Wheel-rail contact problem is a complex nonlinear problem,and there are many factors involved in rail grinding profile optimization,and the evaluation indexes of optimization are also different.At present,the optimization design of rail grinding profile is mainly aimed at the standard grinding profile,achieving single or less objective optimization under specific working conditions,and the research object lacks universal adaptability.The optimization design of rail grinding profile needs to be combined with the actual line characteristics and vehicle running conditions.It should be a complete system process from profile measurement,profile data processing,optimization design and performance test of grinding profile.Currently,there are few perfect optimization design schemes.Aiming at these problems,this paper systematically studies the multi-objective optimization design method of rail grinding profile.The following research contents have been completed:(1)A pre-processing method for measuring rail profile curve data is proposed,that is,the measured sample profile curve is smoothed by local weighted regression method,and then the sample profile is fitted into a sample representative profile that can comprehensively characterize the characteristics of railway diseases by using arithmetic average method.The representative profile is evaluated from four indicators:normal value,wear area,vertical wear value and side wear value,and finally a cubic NURBS curve parametric model of the representative profile is established.(2)The geometric relationship of wheel-rail contact is analyzed by the trace method,and the geometric parameters of wheel-rail contact,such as contact point position distribution,rolling circle radius difference and contact angle difference,are calculated.Based on the three-dimensional non-Hertz rolling contact theory and the finite element analysis method,a three-dimensional simulation model of the contact between standard LMA wheels and 60 kg/m rails is established.The static analysis of wheel-rail contact is carried out in ANSYS,and the mechanical performance parameters such as equivalent stress,contact spot and contact pressure of wheel-rail contact are calculated.A vehicle-track coupling dynamics simulation model is established in SIMPACK,and the performance parameters such as normal stress of wheel-track contact,radius difference of rolling circle and wear work are calculated.(3)Using Isight software for experimental design,the RBF proxy model of wheel-rail contact geometry and wheel-rail contact mechanics is established.Combining with NSGA-II algorithm,the rail parametric modeling process,wheel-rail contact matching performance analysis process and wheel-rail contact performance multi-objective optimization process are organically combined,and a complete rail grinding target profile optimization system process is established,which solves the problems of large calculation amount and low calculation efficiency in traditional calculation of wheel-rail contact geometry and contact mechanics.Using personalized profile grinding technology and asymmetric grinding profile design principle,the grinding target profiles of straight line section and small radius curve section are optimized.(4)Taking the adjustable NURBS weight factorωas the design variable,the average value of wheel-rail cumulative contact stressσms,the average maximum contact stress valueσmm ax and the wear workWt as the optimization indexes of the representative profile of the straight line section,and the radius differenceΔr of the rolling circle,the normal contact stress valueσm of the wheel-rail and the wear work Wt as the optimization indexes of the representative profile of the curve section,the multi-objective performance optimization of the representative profile of the straight line section and the curve section with small radius is carried out to obtain the optimal grinding target profile,The optimal Pareto solution and the corresponding weight factor solution set of wheel-rail contact performance index are obtained.The accuracy of the optimization results and the optimization effect of rail grinding target profile are preliminarily verified by actual simulation. |