Irregularities in the geometry and flexibility of railway crossings cause large impact forces, leading to rapid degradation of crossings. Precise stress and strain analysis is essential for understanding the behavior of dynamic frictional contact and the related failures at crossings. In this research, the wear and plastic deformation because of wheel–rail impact at railway crossings was investigated using the finite-element (FE) method. The simulated dynamic response was verified through comparisons with in situ axle box acceleration (ABA) measurements. Our focus was on the contact solution, taking account not only of the dynamic contact force but also the adhesion–slip regions, shear traction, and microslip. The contact solution was then used to calculate the plastic deformation and frictional work. The results suggest that the normal and tangential contact forces on the wing rail and crossing nose are out-of-sync during the impact, and that the maximum values of both the plastic deformation and frictional work at the crossing nose occur during two-point contact stage rather than, as widely believed, at the moment of maximum normal contact force. These findings could contribute to the analysis of nonproportional loading in the materials and lead to a deeper understanding of the damage mechanisms. The model provides a tool for both damage analysis and structure optimization of crossings.

No abstract

No abstract

In this paper, we investigate the capability of an axle box acceleration (ABA) system to evaluate the degradation at railway crossings. For this purpose, information from multiple sensors, namely, ABA signals, 3D rail profiles, Global Positioning System (GPS) and tachometer recordings, was collected from both nominal and degraded crossings. By proper correlation of the gathered data, an algorithm was proposed to distinguish the characteristic ABA related to the degradation and then to evaluate the health condition of crossings. The algorithm was then demonstrated on a crossing with an unknown degradation status, and its capability was verified via a 3D profile measurement. The results indicate that the ABA system is effective at monitoring two types of degradations. The first type is uneven deformation between the wing rail and crossing nose, corresponding to characteristic ABA frequencies of 230–350 and 460–650 Hz. The second type is local irregularity in the longitudinal slope of the crossing nose, corresponding to characteristic ABA frequencies of 460–650 Hz. The types and severity of the degradation can be evaluated by the spatial distribution and energy concentration of the characteristic frequencies of the ABA signals.

In this paper, we present a method for evaluating the performance of railway crossing rails after long-term service. The method includes 1) 3D profile and hardness measurements; 2) finite element simulation of wheel/ rail interaction; and 3) numerical prediction of rail degradation. We conducted a case study on a crossing that had been in service for several years. The results indicate that the crossing experienced a run-in process in the major traffic direction, manifested as a widening of the running band, an enlargement of the contact patch size, a decrease in contact stress and eventually a reduction in plastic deformation and wear. However, the wheel/rail interaction was exacerbated in the minor traffic direction which induced more severe plastic deformation and wear.

This paper describes an approach for characterizing the dynamic behavior of the vehicle/track interaction at railway crossings. In the approach, we integrate in situ axle box acceleration (ABA) measurements with roving-accelerometer hammer tests to evaluate the influence of train speed, train moving direction (facing and trailing directions), sensor position (leading and rear wheels of a bogie), and the natural response of track structure on ABA signals. The analysis of data from multiple sensors contributes to the following findings: the major frequency bands of the vertical ABA are related to the natural frequencies of the crossing; thus, these ABA frequency bands are not greatly affected by variations in train speed, moving direction, and sensor position. The vibration energy concentrated at the major ABA frequency bands increases at higher train speeds, along the facing moving direction and from the leading wheel. The crossing rails vibrate as a combination of bending and torsion rather than solely bending at the major ABA frequency bands, since the vibrations of the wing rails are not synchronized. These results help enhance our understanding of the vehicle/track interaction at crossings and can be used to improve the dynamic response-based system for monitoring the condition of crossings.

No abstract

To improve the quality of track maintenance work, it is a desire to estimate vehicle dynamic behavior from track geometry irregularities. This paper proposes a deep learning model to predict vehicle responses (e.g., vertical wheel-rail forces, wheel unloading rate, and car body vertical acceleration) using deep learning techniques. In the proposed CA-CNN-MUSE model, convolutional neural networks (CNNs) are used to learn features of track irregularities, and multiscale self-attention mechanisms (MUSE) are employed to capture the long-term and short-term trends of sequences. Coordinate attention (CA) is introduced into CNN to focus on important interchannel relationships and important spatial mileage points. The experiments were performed on a multibody simulation model of the vehicle system and the measured data of the actual high-speed line. The results show that the CA-CNN-MUSE has high prediction accuracy for vertical vehicle responses and fast computation speed. The predicted time-domain waveforms and power spectral densities (PSDs) agree well with the actual vehicle responses. The main features of the lateral vehicle responses can also be captured by the proposed method, yet the results are not as good as the vertical ones.