Wheel polygonal wear induces strong periodic wheel-rail vibrations, threatening ride security and increasing maintenance costs. Data-only deep-learning diagnostic solutions often generalize poorly across operating conditions and rely heavily on large labeled datasets, largely because they ignore the frequency-dependent and multi-wheel coupling physics characteristics of wheel-rail dynamics. We propose Co-FRF-NN, a physics-guided network that embeds wheel-rail dynamic priors through a learnable 4 & times; 4 complex MIMO FRF-Net, enabling frequency-dependent and cross-wheel coupling effects to be modeled directly within the architecture. A threedomain decoupled design-frequency features (C-CNN), spatial MIMO transfer (FRF-Net), and temporal evolution (C-LSTM)-aligns learning with physical mechanisms, substantially improving robustness and data efficiency. Simulation and field results show that Co-FRF-NN markedly surpasses data-only baselines, delivering higher accuracy, stronger cross-speed generalization, substantially improved data efficiency (e.g., matching baseline performance with only 20% of the data), and robust suppression of cross-wheel interference. These results show that physics-guided deep learning is an effective and scalable strategy for wheel-rail condition monitoring, and suggest its potential applicability to intelligent fault diagnosis in other complex electromechanical systems.

Accurate assessment of high-frequency wheel-rail forces (HWRFs) is essential for ensuring operational safety and supporting structural health monitoring in railway systems. Vehicle dynamic response-based force identification offers a more practical and economical alternative to direct measurements. However, conventional identification models typically assume a rigid wheelset, neglecting structural modal vibrations and thus resulting in significant estimation errors in the high-frequency range. To overcome this limitation, this study proposes an improved dynamic load identification model incorporating both rigid-body motions and structural modal vibrations of the wheelset. A frequency response analysis of axle box acceleration (ABA) is first conducted to determine the essential wheelset modes required for accurate HWRFs estimation. Subsequently, a high-fidelity virtual test vehicle model is developed to simulate ABA responses under wheel polygonization and rail weld joint excitations. An augmented Kalman filter, integrated with dummy displacement measurements, is designed to reconstruct HWRFs from ABA signals. Validation results demonstrate that the proposed rigid-flexible coupled identification model accurately estimates HWRFs, achieving an average reduction in estimation error of approximately 30 % within the 30 similar to 1000 Hz frequency range compared to a conventional rigidbody-based approach. These findings confirm the effectiveness of the proposed approach and highlight its promising potential for practical engineering applications.

Tread modification is an emerging wheel profile maintenance technique. Compared to wheel re-profiling, it performs incremental corrections during vehicle operation, extending service life while reducing maintenance costs. However, suboptimal operational strategies may negatively impact its effectiveness. Therefore, refining optimization methodologies for tread modification strategies is crucial. Addressing this, based on the wheel profile comprehensive wear (WPCW) prediction model incorporating tread modification, this study reveals the initial desynchronization phenomenon (IDP) and analysis its mechanism. Subsequently, a novel adaptive optimization strategy is proposed, overcoming the limitations of conventional fixed-parameter approaches. Results demonstrate that the strategy fosters more intelligent tread modification implementation, effectively balancing vehicle dynamic performance with wheel maintenance economic efficiency.

Tread modification has gained significant attention in recent years as a means to address the issue of wheel hollow wear.The wear resulting from tread modification can alter the wheel profile,thereby impacting the wheel-rail contact relationship and vehicle dynamics performance.Consequently,it is crucial to understand the influence of tread modification on wheel wear.This study proposes a prediction method for the wheel profile's comprehensive wear(WPCW)for high-speed trains,considering the impacts of both the wheel-rail interaction and the tread modification on the wheel profile comprehensive wear.First,simulation models are established to quantify wheel wear resulting from wheel-rail interactions and tread modification.Subsequently,the coupling relationship between the two models is subsequently strengthened by incorporating iterative calculation processes,resulting in the prediction model of the wheel profile comprehensive wear.Finally,the prediction method is calibrated and verified through measured data.The simulation results obtained using this method align with the measured results,confirming the feasibility of the proposed prediction method and its applicability in predicting the WPCW for high-speed trains.

