This paper presents the application of a novel AI-based approach, Neural Physics, to produce high-fidelity simulations of train aerodynamics. Neural Physics is built upon convolutional neural networks (CNNs), where the weights are explicitly determined by classical numerical discretisation schemes rather than by training. By leveraging the power of AI technology, this recent approach results in code that can run easily on GPUs and AI processors, achieving high computational speed without sacrificing accuracy. The approach uses an implicit large eddy simulation method based on a non-linear Petrov-Galerkin method to model the unresolved turbulence. Furthermore, for higher-order finite elements, the convolutional finite element method (ConvFEM) is used, which greatly simplifies the implementation of higher-order elements within the NN4DPEs approach. We demonstrate the capability of Neural Physics by simulating a freight Locomotive Class 66 and a partially loaded freight train operating in an open field environment with and without cross wind. This is the first time that ConvFEM has been applied to high-speed fluid flow problems in complex geometries. The results are validated against existing numerical results and experimental measurements, and show good agreement in terms of pressure and velocity distributions around the train body.

A train body’s cross-sectional shape has a significant impact on aerodynamic drag and operational safety in high-speed trains (HSTs). This study extracts five design variables from a real-world HST body: height, width, side arc radius, arc radius at the connection between the side and the roof, and arc radius at the connection between the side and the train’s bottom. The cross-validated Kriging surrogate model and the genetic algorithm are used to perform two types of aerodynamic optimization, with the cross-sectional area as a constraint. Cross-sectional shapes are optimized in both windless and windy conditions. Numerical results indicate that in a windless environment, the aerodynamic drag coefficient of the whole train is reduced by 2.4%; in a windy condition, the aerodynamic drag coefficient of the entire vehicle is reduced by 2.4%, and the aerodynamic lateral force of the leading car is reduced by 37.8%. These suggest that a flat and wide shape helps to reduce not only overall aerodynamic drag in a windless environment but also aerodynamic load in a windy environment, which can be accomplished by reducing the area of the side wall and top region, lowering the train body’s height, increasing its width, and lowering the radius of the side and top arcs.

The suspension gap is a critical operational parameter for high-speed maglev trains and significantly impacts their aerodynamic performance. Based on an engineering prototype of the high-temperature superconducting (HTS) pinning maglev train, this study established a detailed three-dimensional model, and then the aerodynamic characteristics of the HTS maglev train at 600 km/h with suspension gaps of 10 mm, 20 mm, and 30 mm were simulated based on the improved delayed detached eddy simulation (IDDES) turbulence model and SST k-ω two-equation. The results demonstrated that the underbody design of the HTS maglev train leads to unique aerodynamic drag and aerothermal distribution phenomena. The head car experiences the smallest drag, while the tail car experiences the largest. The aerothermal temperature on the train’s bottom surface progressively increases from the head to the tail. Additionally, the U-shaped track significantly constrains the flow around the train body, forming strong vortex structures. As the suspension gap increases from 10 mm to 30 mm, the airflow velocity in the train-track gap rises, reducing the underbody pressure and decreasing the lift of the head car by 12.43%. The drag of the head car increases by 10.98%, primarily due to changes in pressure drag. Additionally, the temperature at the underbody of the tail car rises further due to significant airflow deceleration. These findings provide valuable insights for advancing the engineering design and application of the high-speed HTS maglev technology.

Aerodynamic drag is the dominant factor contributing to energy consumption as the operational speed of high-speed trains increases, necessitating effective aerodynamic optimization strategies. This study investigates the aerodynamic characteristics of the bogie region under two bogie fairing configurations: baseline bogie fairing (BBF) and full bogie fairing (FBF). Both stationary and rotating wheelset conditions are considered. Wind tunnel experiments were conducted on a full-scale bogie model equipped with a wheelset drive system to simulate wheelset rotation. Additionally, numerical simulations were employed to analyze flow structures. Results indicate that the FBF configuration promotes a more uniform front-to-rear pressure distribution in the bogie region. The rotation of the wheelset notably affects the airflow near the wheels and extends its influence throughout the entire bogie region. Specifically, wheelset rotation reduces drag by 6.38% in the BBF configuration but increases drag by 3.5% in the FBF configuration. Further analysis reveals that, in the FBF configuration, aerodynamic drag primarily originates from the wheelsets. The rotating wheelset increases the aerodynamic drag by 18.8% for the rear wheelset, which is attributed to the shift in the pressure curve on the wheelset in the rotating direction. Therefore, the impact of wheelset rotation on aerodynamic characteristics should not be overlooked.

