Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
10677763 | Applied Mathematical Modelling | 2015 | 40 Pages |
Abstract
In practice, there are many engineering problems characterized by complex implicit performance functions. Accurate reliability assessment for these problems usually requires very time-consuming computation, and sometimes it is unacceptable. In order to reduce the computational load, this paper proposes an efficient reliability method combining adaptive importance sampling and Kriging model based on the active learning mechanism. It inherits the superiorities of Kriging metamodel, adaptive importance sampling and active learning mechanism, and enables only evaluating the interested samples in actual performance function. The proposed method avoids a large number of time-consuming evaluation processes, and the important samples are mainly predicted by a well-constructed Kriging metamodel, thus the calculating efficiency is increased significantly. Several examples are given as validations, and results show that the proposed method has great advantages in the aspect of both efficiency and accuracy.
Related Topics
Physical Sciences and Engineering
Engineering
Computational Mechanics
Authors
Hailong Zhao, Zhufeng Yue, Yongshou Liu, Zongzhan Gao, Yishang Zhang,