论文部分内容阅读
论文采用Kriging模型代替结构真实功能函数,引入主动学习函数,序列选择最佳样本点,在每次迭代中加入最佳样本点更新Kriging模型.与直接的蒙特卡洛方法相比,主动学习Kriging模型仅需要少量的结构分析就能够得到精度较高的可靠度结果,适用于实际工程具有隐式功能函数的结构可靠性分析.论文通过三个数学算例,从最佳样本点的分布情况、功能函数的拟合程度及可靠度计算结果出发对四种学习函数进行对比研究,最后对具有隐式功能函数的悬臂板进行可靠度分析.结果表明,主动学习函数的引入,合理选择了Kriging模型所需的样本,提高了计算效率,同时,学习函数的选择对结构可靠性分析结果也存在影响.
In this paper, the Kriging model is used instead of the real function of structure, the active learning function is introduced, and the best sample point of sequence selection is selected, and the best sample point update Kriging model is added in each iteration.Compared with direct Monte Carlo method, active learning Kriging model Only a small amount of structural analysis is needed to get the result of reliability with high accuracy, which is suitable for the structural reliability analysis of implicit function in practical engineering.Through the three mathematical examples, the distribution of the best sample point, the function The fitting degree of the function and the reliability calculation results, the reliability of the four kinds of learning functions is studied, and finally the reliability of the cantilever plate with implicit function is analyzed.The results show that the introduction of the active learning function, the rational choice of the Kriging model The required samples improve the computational efficiency. At the same time, the choice of learning function also has an impact on the results of structural reliability analysis.