The increase in aerodynamic drag brings high energy consumption, which is a critical issue in the development of high-speed trains. Inspired by the excellent hydrodynamic characteristics of fish movement in nature, a two-dimensional numerical simulation method based on spring-smoothing model and adaptive mesh technology was utilized to explore the effects of different fishtail structures and two flexible motion modes (Eel mode and Lunate-tail mode) on the wake of high-speed trains, and to assess their potential for aerodynamic drag reduction. Results indicate that the biomimetic fishtail successfully suppresses the alternating shedding of vortices in the wake, and induces the aerodynamic drag fluctuation period to align with the fishtail oscillation period. The fishtail length, oscillation mode, and frequency have a significant impact on the wake flow and aerodynamic drag of the train. Among these, a 1850 mm Eel fishtail with parameters of λ = 1 and T = 8 s achieves the optimal drag reduction effect, with drag reduction rates of 39.12% and 26.00% for the tail car and the entire train, respectively. These findings provide a theoretical basis for the design of new low-resistance railway trains, promoting the sustainable development of rail transit towards goals of high-speed and energy-efficient.

This paper investigates the influence of numerical methods and mesh resolution on the prediction accuracy of the aerodynamic behaviors of a 1/20 scaled generic high-speed train (HST) model. A thorough comparison is made between partially averaged Navier-Stokes (PANS), large eddy simulation (LES), and wind tunnel experiments, covering aerodynamic forces, surface pressure, velocity distribution, and Reynolds stress and turbulent kinetic energy in the wake region. The Reynolds number for both simulations and experiments is set to 4.75×10⁵. The results show that the PANS approach accurately predicts flow characteristics observed in experiments and fine LES calculations, even with a low-resolution grid. PANS exhibits a distinct advantage over LES when grid resolutions are insufficient for resolving near-wall flow structures around the HST, both in open-air conditions and crosswind environments. Additionally, grid refinement improves the predictive accuracy of the HST’s aerodynamic performance, particularly in the presence of small yaw angle.

The pantograph region constitutes one of the dominant aerodynamic sound sources in high-speed trains. In this study, a 1:3 scaled model of a representative pantograph structure was constructed, explicitly accounting for the geometric configuration of its rod components. To achieve noise mitigation, the pantograph design incorporated aerodynamically optimized cylindrical rods with bio-inspired seal-vibrissa-shaped profiles, perforated geometries, and elliptical cross-sections, etc. The flow dynamics and aeroacoustic characteristics within the pantograph region were systematically investigated through the wall-adapting local eddy-viscosity large-eddy simulation coupled with the Ffowcs Williams-Hawkings (FW-H) acoustic analogy method. Results showed that the structural optimization of the pantograph key components greatly attenuated the vortex shedding intensity in the rod assemblies, inhibiting the initiation and evolution of large-scale Kármán vortex streets, reducing the surface pressure fluctuations, and enhancing the overall aerodynamic performance. In the optimized model of pantograph, the noise level at first tonal peak around 850 Hz is greatly mitigated and the second harmonic peak at 1750 Hz identified in the original model is absent, with overall sound pressure levels reduced by 6.3 dB(A) and 6.6 dB(A) along the streamwise and vertical planes, respectively. These findings validate the efficiency of the noise reduction methods introduced for the optimized pantograph structure.

