在侧面碰撞中电动汽车电池模块破损的预测  

Prediction of battery-module damage in electric-vehicle side-collisions

作　　者：王居闯 曹清林 邱睿 宋刘伟 郭平安 赵港 WANG Juchuang;CAO Qinglin;QIU Rui;SONG Liuwei;GUO Ping'an;ZHAO Gang(School of Mechanical Engineering,Jiangsu University of Technology,Changzhou 213001,China;School of Mathematics and Statistics,Anhui Normal University,Wuhu 241000,China;China Machinery Industy Technology Research Institute of Precision Forming,Wuhu 241000,China)

机构地区：[1]江苏理工学院机械工程学院,常州213001 [2]安徽师范大学数学与统计学院,芜湖241000 [3]中机精密成形产业技术研究院(安徽)股份有限公司,芜湖241000

出　　处：《汽车安全与节能学报》2024年第2期169-177,共9页Journal of Automotive Safety and Energy

摘　　要：为增强电动汽车在侧面碰撞事故中的电池安全性,以某新能源汽车电池箱为研究对象,创建侧面碰撞情况下的有限元模型。通过LS-DYNA进行5种速度侧面碰撞仿真,提取电池箱侧壁几何中心点的应力曲线以及电池模块破损情况,根据两者之间的相关关系,建立预测电池模块碰撞破损的反向传播(BP)神经网络模型。模型的输入量为应力曲线,输出向量为模块破损情况。结果表明:5种速度碰撞后预测错误3块,其余177块预测均正确;准确率达到98.33%。因而,通过对算法的设计可预测出电动汽车在受到侧面碰撞时将要破损的具体模块,有利于提高电动汽车安全性。A finite element model was developed to simulate side-collision scenarios on a new type of energyvehicle(EV)battery-pack to enhance the battery safety of EVs in the side-collision accidents.Using LS-DYNA,five different collision simulations were performed at various speeds.The stress curves at the geometric center of the battery pack's side wall and the battery module damage conditions were extracted.A predictive neural network model from the back propagation(BP)was established for battery module collision damage based on the correlation between the stress curves and the battery-module damage-conditions factors.The model's input quantity was the stress curves,and the output vector was the module damage conditions.The results show that three blocks at five different speeds are predicted incorrectly after collisions,while the remaining 177 blocks are predicted correctly with an accuracy rate of 98.33%.Therefore,this algorithm's design enables the identification of specific modules prone to damage in electric vehicles during side collisions,which holds significant implications for enhancing overall electric vehicle safety.

关 键 词：电动汽车(EV) 电池模块 侧面碰撞 破损预测 反向传播(BP)神经网络 有限元(FE)仿真 

分 类 号：U467.14[机械工程—车辆工程]