基于神经网络算法的汽车前纵梁碰撞性能优化设计  

Optimization Design of Crash Performance for Vehicle Front Longitudinal Beam Based on Neural Network Algorithm

作　　者：王甲畏 WANG Jia-wei(Altair Engineering Software(Shanghai)Co.,Ltd.,Shanghai 200070,China)

机构地区：[1]澳汰尔工程软件(上海)有限公司,上海200070

出　　处：《价值工程》2024年第5期97-99,共3页Value Engineering

摘　　要：随着汽车碰撞安全法规的升级,对车辆的安全性能要求越来越高,这就需要对不同的法规工况采用合适的碰撞吸能策略。汽车前纵梁结构,作为整车结构中的核心吸能结构件,吸收能量的多少关系到车辆的安全性能指标以及NCAP星级评价。前纵梁在碰撞过程中的变形模式是在产品设计和碰撞分析中比较难以控制的,本文根据作者多年的实践经验,采用神经网络方法,对可能的变形模式进行神经网络聚类分析,然后对预采用的变形模式再进行结构优化。这样最后设计的结构不仅仅达到了预期性能目标,还满足了可控的变形模式预期。对汽车碰撞性能设计提供了一个完整的方法和流程,很有实际工程借鉴价值。With the upgrading of vehicle crash and safety regulations,the requirements for vehicle safety performance are becoming increasingly high,which requires the adoption of appropriate crash energy absorption strategies for different regulatory conditions.The front longitudinal beam structure of a vehical,as an energy absorbing core component in the vehicle structure,the amount of energy absorbed is related to the safety performance indicators and NCAP star rating evaluation of the vehicle.The deformation mode of the front longitudinal beam during collision is difficult to control in product design and crash analysis.Based on the author's years of practical experience,this article uses neural network methods to cluster and train possible deformation modes,and then optimizes the pre adopted deformation modes.The final designed proposal not only met the performance target,but also achieved the expected and controllable deformation mode.This provides a complete method and process for vehicle crash performance development,which has practical engineering reference value.

关 键 词：优化设计 神经网络 前纵梁 碰撞 

分 类 号：U463.326[机械工程—车辆工程]