This paper aims to explore the influence of different noise barrier heights on the sound source generation mechanisms of higher-speed trains (400 km/h) using a combination of delayed detached eddy simulation (DDES) and Ffowcs Williams-Hawkings (FW-H) equations. Four cases are investigated and compared, i.e. 1) no barrier, 2) 2.3 m, 3) 3.3 m, and 4) 4.3 m single-side barriers on a bridge. Numerical results show that the presence of noise barriers causes an increase in sound source intensity ranging from 2.1 to 2.8 dB(A). However, the relationship between the barrier height and the increase in sound source intensity varies across different parts of the train. Compared with the head and frontmiddle cars, the boundary layer is thicker around the rear-middle and tail car areas. A thick boundary layer introduces the influence of the crash wall, causing asymmetry and increases in sound source intensity. This is due to the deceleration region formed between the crash wall and the rail surface, as well as the acceleration region formed by the contraction of the flow channel in the noise barrier, both of which influence the sound source’s characteristics. In addition, higher barriers exacerbate asymmetry and increases in sound source intensity.

Tunnel-induced noise amplification has become a major constraint for high-speed trains. This study employs a 1/10 scale three-coach high-speed train model, using the improved delayed detached eddy simulation (IDDES) method coupled with the perturbed convective wave model to investigate the unsteady flow evolution, aerodynamic noise source distribution, and near-field acoustic characteristics of high-speed trains under open-air and tunnel conditions. The results show that the blocking effect of the tunnel wall enhances flow compression, increases local velocity, and aggravates flow disturbances and pressure fluctuations near the pantograph and tail car. In the tunnel, the total sound source energy reaches 1.14 × 10¹² N²/s², 5.26 times higher than in open air, with significant increases in the tail car, bogies, and pantograph. Bogie noise concentrates in the 50 to 1000 Hz range, while pantograph noise dominates from 1500 to 2500 Hz. Tunnel conditions further enhance peak distributions in the low and medium frequency bands. Although pressure disturbances on the train surface are mainly dominated by hydrodynamic effects, the radiated acoustic energy of the sound pressure levels on the roof and side surfaces is amplified by 33.3 and 22.6 times, far exceeding hydrodynamic energy amplification factors of 8.6 and 6.3. The study reveals coupled flow and acoustic mechanisms in tunnels, supporting noise reduction design for high-speed trains.

This study introduces a novel flow-through cowcatcher with integrated inlet and outlet channels as an aerodynamic noise mitigation strategy for the nose car of a high-speed train. The wall-adapting local eddy-viscosity large-eddy simulation (WALE-LES) combined with the Ffowcs Williams-Hawkings (FW-H) acoustic analogy approach is employed to evaluate its impact on the aerodynamic and aeroacoustic characteristics of the leading bogie region. Compared with the conventional closed cowcatcher, results show that the flow-through structure suppresses the flow separation, promotes more stable vortex evolution within the bogie cavity, and reduces the spatial extent of high-amplitude wall pressure fluctuations up to 40%, mitigating effectively the generation of aerodynamic noise. Semi-anechoic wind tunnel experiments validate the simulation results and demonstrate that the sound pressure levels at the far-field observers decrease by 0.4–0.6 dB(A) with the flow-through cowcatcher applied underneath the nose car. The dominant sound source around the leading bogie region is shrunk with intensity reduced about 1.0 dB(A). These findings confirm the effectiveness of the flow-through cowcatcher in reducing the aerodynamic noise produced from the leading bogie region, providing both theoretical insight and engineering guidance for structural optimization and low-noise design of the nose car in a high-speed train.