The hunting stability and curve-passing ability of railway vehicles are inherently contradictory problems, and hydraulic bushings can effectively improve these contradictions due to their strong nonlinear characteristics. This paper presents a refined nonlinear mechanical model for hydraulic bushings, incorporating both frequency-dependent and amplitude-dependent stiffness characteristics, based on their structural parameters and operating principles. The model is developed by combining a rubber spring model, an inertia channel model, and a volume flexibility model to capture the complex dynamic characteristics of the bushing. To validate the model's accuracy, dynamic characteristic experiments are performed on hydraulic bushings using a suspension component performance test bench. The experimental results show good agreement between the simulated dynamic stiffness and equivalent damping values and the measured data. The proposed model is then integrated into vehicle dynamics simulations to evaluate the impact of hydraulic bushings on vehicle stability and curve-passing performance. Compared to traditional rubber bushings, vehicles equipped with hydraulic bushings demonstrate a significant increase in nonlinear critical speed on straight tracks and improvements in wheelset yaw angle and lateral wheel force when passing through curved tracks. These results highlight the ability of hydraulic bushings to effectively enhance both vehicle stability and curve-passing performance, offering potential advantages in vehicle suspension design.

To enhance the operational quality and safety of electric multiple units (EMUs), this study investigates parameter optimization for their underfloor suspension dampers based on vertical dynamic performance and addresses service state management. A rigid-flexible coupled vehicle dynamic model was established, which accounts for the elastic vibration characteristics of car bodies to accurately simulate relevant behaviors. On this basis, the influence patterns of design parameters on the vehicles' operational quality were systematically analyzed, from aspects such as center-of-mass offsets resulting from the structural layout of suspension dampers, damping ratio matching in resonance models, and damping characteristics. From the analysis results, an optimized matching scheme for key suspension parameters was derived. To reveal the service state of these dampers, an in-depth analysis was conducted to examine the effects of performance degradation and typical faults of key components on the vehicles' vertical dynamic performance, which verified the optimized scheme provides sufficient safety margins to accommodate performance degradation. Research results show that parameter optimization significantly improves the vehicles' vertical dynamic performance and underscore the need for effective performance degradation monitoring during service. This study provides a theoretical basis for refined parameter design and performance optimization of underfloor suspension dampers for high-speed trains, with important implications for ensuring the long-term safe and stable operation of EMUs.

To investigate the possible formation and evolution mechanisms of rail corrugation related to the resonances of wheel and rail, a wave approach is adopted to characterize the wave propagation and vibration behaviour of the track. Through the receptance coupling method, the interaction model of multiple wheels and track is established taking into account the P2 resonance and resonances due to the wave interaction, which is applied in combination with the wheel/rail rolling contact model and the Archard wear equation to demonstrate the formation and development of the rail corrugation in both low-speed metro and high-speed rail systems. The numerical results show that the resonances due to the wave interaction between adjacent wheels can lead to short-pitch corrugation with a wavelength of 20 mm-30 mm in the metro system and long wavelength corrugation of 110 mm-160 mm in the high-speed railway, which is supported by corresponding field measurements. It is also noted that the short-pitch corrugation in the metro and the long wavelength corrugation in the high-speed railway share the common frequency range of 550-700 Hz, which is explained from the perspective of the quasi-resonance phenomenon of the multiple wheels and track coupled system.

Accurately assessing dynamic loads on railway vehicles is crucial for ensuring safety, refining vehicle design, enhancing passenger comfort, and streamlining maintenance procedures. However, direct load measurement methods often incur high costs due to challenges associated with precise force transducer positioning and the necessity for numerous sensors. Therefore, leveraging a limited set of accessible vibration response signals for real-time dynamic load identification proves to be a valuable approach. Real-time load identification algorithms, such as the augmented Kalman Filter (AKF), typically require a full-rank observability matrix to achieve stable results, which often involves measuring some or all displacements of each component. Nevertheless, accurately measuring absolute displacements is challenging due to the inherent mobility of railway vehicles along the track. To overcome this challenge, a state reconfiguration (SR) method has been developed within the state-space equation framework. This method transforms absolute displacements of all Degrees of Freedom (DOF) into readily measurable relative displacements, effectively meeting observability requirements. Furthermore, a refined optimal sensor placement strategy (OSPS) has been devised based on observability criteria and minimizing the normalized steady-state error covariance of the estimated loads. The proposed SRAKF combined with OSPS has been validated using a simple numerical vehicle model under three load conditions. Subsequently, a virtual test utilizing a detailed railway vehicle model was conducted, demonstrating the applicability and feasibility of the proposed method in engineering applications.

