Research On Metal Re-Profiling Decision Based On Multi-Objective Optimization Algorithm

With the high-speed development of Chinese high-speed rail ’Belt and Road Initiative’,the safety maintenance management of domestic trains is critical Since the trains wear the wheel pair in operation to cause that the wheel is deformable to the outer profile and wheel deraille at last,it is necessary to adopt a wheel re-profiling method,which is no unload re-profiling The fixed tuning template will not only waste the wheel to make the wheel scrapped in advance,but also it is no universality,so the meticulous re-profiling strategy is important for increasing the wheel on cutting economics and extending service life.In this paper,the CRH380D train is main research object Through the analysis of the historical measured data of the wheel,the wear model and multi-objective model are established by size data of wheel.The Monte Carlo method is used to simulating fine re-profiling strategies that are prolonging the wheel life,reducing the number of revolutions,and reducing cutting costs Compared to fixed re-profiling strategy,the nine strategies selected by traditional evaluation methods have increased by at least 4.4177%over the service life However,the traditional evaluation method has a singleness lacking comprehensive and objective,in this paper,a combination weighted TOPSIS method is introduced.This way is an angle combined with subjective weight and objective weight based on multi-index evaluation strategy portfolio.The results displays that compared with the traditional evaluation method,stability of combined weight TOPSIS has increased by 31.6%,and it is more suitable for the round to repair multitarget decision making.For rounds to multi-objective optimization,the traditional Monte Carlo method is not substantially solved with multi-objective optimization problems and the search target space is limited.The optimization algorithm NSGA-II for multi-objective optimization issues is applied to the wheel re-profilinf problem in this paper,the results indicate that NSGA-Ⅱ looking for optimal wheel re-profiling strategies is batter than Monte Carlo However,the convergence and distribution of the NSGA-II algorithm are poor when it is solving the three-dimensional multiobjective problem.In order to improve the original algorithm,the adaptive mixed cross operator and the adaptive mutation operator areimproved at first,following,an external archive method is added to NSGA-Ⅱ to ensure excellent individual retention,and BFE calculation methods is added to the Elite Policy of NSGA-Ⅱ finally.The improved NSGA-Ⅱ algorithm comparedwith other algorithms in many test problems with comprehensive performance IGD,the experimental results show that the improved NSGA-Ⅱ algorithm has its batter optimistic ability,distribution,and convergence than NSGA-Ⅱ.In the three-dimensional target problems of wheel re-profiling,nine lathing strategies of the improved NSGA-Ⅱ is better than all of NSGA-Ⅱ with evaluation of strategies improved 94%,and the optimal re-profiling strategy indicates that during the removal,the standard curve of the LM34.1 re-profiling profile can not only increase the wheel life by 16.15%,but also can improve the economics of the wheel pair cutting.