The increasing aerodynamic noise caused by high-speed maglev trains (HSMTs) contributes substantially to environmental pollution and passenger discomfort. Numerical studies were performed to examine the effect of air blowing/sucking modes, positions and velocities on the flow field change and their potentials in mitigating the aerodynamic noise produced by HSMTs. The results indicate that the aerodynamic noise can be effectively mitigated by implementing air-blowing in the transition region between the streamlined tail nose and constant cross-sectional body (Scheme 1) and the wake vortex shedding area near the tail nose (Scheme 3) at speeds below 0.3U (train speed), as well as in the side edge area (Scheme 2) at various speeds (0.1U–0.5U), primarily due to the suppression in wake vortices. The optimal noise reduction value of 1.53 dB(A) is achieved when blowing in Scheme 1 at a speed of 0.1U, while the efficacy of the air-sucking mode is inferior with a smaller noise reduction value less than 0.84 dB(A). Additionally, simultaneous reductions in aerodynamic noise and drag can be achieved when sucking in Scheme 2 at speeds below 0.2U and blowing in Scheme 3 at speeds below 0.3U. These findings offer valuable insights for the application of active flow control technology in the design of low-resistance and low-noise HSMTs.

The pantograph cavity coupling system (PCCS) of high-speed trains, as a representative region for aerodynamic noise generation, merits further investigation into its scale effects. In this paper, the large-eddy simulation (LES) and the Ffowcs Williams-Hawkings (FW-H) integral equation are used to calculate and analyze the sound energy intensity distribution pattern and spectral characteristics of the PCCS at different scales (1/1, 1/2, 1/4, 1/8, 1/16, 1/25, 1/50). The research shows that as the scaled model decreases, the relative area of the pantograph submerged by the vehicle boundary layer increases, and its inflow velocity decreases, thereby reducing the overall radiated sound pressure level in this area. For the segments 1/1–1/2 and 1/4–1/16, the dominant scale of sound generation is typical pure tone noise, with distinct similar features in the spectral discrete scales. For the segments 1/25–1/50, the turbulent fluctuation characteristics of the vehicle boundary layer mask the peak features, and the spectrum is dominated by broadband characteristics. Combining the PCCS sound source energy scale correction model and the dimensionless spectrum correction function, a scale correction model for the sound power spectrum of the sound source is obtained, so that the noise results of the reduced-scale model can be corresponded to the full-scale model. This work advances the comprehension of high-speed train aerodynamic noise generation mechanisms and offers critical references for developing precision noise control technologies.

Maglev trains experience significant aerodynamic effects when passing through tunnels. A moving model test was conducted to explore the practical effects of speed reduction and entrance buffer structures on mitigating tunnel/maglev aerodynamic effects. It is found that both have an overall positive effect on mitigating the aerodynamic environment inside and outside the tunnel. Trains operating at 200 km/h show a 49.8% decrease in peak-to-peak pressure and a 50.7% decrease in transient pressure instability on inner walls compared to those at 280 km/h. Lower speeds resulted in a 65.6% decrease in amplitude and a 24.5% decrease in decay rate, both of which are parameters for exponential fittings of pressure peaks that decay naturally after the train leaves. The buffer structures result in a reduction of up to 25.7% in the maximum positive pressure and a 29.0% decrease in transient pressure instability. Additionally, a reduction in amplitude of up to 21.2% and a 32.2% increase in decay rate were observed with the use of buffer structures. Nevertheless, it is difficult to conclude direct correlations between the maximum pressure, peak-to-peak values, etc., and the speeds or buffer structures due to the complex wave propagation in tunnels. However, speed reduction and buffer structures are proven to be effective in reducing the micro-pressure wave levels with a simpler monotonic relationship.

The airflow around a vacuum tube maglev train operating at high speeds is complex. In addition, the effect of relevant parameters in such a transportation system on aerodynamic characteristics is crucial in the design and safety of the system. A three-dimensional (3D) vacuum tube train model is established based on a vacuum tube test platform for rail transit. The effects of the initial ambient temperature and scale ratio on the aerodynamic characteristics are analyzed during the whole operational process in this study. The results mainly focus on each process’s variations in the shock waves, choked flow, and drag. During acceleration, shock wave generation is advanced or delayed under different system parameters, which vary the aerodynamic drag. While the train runs at a constant speed, the time that a standard shock is generated and the length of the choked flow differ under the effects of the varying system parameters. In braking, the disappearance of shock waves and reflections of the expansion wave suddenly decreases the aerodynamic drag either earlier or later due to the varying system parameters.

