改进神经网络的汽车车身结构动力学建模分析  

Dynamic Modeling and Analysis of Automobile Body Structure Based on Improved Neural Network

作　　者：包亚杰 张伟社[1] BAO Ya-jie;ZHANG Wei-she(Chang'an University,School of Contruction Machinery,Xi'an Shaanxi 710000,China)

机构地区：[1]长安大学工程机械学院,陕西西安710000

出　　处：《计算机仿真》2022年第11期189-192,471,共5页Computer Simulation

摘　　要：车身结构承载了行驶过程中的大部分载荷,为提高车身动态特性,设计一个基于神经网络的汽车车身结构动力学模型。对车身结构进行详细设计,主要包含前围板组合件、左右侧围、两门汽车侧围等,采用神经网络方法构建动力学方程,并采用多学科优化方法对汽车车身结构进行优化,完成基于神经网络的汽车车身结构动力学模型的构建。实验结果表明,经所研究模型优化后的车身结构,不仅较车身优化前B柱加速度峰值低,还提高了车身扭转角度、降低了车身左侧、右侧纵向位置变形量。The body structure bears most of the loads of automobiles in the course of driving.In order to improve the dynamic characteristics of automobile body,this paper designed a kinetics model of automobile body structure based on neural network.Firstly,the automobile body structure was designed in detail,mainly including the cowl panel assemblies,left and right-side panels,side panels of two-door vehicle.Secondly,neural network method was used to construct dynamic equations,and then multi-disciplinary method was adopted to optimize the automobile body structure,and thus to complete the construction of model.Experimental results show that the automobile body structure optimized by the designed model not only has lower peak acceleration of the B-pillar,but also improves the body torsion angle and reduces the longitudinal deformation on the left and right sides of automobile body.

关 键 词：神经网络 车身结构 动力学 速度 多学科 灵敏度 

分 类 号：TH122[机械工程—机械设计及理论]