The bogie system, a critical component of railway vehicles, endures complex cyclic loads from wheel-rail interactions, vehicle motion, and traction-braking forces. Progressive wheel out-ofroundness (OOR) amplifies dynamic loads, induces high-frequency resonance, and accelerates fatigue damage in bogie sub-components such as the sanding device. This study proposes a fatigue life prediction framework that integrates axle box vibration spectrum degradation and wheel-rail coupled vibration into frequency-domain damage estimation. The vibration spectrum evolution is modeled and predicted using a nonlinear Wiener process, capturing the stochastic nature of spectral degradation. The framework combines rigid-flexible coupled simulations, field vibration tests, and frequency-domain fatigue algorithms to quantify the impact of spectral shifts on damage accumulation. The results show that considering both spectral evolution and wheel-rail coupled vibration leads to the predicted earlier failure, with the sanding device's remaining useful life (RUL) reduced by up to 75 % under severe OOR conditions. This method enables realtime fatigue prediction and iterative recalibration, supporting condition-based maintenance and fatigue-resistant design, ultimately enhancing the reliability and economic efficiency of railway vehicle operations.

Axle box vibration serves as the main source of excitations for rail vehicles. Due to the wear of wheel/rail contact and the re-profiling procedure, the axle box vibration usually degrades periodically with the increased mileage in the service. This could significantly impact the estimation of vibration fatigue when the component is subjected to the axle box vibration. This paper develop a method to describe the periodic evolution of axle box vibration spectrum to better characterise the vibration spectrum of axle box. In this study, a Wiener process incorporating with four random parameters was employed to model the non-linearity of the degradation process. The maximum likelihood estimation (MLE) algorithm is used to estimate the initial values of the random parameters, and a Bayesian approach is employed to update the parameters based on newly obtained data. Finally, the proposed methodology is tested using long-term field test data from a high-speed train, and the results demonstrate that it accurately estimates the evolution of the axle box acceleration spectral density (ASD) spectrum. This could aid in predicting the residual service life of structures subjected to axle box vibration and further contribute to the development of maintenance strategies and top-down design of the structure.

This study was primarily focused on mitigating high-order polygonal wear (PW) in China's high-speed trains. To achieve this, a fast PW trend analysis strategy and a long-term PW iterative simulation were developed. The rail localized bending modes(RLBM) were identified as critical factors influencing the amplitude-frequency and phase-frequency properties of the wheel/rail force response, which in turn contributed to the occurrence of high-order PW. It was found that the phase lag of the wheel/rail force response relative to wheel excitation played a significant role in determining the growth trend of PW. Therefore, the design of a rail-tuned dynamic vibration absorber (DVA) was crucial to manipulate the phase lag and force it out of the positive growth region, specifically between 90° and 270°. Detailed simulations found that a rail DVA with a tuned frequency of 450 Hz effectively slowed down the development of the 20th order PW (578 Hz). This research provides valuable insights into the design principles for rail DVAs aimed at mitigating periodic wear between the wheel and rail.

The stress intensity factor is essential to determine the residual fatigue lifetime of cracked components. However, the traditional method used to determine the stress intensity factor is very difficult to characterize the dynamic behaviors of structure arising from the fatigue crack. This paper proposed a methodology to determine the dynamic stress intensity factor (DSIF) of a crack incorporating with the rigid-flexible coupled dynamics. Firstly, a rigid-flexible coupled dynamic model of cantilever beam representing the typical structure of bogie frame with a straight crack was developed based on SIMPACK platform. In the model, the equivalent spring - contact element was employed to model the contact behaviors of crack interface. Subsequently, the DSIF considering the crack closure effect was calculated by the node displacement extrapolation. The validity of proposed method was demonstrated through comparing the DSIF with those obtained by other analytical methods. The results suggest that the proposed methodology can characterize the closure effect of crack, and yield more accurate estimation for the DSIF. This method establishes a link between the multibody system dynamics and the fracture mechanics, which enable us to calculate the DSIF of crack under the operating condition using a rigid-flexible coupled dynamic model.