Evacuated tube transportation (ETT) offers a promising high-speed transport solution, but trains operating at supersonic speeds within a sealed tube can induce complex aerodynamic phenomena that impact safety and reliability. This study utilized the Reynolds-averaged Navier-Stokes (RANS) shear stress transport k-ω (SST k-ω) turbulence model for steady-state simulations and the improved delayed detached eddy simulation (IDDES) SST k-ω model for unsteady-state simulations, both coupled with the advection upstream splitting method (AUSM). Four tunnel cross-sectional areas (49 m², 64 m², 81 m², and 100 m²) with corresponding blockage ratios (β) (0.253, 0.192, 0.150, 0.121) were analyzed to explore shock wave formation and its dependence on blockage ratios, along with surface pressure distribution and aerodynamic loading. Results show that higher blockage ratios increase shock wave intensity, while larger tunnel areas reduce this intensity, improving flow structure and wake effects. Moreover, as the blockage ratio decreases, the total drag coefficient of the entire train decreases linearly. When the blockage ratio decreases from 0.253 to 0.121, the total drag coefficient of the entire train decreases by 46.2%, with the head carriage and tail carriage drag coefficients decreasing by 23.3% and 32.7%, respectively, while the drag coefficient of the middle carriage remains nearly unchanged. The percentage of the total drag coefficient contributed by the head carriage decreases from 51.1% to 40.9%, while the percentage for the tail carriage increases from 47.0% to 56.6%. These findings enhance understanding of ETT fluid dynamics and performance.

Water-rich cracks represent common tunnel defects. Intense pressure waves generated by trains traveling through tunnels may undergo enhancement within water-rich cracks. Using the re-normalization group (RNG) k - ε turbulence model and volume of fluid (VOF) method, this study analyzes the spatiotemporal distribution, spectral features, and influencing factors of pressure wave propagation in water-rich cracks when two high-speed trains intersect in a tunnel. The flow mechanisms underlying the pressure enhancement within water-rich cracks are also revealed. The main conclusions are as follows: 1) The positive and negative peak pressure coefficients in water-rich cracks are 1.34 and −2.36, with corresponding pressure gradient peaks of 31.41 kPa/s and −34.01 kPa/s. Compared to the tunnel wall, the peak pressure coefficients and gradients exhibit increases of 34.41%/44.63% and 31.61%/60.46%, respectively. 2) The dominant frequency of the pressure wave power spectral density (PSD) at the crack tip is 26.97% higher than that in the tunnel. The PSD peak value continuously increases with depth and is the largest at the crack tip, representing an increase of 9.36% compared to the tunnel. 3) An increase in crack width reduces the peaks of pressure waves, pressure gradients, and PSD, while increases in vertical and transverse depths amplify these peaks. Crack width has the most significant impact on pressure waves and pressure gradients, while transverse depth has the most significant effect on PSD peak values. 4) Driven by inertia and pressure differences, the water body oscillates variably, enhancing pressure fluctuation amplitude at the crack tip. The higher the water body’s movement velocity, the greater the pressure gradient at the crack tip. The above research results may provide a reference for crack harnessing in high-speed railway tunnels.

The aerodynamic pressure disturbances induced by middle air shafts and bypass ducts in subway tunnels pose significant challenges to enhancing train operational speeds. A comprehensive series of full-scale experiments are employed to examine the impact of these structural elements on the aerodynamic pressure characteristics of platform screen doors (PSD) in high-speed subway stations. The experimental results reveal that peak pressures manifest on PSD surfaces during two distinct scenarios in high-speed subway systems equipped with middle air shafts. One is compression pressure waves propagated from trains traversing the air shaft, and the other is train nearby flow when trains pass the PSD directly. The peak positive pressures caused by train passing PSD is much greater than compression pressure waves. Closing middle air the shaft can reduce the passing pressure waves. The installation of bypass ducts at overtaking station entrances effectively mitigates peak negative pressures during train-PSD interactions, achieving a maximum reduction efficiency of 8%. These findings provide valuable insights for optimizing the structural design of high-speed subway tunnel systems.

