基于RBF近似模型的汽车前部部件耐撞性优化  

Simulation and Optimization of Crashworthiness of Vehicle Front Parts Based on RBF Neural Network

作　　者：辛丕海 汪伟康 王宇[1] 

机构地区：[1]合肥工业大学机械与汽车工程学院,安徽省合肥市230009

出　　处：《农业装备与车辆工程》2016年第4期51-55,共5页Agricultural Equipment & Vehicle Engineering

摘　　要：为提高车身的安全性,采用Isight与Ls-dyna集成优化平台对汽车前部吸能部件进行研究,以保险杠横梁前后板、吸能盒内外板、前纵梁内外板及其加强板7个板的厚度作为设计变量,运用优化拉丁超立方设计方法进行试验设计,建立起了径向基函数神经网络模型,运用遗传算法对其进行优化,最终达到提高车身前部耐撞性的目的。To improve the crashworthiness of the automotive body,this paper chooses 7 variables,which are,the thickness of the bumper beam＇s front and rear plates,the crash-box＇s inner and outer plates,the front longeron＇s inner and outer plates,and the reinforcing plate,as design variables,uses optimal Latin hypercube design method to design the experiments,builds RBF model,and uses genetic algorithm to optimize the problem in the integrated optimization platform of Isight and LS-DYNA.The ultimate objective is to improve the crashworthiness of the front body.

关 键 词：汽车前部 吸能结构 耐撞性 RBF模型 遗传算法 

分 类 号：U462.3[机械工程—车辆工程]