Research On Multiobjective Optimization Design Of Wheel Profile

With the rapid development of high speed and heavy haul train in our country, the problems of match of wheel and rail profile were more and more important. The wheel profile optimization is one important mean to improve match of wheel profile and rail profile. This paper implements the wheel profile optimization research in aspects including choosing design variables, constructing multiobjective optimization model, choosing solution algorithm of multiobjective problems and engineering application.The main researches are as following.Applying the multiobjective optimization theory, the objective functions, design variables and constrain condition of the wheel profile optimization design model were investigated. On the basis of elaborate analysis of the geometrical character of typical wheel profile at home and abroad, this paper introduces radius and center coordinate of arc of arc wheel profile as the design variables, the new design variables have the advantage of less amount of calculation, high accuracy, clear physical significance and the profile design by the new design variables satisfies the monotonicity and convexity-concavity in railway engineering.The convergence and update of sub region of response surface method is elaborate studied which is a global, nonconvex and nondifferentiable optimization method and the Gauss Radial Basic Function based response surface based multiobjective optimization method is introduced to solve the wheel profile multiobjective problem. Comparing with the conventional optimization method, genetic method and sequential quadratic programming method, the new optimization method has the advantages of less amount of calculation and rapid convergence. Taking the wheel profile optimization of ERRI 200 km·h-1 standard passenger vehicle as example, the comparison application analysis of polynomial based response surface method and Gauss radial basic function based response surface method is implemented. The optimization result shows that Gauss radial basic function based response surface wheel profile optimization method is superior to polynomial based response surface wheel profile optimization method. The optimization software template of gauss radial basic function response surface based multiobjective optimization method applying to wheel profile optimization is programmed in C++ language. The optimization software template is able to call the multibody dynamics software ADAMS/Rail to analyze dynamic simulation of railway vehicle system in C++ environment and feed the useful data from dynamic simulation analysis back to the wheel tread optimizing software, to complete the cycle process of the whole optimization design. The whole optimization cycle is implemented in computer without human intervention. In order to validate the program in this paper, taking the dynamic model of ERRI 200 km/h passenger car as example, the validation investigation of multiobjective optimization program is implemented by optimization the LMA profile.The simulation model of CRH₂ high speed multiple units are constructed using dynamic simulation software ADAMS/Rail of railway vehicle system dynamics, and the model is validated by impapplying the field data in Jing-jin line. According to the characteristic of high speed multiple units, choosing three functions of the critical speed, wheel/rail wear and lateral comfort as the objective functions of wheel profile multiobjective optimization model. After wheel profile optimization, the critical speed of CRH₂ high speed multiple units increases 22.6% and the wheel/rail wear index of leading wheelset reduced 7.1%.The simulation model of HXD₂ electric locomotive are constructed using dynamic simulation software ADAMS/Rail of railway vehicle system dynamics, and the model is validated by impapplying the field data in Da-Qin line. According to the characteristic of electric locomotive, choosing three functions of the wheel/rail wear, max contact press, and lateral wheel axis force as the objective functions of wheel profile multiobjective optimization model. After wheel profile optimization, the wheel/rail wear index of leading wheelset of HXD₂ electric locomotive reduced 7.1229%, the max contact press of outer side wheel/rail on curve of 800 m radius reduced 28.7%, the max contact press of outer side wheel/rail on curve of 1 000 m radius reduced 31.4%, the lateral axis force of leading wheelset on curve of 800 m radius reduced 22.6% and the lateral axis force of leading wheelset on curve of 1 000 m radius reduced 18.6%. Comparion and analysis of wheel/rail contact press of HXD₂ electric locomotive before and after optimization using the finite element software ABQUS show that after optimization the maximum Mises profile reduced 23.3% and maximum contact press reduced 29.1%.Wheel profile optimization design is one multiobjective engineering problem relating to many track vehicle system dynamics, wheel/rail materials match and geometrical match and the important research project. It need the new theory and method of computer science to combine with the field test and laboratory test to comprehensive system research and field test validation. The work in this paper is a begin of long term research work, much work is needed to be research in future.