To address the severe aerodynamic effects caused by a 600 km/h superconducting maglev train passing through a tunnel at full speed, this study systematically investigates the coupled influence of auxiliary facility parameters including the shaft (location L, cross sectional dimension W, height h), tunnel portal (cross sectional area S), and openings (spacing D, side length F) on the evolution of tunnel aerodynamic effects. By integrating three dimensional unsteady flow field numerical simulations with a dynamic model testing system, the research notably reveals the regulatory mechanisms of these parameters on the evolution characteristics of the initial compression wave pressure gradient and the multi peak structure of micro-pressure waves. The results show that shaft parameters significantly affect the initial compression wave. Both the wave amplitude and gradient exhibit a linear negative correlation with cross sectional dimension W and a linear positive correlation with location L, while demonstrating a nonlinear relationship with height h, the amplitude follows a cubic polynomial trend, and the gradient initially increases before plateauing. Under the configuration W=8 m, L=50 m, and h=20 m, substantial reductions in both compression wave amplitude and gradient were achieved. The portal cross sectional area S shows a “U-shaped” relationship with the compression wave gradient, with the maximum gradient reduction of 53.24% occurring at S=210 m², a result comparable to that achieved with optimized opening parameters (D=15 m, F=3.5 m, 53.96%). Regarding micro-pressure waves, the amplitude measured 20 m from the tunnel exit shows a linear positive correlation with shaft parameters L and W, while the influence of h saturates beyond 50 m. Reductions exceeding 54% were achieved with portal parameters, either at S=210 m² or using the optimized opening configuration. Furthermore, micro-pressure waves near the portal exhibit a consistent dual peak structure: the first peak originates from the train entry compression wave, and the second results from further wave compression after tunnel exit. The opening location governs selective peak regulation openings near the portal entrance primarily suppress the first peak with minimal impact on the second, whereas centrally located openings reduce the first peak but can amplify the second by up to 3%. Based on these insights, an optimized parameter configuration is proposed: a shaft with a cross-sectional dimension ⩾8 m located 50 m from the portal, a portal cross sectional area of 210 m², and openings spaced at 15 m intervals. This configuration can reduce the initial compression wave gradient by over 50%. The results provide a theoretical foundation for controlling aerodynamic effects of superconducting maglev train.

This study innovatively employs functional near-infrared spectroscopy (fNIRS) technology to investigate passengers’ brain responses to various external stimuli during high-speed train operations, assessing their impact on passenger comfort. Three stimuli are examined: passing through tunnels, sonic booms at tunnel exits, and two trains meeting within the tunnel. The analysis of environmental variables, including cabin noise, cabin-to-external pressure, and cabin-to-body acceleration, reveals that changes in auditory and pressure levels during the tunnel experience led to an 87% increase in oxygenated hemoglobin (HbO) levels in the temporal lobe (TL). This reflects a brief discomfort that subsides as passengers adapt, with HbO levels nearly returning to pre-tunnel levels upon exit. Among the stimuli, the sonic boom triggered the most significant neural response, with HbO fluctuations increased by 175%. In contrast, the impact of train meetings was minor, yielding an average HbO increase of only 14.21%. Connectivity analysis further shows significant enhancements in brain functional connectivity during tunnel entrance and sonic boom scenarios, with increases of 52% and 80%, respectively. Our findings contribute to passenger comfort assessment by establishing objective neurophysiological measures that quantify previously subjective experiences. The application of fNIRS in this dynamic environment creates new possibilities for evidence-based comfort optimization in railway design.

