined material-stiffness ratios and particle resolutions. For the examined stiffness ratios, peak contact-pressure errors are approximately 1%, while penetration ratios remain near 1% across the benchmark cases. The GPU-accelerated contact routine achieves overall speedups of 99.1–263.0 relative to the corresponding serial CPU implementation. Impact-contact, frictional-contact, and rolling-contact examples further assess the method. The impact-contact results obtained using time steps below the estimated critical value reproduce the experimentally observed fracture pattern and support the contact-related critical time-step estimate for selecting a stable fixed time step. The frictional-contact results agree with finite-element results, whereas the rolling-contact results agree with the corresponding Hertz and Carter solutions. The method also captures damage-induced contact-force redistribution and the effects of creepage, surface cracks, and surface irregularity, and resolves two simultaneous, spatially separated rolling-contact regions.

Traditional railway track fastener maintenance relies primarily on manual labour, which is not only inefficient but also poses significant safety risks. Precise fastener positioning is fundamental to achieving automated fastener maintenance. To ensure automated fastener maintenance within the complex environment of ballasted railway tracks, this paper proposes a two-stage positioning optimization method for ballasted railway track fasteners (TPOBRTF). In the first stage, the edge information of the track image is extracted. The Hough transform and mask processing are adopted to enhance the edge of the sleeper. Then, the position information of the left and right boundaries, as well as the approximate position information of the top and bottom boundaries of the sleeper fasteners, is obtained through gray integral projection. In the second stage, based on the approximate position information of the top and bottom boundaries of the fasteners obtained in the first stage, and considering the constraint of the center position of the quasi-circular shape, the top and bottom boundaries of the fasteners are optimized to achieve precise positioning. Experimental results demonstrate that, compared to template matching methods, TPOBRTF achieves a 21.5% higher mean intersection over union (mIOU) for fastener positioning. The mean root mean square error (mRmse) of the fastener boundaries localized by TPOBRTF was reduced by 80.4% compared to the template matching method. Furthermore, the deviation accuracy between TPOBRTF and the reference fastener boundaries offset was superior to that of template matching. The proposed TPOBRTF method maintains a mIOU exceeding 0.8 when processing noise-contaminated images, demonstrating precise fastener position capability within the complex environment of ballasted railway tracks. The research results may provide theoretical and technical foundations for intelligent maintenance operations of railway tracks.

Train passage through turnouts induces significant vibrations in urban rail transit systems, particularly under diverging-route operation with small-radius curves. Such curves generate pronounced centrifugal forces, amplifying structural responses. To investigate vibration sources and transmission characteristics of turnouts, full-scale field tests were conducted for both through and diverging routes. Vertical and lateral accelerations were measured at rails, track slabs, and tunnel walls. Time-frequency analysis was applied, and new indicators were developed to quantify vibration transmission efficiency and energy distribution in the variable cross-section turnout. Vehicle-turnout rigid-flexible coupled dynamic simulations were used to interpret the observed vibration behavior. Results show that diverging-route vibrations are dominated by low-frequency components (1-30 Hz) induced by centrifugal forces, with higher amplitudes than in through-route operations. Vertical vibrations exhibit higher transmission efficiency than lateral vibrations. These findings highlight the mechanisms of low-frequency vibration generation and provide guidance for vibration control in urban rail turnout zones.

In ballasted track maintenance, key processes like tamping and stabilization lack real-time, quantitative feedback on ballast bed conditions. This forces reliance on operator experience and mechanical parameters, causing under-tamping, over-tamping, and ineffective replacement. To address this, the paper proposes a non-contact vision-based monitoring method for tracking the six-degrees-of-freedom (six-DoF) of ballast particles. A redundant ArUco marker cluster establishes a dynamic mapping between marker poses and particle centroids, reconstructing particle motion. For challenges like variable lighting and occlusion, a multi-objective optimization adjusts marker colors and imaging parameters to improve detection robustness. Experiments achieve a marker recognition rate of 97.2% under difficult conditions. Loading tests confirm stable and continuous output of ballast particle motion trajectories. The results validate the redundant marker system and adaptive visual strategy in enhancing environmental adaptability and data reliability. This paper offers a new approach for real-time ballast bed monitoring and supports intelligent development of ballasted track maintenance.

