A robust structural design method using the Kriging model to define the probability of design success
A robust structural design method using the Kriging model to define the probability of design success
复制标题
使用克里金模型定义设计成功概率的稳健结构设计方法
DOI：
10.1243/09544062jmes1736
复制
发表时间：
2010
期刊：
影响因子：
--
通讯作者：
K
中科院分区：
文献类型：
--
作者：
K
摘要通过引入Kriging近似模型和定义设计成功概率，提出了一种稳健优化方法。稳健优化中的一个关键问题是响应的均值和方差不容易计算。本研究提出了一种基于Kriging元模型的近似统计矩方法的实现。在此基础上，采用二阶统计逼近方法进行统计，避免了局部稳健最优。因此，设计成功概率被定义为满足所施加的设计要求的概率，它被表示为近似均值和方差的函数。稳健优化的公式可以定义为每个响应的设计成功的概率。为了验证所提方法的有效性，对双杆结构和微陀螺仪的数学问题和设计问题进行了研究。
Abstract In this study, a robust optimization method is proposed by introducing the Kriging approximation model and defining the probability of design-success. A key problem in robust optimization is that the mean and the variation of a response cannot be calculated easily. This research presents an implementation of the approximate statistical moment method based on the Kriging metamodel. Furthermore, the statistics using the second-order statistical approximation method are adopted to avoid the local robust optimum. Thus, the probability of design-success, which is defined as the probability of satisfying the imposed design requirements, is represented as a function of approximate mean and variance. The formulation for the robust optimization can be defined as the probability of design-success of each response. The mathematical problem and the design problems of a two-bar structure and microgyroscope are investigated for the validation of the proposed method.