Multi-objective optimization of a metro wheel profile to reduce wheel flange wear and RCF considering the distribution density of wheel−rail contact points

Abstract In this study, the problem of severe wheel flange wear and tread rolling contact fatigue (RCF) of metro wheels with two wheel profiles, i.e., the LM-30 profile and the DIN5573-30 profile, is investigated. The contact characteristics of the two profiles are compared to analyze the causes of abnormal wheel flange wear and tread RCF. Based on the two-dimensional Gaussian kernel density estimation method, indices combining the density distribution of contact points are proposed to evaluate the wheel flange wear and tread RCF. The optimization region of the wheel profile is defined by eight arc parameters, which is derived from the tangency of the arcs. A penalty function is used to ensure tangency between the arcs during the optimization process. A back-propagation neural network is constructed as a surrogate model for the optimization parameters and the objective function. Four intelligent optimization algorithms are compared and used to solve the optimization problem. The vehicle‒track dynamic model and the wheel wear prediction model are used to verify the performance of the optimized profile. The results show that the critical speed of the optimized profile is between that of the LM profile and that of the DIN5573 profile, and the curving performance of the optimized profile is significantly improved. At the same time, the optimized wheel profile can reduce wheel flange wear and tread RCF. The loss of wheel flange thickness is reduced by 41.9% compared with that of the DIN5573 profile, and the peak value of wheel tread RCF damage is reduced by 33.6% compared with that of the LM profile.