返回
A Kriging-based adaptive adding point strategy for structural reliability analysis
DOI:10.1016/j.probengmech.2023.103514.png)
摘要
En 中文
Aiming at the problem of structural reliability analysis with complex performance function in practical engineering, an adaptive adding point strategy for structural reliability analysis is proposed by combining Kriging surrogate model and learning function. In the process of structural reliability analysis, in order to reduce the computational cost, the surrogate model is usually used to fit the implicit performance function. Existing learning functions rarely take into account the problem of sample point aggregation due to too many sample points selected from the failure domain (or safety domain) during the fitting process. In order to overcome this defect, a new method can ensure that the added sample points are evenly distributed on both sides of the limit state function, preventing the aggregation of sample points and causing information redundancy, thereby improving the fitting accuracy of the model, accelerating the convergence speed of the sample points and saving the sample space. Through the analysis of four-branch series system, nonlinear oscillator and truss structure, the results show that the algorithm needs less samples than other methods, and the reliability calculation accuracy is higher, which verifies the correctness and efficiency of the proposed method.
Keyword:
Reliability analysis
Kriging model
Control sample size difference
Adaptive adding point
期刊
IF:
3.5
论文数:
1.7K
被引数:
4.1K
机构
引用论文
An active-learning algorithm that combines sparse polynomial chaos expansions and bootstrap for structural reliability analysis结合稀疏多项式混沌展开和bootstrap进行结构可靠性分析的主动学习算法
STRUCTURAL SAFETY
IF6.3
Application of line sampling simulation method to reliability benchmark problems线抽样模拟法在可靠性基准问题中的应用
STRUCTURAL SAFETY
IF6.3
The stepwise accuracy-improvement strategy based on the Kriging model for structural reliability analysis基于Kriging模型的结构可靠性分析的逐步提高精度策略

