Structure Optimization Design Of CRH380 EMU

The high-speed EMUs have grown rapidly in a short period of time and become a symbol and business card of China’s image since they introduced to our country at the end of the last century.The trains play a vital role in railway passenger transport and even in the entire transportation industry.They have become an important content of the research on the lightweight of car body in order to actively respond to the energy conservation and emission reduction,and improve the efficiency of energy utilization.This paper analyzes the vehicle characteristics and structure of the current CRH380B-type EMU,and simulates and simplifies the structure,and uses the Hyper Mesh,finite element pre-processing tools,to establish the finite element model.The main research goal is the body underframe,and the total mass is the objective function.This paper selects the Response Surface Model(RSM)as the main strategy to complete the approximate model among the various methods of establishing the approximate model.There are two ways of using the built-in basic algorithm of the Isight platform and the asynchronous optimization particle swarm optimization algorithm discussed at the second chapter.Then evaluate the results of the two algorithms with three evaluation indicators and select the optimal method and formulate the structure lightweight strategy of the car body.The main contents of this paper are as follows:(1)The linear adjustment function is added to the particle swarm optimization algorithm,and discussed a new algorithm.There are prone to occur in the iterative process of particle swarm optimization with the "premature" phenomenon,slow convergence speed,poor robustness,and low precision.This paper proposes to construct a linear function and use the fitness value of the particles to dynamically adjust the learning factor,so that the algorithm is the convergence speed in the initial period of the search is accelerated,and reduced in the later period of the search to maximize the local optimization capability of the algorithm,reduce the time cost,and increase the accuracy and quality of the solution.(2)The whole vehicle structure of the car body is simulated and simplify,and generate the finite element model with the Hyper Mesh,finite element pre-processing tool.Simulate the plate structure of the car body with shell elements,re-simplify the whole vehicle solid model,connection mode,and accessories on and under the car,omit non-load-bearing parts,extract the middle surface model,select the corresponding unit type,and simplify the discretization of the car body,check the mesh density and size of the discretized car body,and generate a finite element model of the car body structure.Based on the finite element model and referring to relevant domestic and foreign standards such as EN12663:2000,14 kinds of combined working conditions are selected,and the stiffness,modal and strength are checked and analyzed,which proves the rationality of the finite element model.(3)Approximate models and formulate optimization strategies are established.Choose the Isight platform,create a sample space of design variables,and use the main effects analysis method to analyze the sensitivity of each design variable,the interaction effect diagram shows that there is no interaction between the design parameters;use the response surface model method to establish the corresponding approximation of the car body Model,by analyzing the accuracy and error of the approximate model,it is obtained that the error between the approximate model and the finite element model is within the specified range.(4)The lightweight design of the car body is based on the approximate model.Based on the approximate model established above,the boundary conditions are the vertical bending stiffness,the first-order vertical bending frequency,static strength,and torsional stiffness in the case of an empty vehicle,and the weight of the empty vehicle excluding all the equipment on and off the vehicle as the boundary conditions.The objective function and the thickness of each section of the profile are design variables.The original particle swarm algorithm based on the Isight platform and the asynchronous optimization particle swarm algorithm discussed in Chapter 2 are used to discuss the two optimization methods.The results of the two methods are collected and compared.With the satisfaction of meeting the normal operation of the train,the results of the two algorithms can meet the requirements,but the latter requires less time and fewer iterations.Based on this method,an optimization plan is formulated.