A saddlepoint approach to estimating joint extreme value distributions for vector non-stationary Gaussian processes. (June 2016)

• Record Type: Journal Article
• Authors: Ambetkar, Vighnesh; Kuppa, Ramakrishna; Gupta, Sayan
• Is Part Of: Applied ocean research. Volume 58(2016:Oct.)
• Page Start: 178
• Page End: 188
• Publication Date: 2016-06
• Subjects: Extreme value distribution -- Non-stationary Gaussian process -- Poisson counting process -- Saddlepoint approximation; Ocean engineering -- Periodicals; 620.416205
• DOI: 10.1016/j.apor.2016.04.005
• Languages: English
• ISSNs: 0141-1187

Abstract

Abstract : Highlights: Methodology developed for time variant system reliability analysis for a vibrating structural system. Analytical approximations for joint EVD for vector non-stationary Gaussian processes. Methodology uses principles of random process theories and crossing statistics. The saddlepoint approximation method used to develop the approximations. Abstract: The focus of this study is on estimating the multivariate extreme value distributions associated with a vector of mutually correlated non-stationary Gaussian processes. This involves computing the joint crossing statistics of the vector processes by assuming the crossings to be Poisson counting processes. A mathematical artifice is adopted to take into account the dependencies that exist between the crossings of the processes. The crux in the formulation lies in the evaluation of a high-dimensional integral, which can be computationally expensive. This difficulty is bypassed by using saddlepoint approximation. The developments are illustrated through two numerical examples and are validated using Monte Carlo simulations. In the second example, reliability analysis is carried out for a multiply supported pipeline on an offshore jacket structure subjected to wave loading using the proposed formulation.