With increasingly stringent requirements for the airtightness of high-speed train bodies, determining appropriate airtightness levels has become critically important. To calculate the airtightness of high-speed train bodies more accurately, based on one-dimensional isentropic flow theory, this study derives cabin pressure calculation models for both positive and negative pressure conditions during static airtightness tests of high-speed train bodies. Since the flow coefficient, which is closely related to the leakage characteristics of the carriage, is influenced by multiple factors including operating pressure conditions (positive/negative), leakage path cross-sectional shape, and size, a flow coefficient calibration method is proposed to achieve high-precision and efficient calibration of the flow coefficient for trains with varying leakage properties. This method generates a series of flow coefficient values for circular and square cross-sectional shapes under both positive and negative pressure conditions across various cross-sectional areas. Furthermore, functional relationships between flow coefficient and leakage path area under positive/negative pressure are established through curve fitting. Using these functional relationships and the cabin pressure calculation model, the pressure variation curves for a static airtightness test are simulated. Specifically, for circular cross-sectional shapes, the theoretical curves under positive and negative pressure conditions exhibited R² values of 0.9936 and 0.9931, respectively, when compared to experimental data, and for square cross-sectional shapes, the corresponding R² values are 0.9928 and 0.9932, validating the accuracy of the proposed theoretical model. The proposed theoretical model effectively evaluates the airtightness of high-speed train bodies with varying performance levels during static airtightness tests, providing a robust theoretical reference for optimizing high-speed train airtightness design.

The pressure comfort of passengers and crew in high-speed trains faces significant challenges under alternating open-tunnel conditions. To better understand the mechanism of pressure transmission and control interior pressure fluctuations in high-altitude regions, this study develops an interior pressure fluctuation model. By establishing the frameworks of the non-ideal gas state equation and the polytropic process equation, gas heat transfer and mass transfer were expressed through the first law of thermodynamics and the continuity equation. Simulation results, evaluated by root mean square error, coefficient of determination, peak-to-peak error, and pressure change rate, show that the proposed model closely aligns with measured signals in both overall trends and local details. Data from various train types and tunnel scenarios further demonstrate the model’s accuracy and practical applicability. This study provides a critical foundation for evaluating interior pressure comfort for high-speed trains in high-altitude regions.

Most studies have analyzed the aerodynamic characteristics and wind-train(vehicle)-bridge coupled vibration response of trains or vehicles on bridges of a certain structural system, while few comparative studies have been carried out on the wind-train-bridge coupled vibration response on bridges of three different structural systems. This paper takes the main span 1120 m dual-purpose highway-railway bridge as the engineering background, and studies the three bridge types of (122+1120+90+92) m suspension bridge, (130+432+1120+432+130) m cable-stayed bridge and (92+210+1120+210+92) m cable-stayed-suspension collaborative system bridge. The trend of the maximum value of the train dynamic response to the wind-train-bridge coupling of the three structural system bridges as well as the speed thresholds are compared and analyzed, and conclusions are drawn: 1) Under the same speed, the maximum value of train safety indexes in three types of bridges increases with the increase of wind speed. 2) Under the same wind speed, the safety and smoothness indicators of trains in three types of bridges without wind barriers rank in the order of cable-stayed-suspension collaborative system bridge>cable-stayed bridge>suspension bridge. 3) At low wind speeds (≤15 m/s), a 3.0 m wind barrier has negligible effect on speed thresholds. The safety ranking of structural systems remains unchanged: cable-stayed-suspension collaborative system bridge>cable-stayed bridge>suspension bridge. 4) At high wind speeds (≥20 m/s), the 3.0 m wind barrier can increase the train speed threshold for bridges within the same structural system. The safety ranking of the three bridge types (3.0 m 30% wind barrier) remains unchanged: cable-stayed-suspension collaborative system bridge>cable-stayed bridge>suspension bridge. This study represents the first systematic comparative analysis of wind speed critical values and performance ratings across three distinct bridge structural systems.