High frequency impacts arising from wheel/rail short pitch irregularities have been widely reported as main casual factor of high frequency vibration of bogie frames recent years. This study aims at exploring characteristics of high frequency vibration of a metro bogie frame and its control methodologies. Firstly, the measurements obtained from a field test were employed to demonstrate the elastic vibration of bogie frame due to rail corrugation-induced impacts. Secondly, a rigid/flexible coupled dynamic model neglecting the wheel/rail contact model was established to simulate the vibration of bogie frame. Subsequently, the control methodologies were proposed to suppress the high frequency vibration of bogie frame. The results showed that the end of bogie frame is predominated by the localized bending vibration mode at 220 Hz, which is close to the passing frequency of rail corrugation at the rubber booted short sleeper track. The structural resonance serves as the main driving force of fatigue failure for the end of bogie frame. The structural improvement through installing a stiffener at end of bogie frame can effectively suppress the local bending of end of bogie frame. The piezoelectric-based actuator could serve as an alternative method to reduce the vibration level for considered frequency range.

The axle box bearing serves as a critical rotational component for high-speed trains. Thus, it is of great significance to obtain accurate loading conditions for axle box bearings for its reliability design. In this study, a bearing-vehicle coupled system dynamic model, consisting of an equivalent bearing model characterized by the time-varying stiffness in the bearing, a nonlinear high-speed train dynamic model, and a slab track model, is developed. The validity of the proposed coupled dynamic model is demonstrated by comparing the simulated results with those obtained in a field test and the literature, respectively. Subsequently, the influence of long- and short-wavelength irregularities on the dynamic responses of bearings is investigated using the proposed coupled model in terms of dynamic loads of the axle box bearing. The results obtained by the proposed model are further compared with those calculated using a traditional vehicle dynamic model that represents the bearing as a revolute joint, so as to illustrate the advantages of the proposed model. The results suggest that the loading responses of the bearing calculated by the two models exhibit differences, which are simultaneously related to the structural damping of the bearing and the mode frequency change of the system..

The performance of the yaw damper for rail vehicles degrades significantly degradation after long-term operation, which directly affects the dynamic behaviour of the vehicle operation. To quickly achieve the performance degradation of the yaw damper and predict the operating mileage under actual service conditions, the operating displacement spectrum of the yaw damper under actual operating conditions is analysed and decomposed using the short-time Fourier transform. Based on the equivalent principle of energy loss during vibration attenuation, an accelerated test method is proposed for the performance degradation of yaw dampers. By conducting the accelerated degradation test of the yaw damper on the test bench, the continuous degradation relationship between the performance parameters of the damper and the loading times is obtained. By combining the operating displacement spectrum and the accelerated degradation test data, the performance degradation trajectory model of the yaw damper is established, taking into account the equivalent energy loss between the service mileage and the loading times. The mapping relationship among loading times, service mileage and performance degradation is constructed. Finally, parameter performance tests are conducted on yaw dampers under different service mileage to verify the accuracy of the performance degradation trajectory model. The results show that the proposed acceleration test method and degradation trajectory model can quickly realise the performance degradation of the damper and accurately predict the performance parameters for different service mileage under actual operating conditions.

The sub-components of railway bogie have been frequently reported to be subjected to the vibration fatigue due to the wheel/rail high frequency vibration. An efficiency numerical model is thus desirable to simulate the high frequency vibration of railway bogie, which can substantially enhance the design efficiency of railway bogie system considering the vibration fatigue. Therefore, this paper aims to proposing a modelling methodology for simulating the high frequency vibration of railway bogies’ subcomponents, and a lifeguard of metro bogie was taken as an example. Firstly, a field test measurement of lifeguard for the metro bogie was primarily introduced to demonstrate the high frequency vibration characteristics and the related failure mechanisms arising from the wheel/rail high frequency impact. Subsequently, a random vibration model for the railway bogie system and the lifeguard based on the rigid/flexible coupled dynamics were developed to simulate the high frequency vibration and the dynamic stress developed at the lifeguard. This method was subsequently employed in the structural optimization for the lifeguard. The results showed that the proposed methodology can effectively simulate the high frequency vibration of lifeguard of bogie system based on the measured axle box acceleration, and the optimized structure can effectively increase the service lifetime under the excitation of wheel/rail high frequency vibration.

