基于集成学习的某重型自卸车车厢多目标优化  

Multi-objective optimization of a heavy dump truck carriage based on ensemble learning

作　　者：兰克俊 蒋春玲 廖警 黄伟[1] LAN Kejun;JIANG Chunling;LIAO Jing;HUANG Wei(School of Mechanical Engineering,Guangxi University,Nanning 530000,Guangxi,China;SINOTRUK Liuzhou Yunli Special Vehicle Co.,Ltd.,Liuzhou 545000,Guangxi,China)

机构地区：[1]广西大学机械工程学院,广西南宁530000 [2]中国重汽集团柳州运力专用汽车有限公司,广西柳州545000

出　　处：《中国工程机械学报》2024年第1期44-49,共6页Chinese Journal of Construction Machinery

基　　金：广西科技重大专项资助项目(桂科AA22068055,桂科AA22068060)。

摘　　要：针对复杂自卸车车厢结构的轻量化设计问题,提出一种集成学习联合多目标优化的设计方法。以某重型自卸车车厢为研究对象,建立了高保真的车厢有限元模型,并进行3个典型静力学工况和自由模态的结构性能仿真分析。通过贡献度分析筛选出重要零部件的厚度为设计变量,以危险工况下的最大位移、最大等效应力及自由模态第1阶固有频率为约束条件,以质量和满载举升0°工况下的最大位移为优化目标,运用极端梯度提升(XGBoost)集成学习近似模型进行自卸车车厢多目标优化设计。经优化决策后的轻量化设计表明:在满足基本静动态结构性能要求下,车厢满载举升0°工况下的最大位移减少了1.98 mm,减少幅度为8.57%,车厢质量减少了0.207 t,减少幅度为6.64%,取得了较好的轻量化效果。Aiming at the lightweight design problem of complex dump truck,an ensemble learning combined multi-objective optimization design method was proposed.Taking a heavy dump truck as the research object,the high-fidelity finite element model of the carriage was established,and the structural performance of three typical static conditions and free modes were simulated.Through contribution analysis,the thickness of important parts was selected as the design variable,the maximum displacement,maximum equivalent stress under dangerous working conditions and the first natural frequency of free mode were taken as the constraint conditions,and the mass and maximum displacement under full load lifting condition of 0°were taken as the optimization objectives.The XGBoost ensemble learning approximate model was used to carry out the multi-objective optimization design of dump truck carriage.The lightweight design after optimization decision shows that,under the basic static and dynamic structural performance requirements,the maximum displacement of the carriage under full lift condition of 0°is reduced by 1.98 mm(8.57%),and the carriage mass is reduced by 0.207 t(6.64%),achieving a good lightweight effect.

关 键 词：自卸车车厢 轻量化设计 集成学习 多目标优化 

分 类 号：U469.4[机械工程—车辆工程]