基于PSO-SVR的汽车前纵梁优化设计  

Optimization Design for Vehicle Front Rail Based on PSO-SVR Technology

作　　者：殷为洋 仲衍慧[1] 郭树文[1] 李向荣[1] YIN Weiyang ZHONG Yanhui GUO Shuwen LI Xiangrong(China Automotive Technology ＆ Research Center,Tianjin 300162,China)

机构地区：[1]中国汽车技术研究中心,天津300162

出　　处：《天津科技》2016年第11期56-62,共7页Tianjin Science & Technology

摘　　要：在汽车碰撞中,前纵梁的吸能特性对整车安全具有至关重要的影响。综述了吸能梁的横截面形状、整体结构形状、诱导槽、焊接方式及壁厚对吸能特性的影响。由于实际生产工艺及与其他部件的装配要求,前纵梁的横截面形状和结构形态通常不做大的改动,因此在给定材料的基础上,板厚优化是前纵梁优化设计的主要内容。分别以焊接双吸能梁和某汽车前纵梁为例,设定板厚为自变量,采用拉丁方试验设计方法,建立基于粒子群优化的支持向量机回归(PSO-SVR)近似模型,结合NSGA-Ⅱ遗传算法进行多目标优化设计,最终匹配出各个部分的最优板厚。结果表明,经过优化设计的焊接双吸能筒和汽车前纵梁吸能特性有显著提高,证明该方法在吸能部件的优化设计中具有一定的工程应用价值。The energy absorption characteristics of front rails have a crucial impact on vehicle safty during the automobilecollision.A summary about how cross-sectional shape,structure,inducing grooves,welding technology and wall thicknessinfluence the energy absorption characteristics was proposed.Because of productive technology and requirements ofassembling with other components,the cross-sectional shape and structure form usually remain the same.Thickness optimizationis the main design content when the material for front rail has been chosen.In order to get the best value of wall thickness,the optimization design of welding beam and front rail are based on the methods below：Taking the thickness as variables,producing sample points based on the latin square,building metomodels with PSO-SVR（Particle Swarm Optimization,Support Vector Machine for Regression）method,and taking the NSGA-II（Nondominated Sorting Genetic AlgorithmⅡ）method into the multi-objective optimization design process.Results show that energy absorbing characteristics of thewelding beam and front rail have been improved with this optimization process.It also proves that the method has a certainengineering value for the optimization design of energy absorbing components.

关 键 词：前纵梁 PSO-SVR 近似模型 板厚 

分 类 号：U463[机械工程—车辆工程]