Switch panels in railway turnouts are critical for route transitions, but their complex nonlinear behavior complicates vibration analysis. This study integrates hammer testing with an explicit LS-DYNA finite element model to compute frequency response functions and systematically assess sensitivity to modeling parameters (mesh, excitation points, hourglass control, damping). Results show that including nonlinear baseplate contact is essential for capturing the dynamic response in the 300 – 1500 Hz range; without it, simulated resonance peaks shift and attenuate. A fundamental frequency bifurcation around 50 Hz is observed, triggered by coupling between locking devices and railhead contact beyond baseplate #12.5. This interaction intensifies nonlinear effects above 500 Hz. Additionally, iron blocks suppress vibrations below 500 Hz, while fillers limit vibration transmission to one sleeper span and reduce high-frequency peaks above 1500 Hz. Operating deflection shape analysis reveals dominant resonant wavelengths and baseplate deformation modes, with significant deviations from linear behavior above 576Hz due to contact nonlinearities. This explicit finite element approach effectively simulates impact dynamics, providing valuable insights for fatigue prediction and turnout design optimization.

This study introduces a dual-phase compression-torsion metamaterial (DP-CTM) concept that integrates load-bearing capability with vibration attenuation for metro track structures. Unlike conventional vibration control approaches that rely on stiffness reduction or local resonance, the proposed DP-CTM enables functional integration through compression-torsion (CT) coupling enabled by the internal skeleton. Under axial loading, the internal compressive-torsional skeleton converts vertical compression into rotational deformation via constrained topology, while the thermoplastic polyurethane (TPU) layer functions as a constrained viscoelastic damping layer, enabling efficient hysteretic energy dissipation through shear-dominated deformation. To achieve reliable design under limited data, a data-efficient sequential surrogate-based multi-objective optimization framework is developed for high-precision parameter tuning. Numerical simulations and experimental tests are performed to validate load-bearing and vibration attenuation performance. Results reveal symmetry-dependent deformation and vibration mitigation characteristics. A CT ratio is used to quantify deformation conversion efficiency, showing stronger torsional activation in the antisymmetric configuration, whereas the symmetric configuration provides improved load-bearing performance with similar to 2.7% lower compressive deformation at 20 kN. Frequency-sweep tests further confirm effective vibration attenuation in the 103-166 Hz range, with average vibration level reductions of 3.1 dB and 3.5 dB for the antisymmetric and symmetric structures, respectively. These findings support a metamaterial-driven alternative to conventional rubber-based isolation solutions for metro track vibration control.

The High-Speed Railway Turnout (HSRT) is a complex, non-periodic assembly structure with asymmetric longitudinal distribution along the track, subjected to broadband wheel-rail impact loads. These features lead to highly intricate mechanical behavior. Gaining insight into the vibration characteristics of the HSRT is essential for understanding and optimizing its structural performance and ensuring the service safety of railway infrastructure. This study, for the first time, conducted a systematic in-situ vibration test to reveal the spatialfrequency distribution patterns of vibration characteristics across different rails of the HSRT and to investigate the influence of rail fastening methods and various connection components on its dynamic behavior. The findings indicate that the vibration characteristics of each HSRT rail vary continuously along the track due to special structures, such as variable rail cross-sections and connection components, which generally amplify lowfrequency responses below 120 Hz. Above 790 Hz, the RPDFRF amplitudes exhibit a "striped" distribution due to the half-wavelengths of "pinned-pinned" and similar modes approximately matching the sleeper spacing, though this pattern is less pronounced in variable cross-section regions. In these regions, the increased railhead width raises the system mass, lowering the natural frequency and causing an overall shift of the RPDFRF amplitudes toward lower frequencies. Different fastening methods mainly affect RPDFRF amplitudes and peaks below 240 Hz, while fastening constraints significantly reduce amplitudes in the 867-1820 Hz range. Various connection components predominantly influence the RPDFRF amplitude and peak distribution within 300 Hz for each rail.

