Improvement Of Particle Swarm Optimization Algorithm And Its Application In Structure Optimization Of CRH3 EMU

As the fastest land transportation vehicle,the high speed EMU plays an important role in the railway passenger transport in China.In accordance with the requirements of "introducing advanced technology,combining design and production,creating Chinese brand",the domestic work of high speed EMU is being carried out smoothly.Vehicle lightweight optimization is an important content of high speed EMU Structural design.Taking the chassis of the car body as the main research object,the lightweight optimization model of the car body is established by using Hypermesh and ANSYS software.An improved particle swarm optimization algorithm based on dynamic inertia weight is proposed.The algorithm structure is made by using MATLAB software,and the test examples of unconstrained optimization and constrained optimization are completed.The improved algorithm is applied to the lightweight optimization design of the CRH3 EMU structure,and the light weight optimization scheme of the car body is obtained.The research contents of this paper are as follows:1.In order to improve the slow convergence rate and easy to fall into premature,the success rate of particle swarm optimization algorithm is constructed.The success rate of population iteration is used as feedback parameter to dynamically change the value of inertia weight,so as to adjust the global and local search ability of PSO,and improve the convergence performance of the algorithm.4 examples of unconstrained optimization and 4 constraint optimization examples are applied to complete the algorithm effectiveness test.The results show that the improved algorithm improves the accuracy of solving multimodal functions significantly,and has strong global optimization ability when solving constrained optimization problems.2.First,the structure of the CRH3 car TC02 car body is fully understood through the drawings and model data,and the geometric model of the car body is simplified.The shell element and beam element are used to express the stress characteristics of the car body.The components of the vehicle are simplified to their actual installation locations by using the quality unit,and the finite element model of the car body is established.Referring to the European standard EN 12663:2000,the static strength,stiffness and modal finite element calculation of vehicle body under 14 working conditions are completed by using the finite element model of the CRH3 car TC02 car.The calculation results show that the structure of the car body meets the standard requirements.3.The design variables are the thickness of the upper and lower cover plate,the thickness of the underframe plate and the thickness of the side beam of the bottom frame.With the output file of ANSYS,the parameters corresponding to the static strength,stiffness and mode of the constrained car body are extracted as the output response.In the ISIGHT optimization platform,the acquisition of sample points is completed by using the optimal Latin hypercube test design method.Based on neural network(RBF/EBF),the approximate model of CRH3 type TC02 car body structure is established,and the approximate model error is calculated.Based on the coefficient table,the multivariate two regression equation of output response on design variables is obtained.4.Based on the approximate model,taking the quality of TC02 car body as the objective function and taking the strength,stiffness and the first vertical resonance frequency as the constraint conditions,the lightweight optimization design scheme of the car body with the smallest car body mass is obtained.The new particle swarm optimization(PSO)and the particle swarm optimization(PSO)in Isight are applied to the lightweight solution of the car body of the CRH3 car TC02 vehicle,and the two optimization results are compared.The results show that the two algorithms can achieve the purpose of lightening the vehicle body under the premise of ensuring the performance of the vehicle,but the improved algorithm has certain advantages in speed and optimization ability,and based on the optimization results of the improved algorithm,it puts forward the optimization plan.