Algorithms for structural reliability

Prentzas, Ioannis; Πρέντζας, Ιωάννης

dc.contributor.author	Prentzas, Ioannis	en
dc.contributor.author	Πρέντζας, Ιωάννης	el
dc.date.accessioned	2022-07-05T07:49:30Z
dc.date.available	2022-07-05T07:49:30Z
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/55360
dc.identifier.uri	http://dx.doi.org/10.26240/heal.ntua.23058
dc.description	Εθνικό Μετσόβιο Πολυτεχνείο--Μεταπτυχιακή Εργασία. Διεπιστημονικό-Διατμηματικό Πρόγραμμα Μεταπτυχιακών Σπουδών (Δ.Π.Μ.Σ.) "Επιστήμη Δεδομένων και Μηχανική Μάθηση"
dc.rights	Default License
dc.subject	Structural reliability	en
dc.subject	FORM	en
dc.subject	Monte carlo	en
dc.subject	Subset simulation	en
dc.subject	Machine learning	en
dc.title	Algorithms for structural reliability	en
heal.type	masterThesis
heal.secondaryTitle	Computational tools and Machine learning Methods	en
heal.classification	Structural engineering	en
heal.language	en
heal.access	free
heal.recordProvider	ntua	el
heal.publicationDate	2022-03-02
heal.abstract	Structural reliability analysis provides a useful tool for safety assessment of engineering structures and enables performance of more rational risk evaluations. It is an alternative approach to traditional deterministic structural design, which takes into account the uncertain parameters characterizing the physical state of structure and its environment. Generally, structural reliability analysis is convenient and straightforward when the limit state function is formulated with an explicit function. However, in practical engineering, the limit state function is generally expressed as implicit function. The implicit limit state function presents great difficulties in structural reliability analysis when the most common methods are used, such as the first-order reliability method (FORM). Typically, when the implicit limit state function is evaluated implicitly using a numerical code, such as the finite element method. Although reliability analysis can be performed using the Monte Carlo simulation or Subset Simulation, a large number of FEM executions for structural analysis is time consuming, especially for large and complex structures with high reliability. Various regression models in combination with reliability methods have been used to solve reliability analysis problems involving the implicit limit state function. Gaussian process regression, and Support Vector machine are Machine Learning algorithms, which have been applied to approximate the limit state function, shortening the computational time, and the failure probability was predicted using reliability methods, such as Monte Carlo. The structural reliability methods have been applied to three structural analysis problems for calculating the probability of failure. The limit state equation is explicit in the first example, while the equation of the other examples includes the finite element method. Moreover, these equations have been approximated using the two machine learning regressions models.	en
heal.advisorName	Fragiadakis, Michalis	en
heal.committeeMemberName	Fragiadakis, Michalis	en
heal.committeeMemberName	Papadopoulos, Vissarion	en
heal.committeeMemberName	Thanopoulos, Pavlos	en
heal.academicPublisher	Εθνικό Μετσόβιο Πολυτεχνείο. Σχολή Πολιτικών Μηχανικών	el
heal.academicPublisherID	ntua
heal.numberOfPages	90 σ.	el
heal.fullTextAvailability	false