The shear resistance of ballast under dynamic loading directly affects the stability and bearing capacity of the ballast bed; however, the influence of load characteristics on direct shear parameters remains insufficiently understood. A direct shear testing scheme for premium-grade ballast was presented in this study, utilising a dynamic direct shear apparatus configured to accommodate varying characteristics of dynamic loading. Twelve loading conditions were tested, including constant normal load shear and dynamic normal load shear at different frequencies and amplitudes. The effects of these parameters on shear strength, vertical deformation, particle breakage, vertical dynamic stiffness, and damping were analyzed. Results indicate that, under identical static loads, dynamic normal loading exerts little influence on the initial shear force but either strengthens or weakens shear strength. High-frequency loading weakens shear strength, whereas high-amplitude loading strengthens it. Dynamic normal loading suppresses the shear dilation effect, leading the ballast towards a stable compressive state during shearing and causing greater particle breakage. Both compressive deformation and particle breakage increase nonlinearly with the frequency and amplitude of dynamic loading. In addition, ballast dissipates energy through intrinsic damping, achieving a relatively stable dynamic stiffness in the later stages of shearing. These findings provide theoretical support for the scientific maintenance and operation of railway ballast beds.

Ballasted tracks at the ends of long-span suspension bridges are subjected to complex cyclic longitudinal displacements and train dynamic loads, which may lead to severe ballast degradation and threaten track stability. To investigate the underlying instability mechanism and evaluate potential mitigation measures, this study develops a multiscale numerical framework by coupling a full-bridge finite element model with a local bridgeend ballast bed model based on the discrete element method and finite difference method. Using this framework, the ballast-bed responses under thermal loading and train loading are compared in terms of particle migration, geometric evolution, and mechanical performance degradation. Two distinct instability mechanisms emerge: thermal loading induces vertical dilatancy dominated by interfacial friction, while train loading triggers intense longitudinal ballast flow due to inertial effects. Quantitative analysis shows that train loading leads to degradation rates of 26.25% (longitudinal resistance) and 24.42% (sleeper support stiffness), both higher than the thermal loading values of 19.26% and 19.84%, respectively. To address this performance deterioration, two mitigation measures, sleeper densification and polyurethane stabilization, are evaluated. Results demonstrate that the latter forms an elastic bonding matrix that limits resistance degradation to 16.11%, markedly outperforming the former. The proposed multiscale approach provides an effective tool for analyzing continuum-granular interaction in bridge-track systems under complex boundary conditions and offers theoretical support for the design and maintenance of stable ballasted tracks at the ends of long-span bridges.

PurposeIn vehicle-track coupled dynamics, the wheel-rail contact model is particularly important because its accuracy and stability determine the effectiveness and reliability of the vehicle-track coupled model. The aim of this article is to propose a model that can satisfy the efficiency and accuracy of the dynamics.Design/methodology/approachThis article proposes an equivalent ellipse model (EEM) that considers non-Hertz geometry gaps. Using CONTACT as a reference, the accuracy and stability of the model are validated in both static contact and dynamic applications.FindingsThe results show that in static contact, EEM exhibits higher accuracy in both the normal and tangential directions, compared to the Hertz theory. In dynamic applications, two specified cases are considered: straight track with lateral irregularity and curved track. The Hertz theory exhibits instability in the curved track compared to the EEM. The critical speed calculated by EEM is also closer to CONTACT than that of Hertz theory, showing great accuracy. Finally, as a rapid algorithm, EEM exhibits computational efficiency comparable to the Hertz theory, being several times faster than CONTACT.Originality/valueThe novelty of this method is that it uses the ratio of length and width obtained from KP method to re-solve the Hertz theory. At the same time, this is different from the Hertz theory, which uses the curvature of the contact point as the input and avoids the error caused by excessive curvature changes near the contact point.