Accelerated subset simulation with neural networks for reliability analysis

Papadopoulos, V; Giovanis, DG; Lagaros, ND; Papadrakakis, M

dc.contributor.author	Papadopoulos, V	en
dc.contributor.author	Giovanis, DG	en
dc.contributor.author	Lagaros, ND	en
dc.contributor.author	Papadrakakis, M	en
dc.date.accessioned	2014-03-01T02:07:34Z
dc.date.available	2014-03-01T02:07:34Z
dc.date.issued	2012	en
dc.identifier.issn	00457825	en
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/29580
dc.subject	Markov chain	en
dc.subject	Metropolis-Hastings	en
dc.subject	Neural networks	en
dc.subject	Reliability analysis	en
dc.subject	Subset simulation	en
dc.subject.other	Computational effort	en
dc.subject.other	Meta model	en
dc.subject.other	Metropolis algorithms	en
dc.subject.other	Metropolis-Hastings	en
dc.subject.other	Network-based	en
dc.subject.other	Numerical example	en
dc.subject.other	Probability of failure	en
dc.subject.other	Proposal distribution	en
dc.subject.other	Reliability prediction	en
dc.subject.other	Robust estimation	en
dc.subject.other	Sub-domains	en
dc.subject.other	Subset simulation	en
dc.subject.other	Algorithms	en
dc.subject.other	Markov processes	en
dc.subject.other	Reliability analysis	en
dc.subject.other	Neural networks	en
dc.title	Accelerated subset simulation with neural networks for reliability analysis	en
heal.type	journalArticle	en
heal.identifier.primary	10.1016/j.cma.2012.02.013	en
heal.identifier.secondary	http://dx.doi.org/10.1016/j.cma.2012.02.013	en
heal.publicationDate	2012	en
heal.abstract	Subset Simulation (SS) is a powerful tool, simple to implement and capable of solving a broad range of reliability analysis problems. In many cases however, SS leads to reliability predictions that exhibit a large variability due to the fact that the robustness of the SS prediction depends on the selection of an adequate width of the proposal distribution when applying the modified Metropolis algorithm. In this work a Neural Network-based SS (SS-NN) methodology is proposed in which NN are effectively trained over smaller sub-domains of the total random variable space which are generated progressively at each SS level by the modified Metropolis algorithm. NN are then used as robust meta-models in order to increase the efficiency of SS by increasing significantly the samples per SS level with a minimum additional computational effort. In the numerical examples considered, it is demonstrated that the training of a sufficiently accurate NN meta-model in the context of SS simulation leads to more robust estimations of the probability of failure both in terms of mean and variance of the estimator. © 2012 Elsevier B.V.	en
heal.journalName	Computer Methods in Applied Mechanics and Engineering	en
dc.identifier.doi	10.1016/j.cma.2012.02.013	en
dc.identifier.volume	223-224	en
dc.identifier.spage	70	en
dc.identifier.epage	80	en