Simulation of wheel and rail profile wear: a review of numerical models

Abstract

The development of numerical models able to compute the wheel and rail profile wear is essential to improve the scheduling of maintenance operations required to restore the original profile shapes. This work surveys the main numerical models in the literature for the evaluation of the uniform wear of wheel and rail profiles. The standard structure of these tools includes a multibody simulation of the wheel–track coupled dynamics and a wear module implementing an experimental wear law. Therefore, the models are classified according to the strategy adopted for the worn profile update, ranging from models performing a single computation to models based on an online communication between the dynamic and wear modules. Nevertheless, the most common strategy nowadays relies on an iteration of dynamic simulations in which the profiles are left unchanged, with co-simulation techniques often adopted to increase the computational performances. Work is still needed to improve the accuracy of the current models. New experimental campaigns should be carried out to obtain refined wear coefficients and models, while strategies for the evaluation of both longitudinal and transversal wear, also considering the effects of tread braking, should be implemented to obtain accurate damage models.

A Method for Damage Detecting of Large Reflector Antennas Wheel-Rail Based on Electromagnetic Ultrasonic Technology

“…Drawing on the previous research ideas, this paper adopts rectangular groove equivalent to the actual wheel-rail crack damage in the simulation analysis [28][29][30]; corrosion usually corrodes downward from the middle point and then expands to both sides, so the simulation adopts inverted triangular groove equivalent to the actual wheel-rail corrosion damage [31]; the wear is usually a concave arc, and elliptical groove equivalent to the actual wheel-rail wear damage is adopted in the simulation [32]. By analyzing and comparing the model results without defects and the simulation results of several defects, the existence of defects and the location of defects can be judged more clearly and intuitively.…”

Study on the Influence of the Modelling Strategy in the Calculation of the Worn Profile of Railway Wheels

“…The global computation of wear is performed using the Simpack wear module, while the local computation is carried out by means of a dedicated Matlab routine, which reads the contact patch data obtained from the dynamic simulation. The Simpack wear module computes the worn volume ΔW in each contact point position for each time step according to eqn (1), in which N is the normal load, H is the hardness of the softer material, s is the total sliding distance and kArch is the wear coefficient, which is a function of sliding speed and contact pressure. The wear coefficient is obtained from the KTH wear map as a function of the contact pressure and of the sliding speed, with a reduction of the nominal values by a factor 7 to account for natural contamination at the wheel-rail contact, as suggested by Jendel [28] and Jendel and Berg [29].…”

“…Currently, the typical strategy adopted to predict wear of wheel profiles is based on numerical tools that include a module for the vehicle-track dynamic simulation, usually implemented in a commercial multibody (MB) software package, and a module that computes the worn profile starting from the outputs of the dynamic simulation [1]. In the past, the wear module was implemented in the MB code used for the dynamic simulation in the form of a user-written routine, while nowadays several wear computation tools rely on co-simulation approaches, with the wear module implemented in a different computational environment [2].…”

Application of the extended FASTSIM for non-Hertzian contacts towards the prediction of wear and rolling contact fatigue of wheel/rail systems

“…This approach, however, is makes use of lookup tables which only provide the total creep forces, hence not allowing to determine the distribution of local stresses and frictional power in the contact region which are generally required for wear and RCF estimations. 15 To overcome this problem, in this paper, we apply an extended version of the FASTSIM algorithm for a SDEC region which we recently proposed, 16 extending its use to a generic non-elliptic contact area and we assess the accuracy of this algorithm in the prediction of wheel wear and RCF using CONTACT as a term of comparison. The numerical results show that the extended FASTSIM provides better agreement to the results of CONTACT than other commonly used existing models for detailed analysis of wheel wear and RCF.…”

“…This approach, however, is makes use of lookup tables which only provide the total creep forces, hence not allowing to determine the distribution of local stresses and frictional power in the contact region which are generally required for wear and RCF estimations. 15…”