稀疏网格模型在变速箱壳体轻量化设计中的应用  

Application of Sparse Grid Model to Lightweight Design of Gearbox Shell

作　　者：刘爽 吉效科 张思[1] 华剑[1] LIU Shuang;JI Xiaoke;ZHANG Si;HUA Jian(School of Mechanical Engineering,Yangtze University,Jingzhou 434023,Hubei,China;General Machinery Manufacturing Factory,Changqing Oilfield Company,Xi′an 710018,China)

机构地区：[1]长江大学机械工程学院,湖北荆州434023 [2]长庆油田公司机械制造总厂,西安710018

出　　处：《机械科学与技术》2022年第10期1515-1523,共9页Mechanical Science and Technology for Aerospace Engineering

基　　金：国家自然科学基金项目(52174018)。

摘　　要：为解决某压裂泵用变速箱壳体壁厚设计不合理的问题,提出了一种稀疏网格近似模型与MOGA遗传算法集成的优化方法。对比分析传统近似模型与稀疏网格模型的预测精度,得出稀疏网格模型的预测精度更高。利用稀疏网格初始化法构建样本数据,搭建出响应面模型,采用MOGA遗传算法搜索最优设计方案。结果表明:在结构性能满足使用要求的基础上变速箱壳体重量减轻了18.1%,为变速箱壳体轻量化设计提供了一种新途径。In order to solve the problem of unreasonable wall thickness design of a gearbox shell for a fractured pump, an optimization method integrating sparse grid approximate model and MOGA genetic algorithm was proposed. The comparison and analysis between the prediction accuracy of the traditional approximate model and the sparse grid model show that the prediction accuracy of the sparse grid model is higher. The sample data is constructed by using the sparse grid initialization method, the response surface model is built, and the MOGA genetic algorithm is used to search for the optimal design scheme. The results show that the weight of the gearbox shell is reduced by 18.1% under the structural performance meeting the requirements of use, which provides a new way for the lightweight design of the gearbox shell.

关 键 词：近似模型 轻量化 稀疏网格 遗传算法 灵敏度分析 

分 类 号：TH164[机械工程—机械制造及自动化]