Institutional Repository

Reliability estimation using Markov chain Monte carlo-based tail modeling

Show simple item record

dc.contributor.author Bayrak, Gamze
dc.contributor.author  Acar, Erdem
dc.date.accessioned 2019-03-16T07:46:31Z
dc.date.available 2019-03-16T07:46:31Z
dc.date.issued 2018-03
dc.identifier.citation Bayrak, G., & Acar, E. (2017). Reliability Estimation Using Markov Chain Monte Carlo–Based Tail Modeling. AIAA Journal, 56(3), 1211-1224. en_US
dc.identifier.issn 00011452
dc.identifier.uri https://arc.aiaa.org/doi/pdf/10.2514/1.J055947
dc.identifier.uri http://hdl.handle.net/20.500.11851/723
dc.description.abstract Tail modeling is an efficient method used in reliability estimation of highly safe structures. Classical tail modeling is based on performing limit-state function evaluations through a sampling scheme, selecting a threshold value to specify the tail part of the cumulative distribution function, fitting a proper model to the tail part, and estimating the reliability. In this approach, limit-state function calculations that do not belong to the tail part are mostly discarded, and so majority of limit-state evaluations are wasted. In this paper, Markov chain Monte Carlo method with Metropolis–Hastings algorithm is used to draw samples from the tail part only so that a more accurate reliability index prediction is achieved. A commonly used proposal distribution formula is modified by using a scale parameter. The optimal value of this scale parameter is obtained for various numerical example problems with a varying number of random variables, and an approximate relationship is obtained between the optimal value of the scale parameter and the number of random variables. The approximate relationship is tested on the reliability prediction of a horizontal axis wind turbine and observed to work well. It is also found that the proposed approach is more accurate than the classical tail modeling when the number of variables is less than or equal to four. For a larger number of random variables, none of the two approaches are found to be superior to another. en_US
dc.language.iso eng
dc.publisher American Institute of Aeronautics and Astronautics Inc.
dc.relation.isversionof 10.2514/1.J055947
dc.rights info:eu-repo/semantics/closedAccess en_US
dc.subject Reliability analysis en_US
dc.subject Reliability en_US
dc.subject Time-dependent reliability en_US
dc.title Reliability estimation using Markov chain Monte carlo-based tail modeling en_US
dc.type article en_US
dc.relation.journal AIAA Journal
dc.contributor.department TOBB ETÜ, Mühendislik Fakültesi, Makine Mühendisliği Bölümü
dc.contributor.department TOBB ETU, Faculty of Engineering, Department of Mechanical Engineering en_US
dc.identifier.volume 56
dc.identifier.issue 3
dc.identifier.startpage 1211
dc.identifier.endpage 1224
dc.relation.tubitak info:eu-repo/grantAgreement/TÜBİTAK/MAG/214M205
dc.contributor.orcid https://orcid.org/0000-0002-3661-5563
dc.identifier.wos WOS:000426601300025
dc.identifier.scopus 2-s2.0-85043253515


Files in this item

Files Size Format View

There are no files associated with this item.

This item appears in the following Collection(s)

Show simple item record