Adaptive structural reliability analysis method based on confidence interval squeezing. (September 2022)

• Record Type: Journal Article
• Authors: Chen, Zequan; Li, Guofa; He, Jialong; Yang, Zhaojun; Wang, Jili
• Is Part Of: Reliability engineering & system safety. Volume 225(2022)
• Publication Date: 2022-09
• Subjects: Structural reliability -- Kriging model -- Adaptive -- Confidence interval -- Parallel additions; Reliability (Engineering) -- Periodicals; System safety -- Periodicals; Industrial safety -- Periodicals; Fiabilité -- Périodiques; Sécurité des systèmes -- Périodiques; Sécurité du travail -- Périodiques; 620.00452
• DOI: 10.1016/j.ress.2022.108639
• Languages: English
• ISSNs: 0951-8320

Abstract

Highlights: The confidence interval of the failure probability for Kriging model is derived. A novel adaptive structural reliability analysis strategy (CIS) is presented. A new global convergence condition is proposed. CIS strategy is suitable for sequence and parallel analysis. Abstract: Kriging-based adaptive structural reliability analysis methods can replace the structural performance function and perform reliability analysis accurately and efficiently by using appropriate learning functions. In this context, a novel adaptive structural reliability analysis strategy, namely, confidence interval squeezing (CIS) method, is presented in this paper. After obtaining the confidence interval of the failure probability on the basis of the given significance level α, CIS considers squeezing the confidence interval at the highest speed possible to improve the accuracy of estimating the failure probability. CIS strives to improve the estimation accuracy of failure probability rapidly rather than paying too much attention to the state of a single candidate sample, which is the primary difference between the CIS method and other learning functions. In addition, a new global convergence condition is proposed based on the confidence interval of failure probability. Considering the difficulty of direct application of CIS, three variants of the CIS method, namely, sCIS method for sequence additions, pCIS method for parallel additions based on a clustering algorithm, and improved pCIS (ipCIS) method, are proposed. Several examples are used to demonstrate that each CIS method can handle the complex limit state function and the engineering problem of implicit functions efficiently and accurately.