The wheel-rail contact geometry strongly affects the dynamic running behavior of a railway vehicle, so it is a key factor for its operational safety. Although wear in the wheel-rail contact is generally inevitable, a particular problem is the occurrence of nonuniformly distributed wear, which changes the wheel profile and thereby often leads to a deterioration of the vehicle's running behavior. Therefore, the restoration of the wheel profile, known as reprofiling, is an important maintenance action for ensuring operational safety. The exact restoration of the original wheel profile, however, often requires the removal of a high volume of material, which can drastically shorten the service life of the wheel. An incomplete restoration of the wheel profile, which requires the removal of less material, extends the wheel's service life and is therefore also known as “economic reprofiling”. Nevertheless, an incompletely restored wheel profile, too, must still fulfill the same requirements as the original profile with respect to the vehicle's running behavior and thereby to its operational safety. While previous work mainly dealt with the impact of flange wear, the present study focuses on the wheel tread frequent contact area (WTFCA). This study presents a long-term wheel wear prediction model for a high-speed train and proposes a design method for the wheel profile in the WTFCA for economic reprofiling. After optimizing the key factors, the utilization of the wheel profile for economic reprofiling can comprehensively consider the dynamic performance, wear behavior, and economic performance of the vehicle, which can be applied to guide the maintenance of the wheel profile.

Potential manufacturing defects and dynamic loads in the service could introduce cracks in the bogie frame. These cracks, in turn, can vary the stiffness matrix of system, thereby the variations in dynamic behaviors of bogie frame. This study thus investigated dynamic behaviors of a bogie frame in the presence of fatigue crack through a field test and a numerical model. In the test, the axle box vibrations were measured, which serves as the excitation of the numerical model of bogie frame. Subsequently, a rigid/flexible coupled dynamic model of vehicle was developed, considering the flexibility of bogie frame. A methodology, based on the equivalent spring and contact elements, was developed to model fatigue cracks in the dynamic model. This enables us to simulate coupling behaviors between the vibration and the crack propagation. Upon the proposed model, the evolution of dynamic behaviors of bogie frame considering the crack propagation process was studied under the services loading. The results show that the cracks significantly alter the vibration behavior of the bogie frame, leading to a noticeable response in the high-frequency components. This facilitates further research on crack damage detection and localization. The dynamic stress intensity factor (DSIF) obtained through displacement extrapolation method accurately describes the dynamic propagation behavior of crack tips. This helps us understand the crack propagation models better, thereby the estimation of residual fatigue life of bogie frame.

With the increasing speed of high-speed trains, the wear of wheel and rail is becoming increasingly serious. It is difficult to take into account the contradiction between high-speed stability and curve negotiation capability of vehicles and the emerging new contradiction by setting conventional yaw dampers with fixed parameters. In order to further improve the comfort and high-speed stability of passenger trains and reduce the wear of wheel and rail, it is necessary to design a semi-active yaw damper solution. In this paper, an innovative design prototype of semi-active yaw damper is proposed and its numerical simulation model is established. The accuracy of the model is verified by bench test. From the perspective of dynamics, a control strategy of the yaw damper is proposed based on a real-time curve identification strategy. In order to verify the effectiveness of the implementation of the solution in solving the above problems at the same time, three different comparative operating conditions are designed through the roller rig and the numerical simulation model.

The axle box bearing is a critical component of high-speed trains. Investigating the thermal characteristics of axle box bearings under in-service conditions is essential for developing an effective temperature monitoring strategy. In this study, a non-linear thermal field model for axle box bearings, considering the vehicle-environment coupling effects, is established based on the bearing-vehicle coupled dynamics and finite element method. The validity of the proposed thermal model is demonstrated by comparing the temperature results with those obtained from a field test and calculated via the traditional method. Furthermore, the influence mechanism of vehicles and environment on bearing temperature was revealed, and the thermal characteristics of the axle box bearing under in-service conditions were analyzed. The results show that the vehicle affects the bearing temperature through the boundary load and wheelset speed, and these two parameters can have a significant impact on the power loss of the bearing and the convective heat transfer coefficient of the grease. The environment influences the bearing temperature by affecting the convective heat transfer coefficient on the surface of the axle box and wheelset. In addition, the track line parameters, such as curve radius and superelevation, may lead to the changes in bearing loads, consequently affecting the thermal behavior of the bearings. The maximum temperature of each bearing component increases with higher vehicle speed and higher ambient temperature. It is thus essential to consider the vehicle-environment coupling effects when analyzing the temperature characteristics of axle box bearings under operating conditions.