| With the rapid development of high-speed railways and high-speed trains in China,the requirements for the dynamic performance of trains have also increased.The current optimization of train dynamics performance has been transformed into a complex non-linear multi-objective optimization problem.The complex physical model established by traditional train dynamics and the traditional single-objective optimization cannot meet the current multi-objective optimization requirements.Therefore,an agent model related to the train suspension design parameters and dynamic performance indicators is established in this paper.The train performance evaluation method is used as a constraint.Based on the agent model,the train suspension parameters are pareto multiplied by the NSGA-II genetic algorithm.Objective optimization,and then solve the current complex nonlinear multi-objective optimization problem.First,taking a certain type of EMU as the research object,the dynamic model of the train is established based on the SIMPACK dynamic analysis software,and the model is verified.Add track excitation and line conditions to the vehicle dynamics model,collect simulation data and calculate vehicle dynamic performance indicators.The dynamic model of the train was used to change the stiffness and damping of the first series of horizontal shock absorbers,the first series of vertical shock absorbers,and the second series of horizontal shock absorbers.Influence,further determine the research scope of design parameters,and provide sample data for the subsequent establishment of the proxy model.Secondly,the optimal Latin hypercube test design method is used to design the test,and a radial basis(RBF)neural network is used to construct a proxy model of the mathematical relationship between the suspension parameters and the dynamic performance indicators of the high-speed train.An error analysis was performed on the agent model through various methods,and it was confirmed that the agent model has sufficient accuracy for subsequent optimization.Finally,Isight optimization software is used to solve the multi-objective optimization problem,and the nonlinear critical speed and derailment of the train The coefficient and lateral stability index were used as the optimization targets.The NSGA-Ⅱ genetic algorithm was used to optimize the suspension parameters of the train for multiple objectives.The optimized train design parameters are input into the dynamics analysis software and compared with the original vehicle data.A comparative analysis of the optimization rates of train dynamic performance indicators in Scheme 1,Scheme 2 and Scheme 3 shows that the optimized vehicle dynamic performance has been greatly improved.Among them,the optimization effect of scheme 2 on the dynamic performance index of the train is higher than that of scheme 1 and scheme 3.The optimization rate of the lateral stability index is the highest,reaching 10.228%.The optimization effect has a better improvement on the train’s smooth running and the stability ofthe serpentine movement.In addition,designers can also choose different optimization schemes according to actual needs,adjust the optimization weight ratio of each goal,and finally be able to choose among many optimization schemes to meet the actual needs. |