The aerodynamic performance of a high-speed train deteriorates sharply under crosswind, severely affecting its operational safety. This paper adopted a three-car high-speed train as the benchmark and established leeward side (LWS) airbag-train models. Based on the three-dimensional steady SST k-ω two-equation turbulence model, this study investigated the aerodynamic characteristics of trains under crosswind at three different airbag’ s installation positions. The results show that the airbags installed on the LWS change the surface pressure distribution on the LWS of the train body, lowering the lateral force coefficient and overturning moment coefficient, and the aerodynamic performance of the train under crosswinds is enhanced. The airbag structure located at the top of the LWS (Model III) shows the most significant improvement in crosswind performance that the lateral force coefficient is reduced by 16.71%, and the lift coefficient is increased by 17.95%, which collectively led to a decrease in the train’ s overturning moment coefficient by 23.65%. The research findings provide a reference for improving the anti-overturning performance of the next generation high-speed trains under crosswind.

The influence of train height on aerodynamic characteristics of high-speed train (HST) is significant in crosswind environments. This study employed the improved delayed detached eddy simulation (IDDES) turbulence model to analyze the aerodynamic characteristics of trains with three different heights under a crosswind of 20 m/s. The numerical model was validated through comparison with wind tunnel experimental data. A comprehensive analysis was conducted on the characteristics of the flow field around trains, surface pressure distribution, and aerodynamic loads for trains with different heights. Results indicate that the side force coefficient increased by up to 61.54% with an increase in train height from 3.89 to 4.19 m. Compared with the 3.89 m case, the roll moment coefficient on the head, middle, and tail cars for 4.19 m cases increased by 18.11%, 24.78% and 34.23%, respectively. The increase in train height widens the impact width of the leading car’ s front vortex on the leeward side and intensifies the helical shedding and coupling interactions of two vortices in the wake, leading to an increase in the intensity and extent of wake flow in both vertical and longitudinal directions. Additionally, the increase in height shifted the flow separation point on the leeward side, moving vortices farther from the train, expanding the back-flow region, and intensifying Reynolds stress and turbulent fluctuations on the leeward side, which adversely impacted train stability and safety. The research findings can provide a reference for the design of train configurations and the assessment of dynamic performance in crosswind environments.

Considering passenger trains’ key role in remote regions, this study employed machine vision technology to monitor five posture parameters of the second car of a conventional passenger train, aiming to investigate the influence of windbreaks and crosswinds along railways on the operating postures of conventional passenger trains. The study found that when passing through the anti-wind tunnel with holes, the amplitudes of posture parameters were smaller than those of other windbreaks, demonstrating the superior performance of this windbreak in maintaining posture stability compared to others. In tunnel sections, larger amplitudes of these parameters were observed for the tail car than the head car, while the opposite occurred in non-tunnel sections. Notably, during tunnel transit, their amplitudes did not increase monotonically with speed but peaked at a specific speed that most adversely affected the operating posture. These conclusions have a great significance for improving operating safety under crosswinds.

The stability of high-speed trains under crosswind conditions has become a key consideration in aerodynamic design. As running speeds continue to increase and car body weight decreases, crosswinds pose a greater risk to train safety, significantly lowering the critical wind velocity. Therefore, developing strategies to enhance crosswind stability is essential. This study focuses on the leeward region adjacent to the train body, where separated flows with large vortices generate significant negative surface pressure. Enhancing this negative pressure distribution is proposed as a potential method to improve a train’s resistance to overturning. To achieve this, winglets are installed on the leeward side as a flow control measure, and their effects at different deflection angles are evaluated. The influence of five deflection angles on the leeward-side flow field and aerodynamic loads is analyzed, considering the head, middle, and tail cars. Results indicate that a deflection angle of 90° optimally reduces the overall overturning moment by 27.6% compared to the baseline model in a three-car configuration. These findings highlight that optimizing the winglet deflection angle to approximately 90° can significantly enhance a train’s resistance to overturning, offering valuable insights for aerodynamic optimization in strong wind conditions.