T型管液压成形加载路径自适应多目标优化

发布时间：2018-06-15 02:41

本文选题：T型管液压胀形 + 加载路径　；参考：《哈尔滨工业大学学报》2017年07期

【摘要】：为解决静态代理模型非线性结构优化效率、精度低的问题,采用最小二乘支持向量回归机(LSSVR)模型进行T型管液压成形加载路径自适应多目标优化研究.用一个数值算例说明本文方法的有效性,以管与背压冲头的接触面积最大及管的最大减薄率最小为优化目标,以接触面积大于对标仿真值、最大减薄率小于实验值、高度大于实验值为约束条件进行多目标优化设计.采用拉丁超立方体设计构造初始支持向量回归模型,用自适应法将每次迭代中获得的额外取样点添加到重建的支持向量回归机模型,得到帕累托最优解集.用理想点法,选择一个最优妥协解以供工程师选用.在成形高度没有变差的情况下,自适应多目标优化结果的管与被压冲头接触面积比实验值提高了32.42%,最小厚度比实验值增加了14.97%.表明自适应迭代LSSVR模型能够在少量样本下保证优化设计精度和计算效率.
[Abstract]:In order to solve the problem of low precision and efficiency of nonlinear structure optimization of static agent model, the least square support vector regression (LSS) model is used to study the adaptive multi-objective optimization of loading path of T-tube hydroforming. A numerical example is given to illustrate the effectiveness of this method. The maximum contact area between the tube and the backpressure punch and the minimum thinning rate of the tube are taken as the optimization objectives, and the contact area is larger than the standard simulation value, and the maximum thinning rate is less than the experimental value. The multi-objective optimization design is carried out under the constraint condition of the height greater than the experimental value. The initial support vector regression model is constructed by using Latin hypercube design. The additional sampling points obtained in each iteration are added to the reconstructed support vector regression machine model by adaptive method, and the Pareto optimal solution set is obtained. By using the ideal point method, an optimal compromise solution is chosen for the engineer to choose. In the case of no variation in forming height, the experimental value of the contact area ratio between the tube and the pressed punch is increased by 32.42, and the minimum thickness is increased by 14.97 to the experimental value. It is shown that the adaptive iterative LSSVR model can guarantee the optimal design accuracy and computational efficiency under a small number of samples.
【作者单位】：吉林大学汽车仿真与控制国家重点实验室;
【基金】：国家自然科学基金(51175218)
【分类号】：TG394;TP18

【相似文献】