基于响应面法的数控机床床鞍尺寸优化

摘要

The top bed saddler of the NC lathe is taken as the object of study. Modal analysis is carried out after the finite element model is built , To get a set of experiment points ,Latin Hypercube experimental de-sign is carried out based on the design variables of the saddle. A RSM of the first order natural frequency and mass to design variables is fitted by neural network based on these points ,then a math model is built . Finally the model is resolved by Genetic Algorithm, and the result is got by size optimization. The optimiza-tion result of this paper makes a guide for the redesign of the bed saddler of NC lathe. The analysis process and methods provide a reference for the designing and manufacturing of similar bed saddler of machine tools.%以数控车床上床鞍为研究对象，构建床鞍的有限元模型后对其进行模态分析。以床鞍的结构设计变量为基础进行拉丁超立方实验设计，得到一组试验点。将这组试验点作为训练数据用神经网络方法拟合出床鞍质量和一阶固有频率相对于设计变量的响应面模型，在此基础上完成数学模型的构建。最后通过遗传算法求解该模型，得到床鞍尺寸优化后的结果。文章的优化结果为改进床鞍的结构提供了理论依据，文中的分析思路和方法为其他机床床鞍的设计和生产提供了参考。