Localised metamodelling techniques for geotechnical reliability-based analysis

• Tipo: Artigo em Livro de Atas de Conferência Internacional
• Data: 2014

Autores

Sónia Marques (Autor) FEUP; António Topa Gomes (Autor) FEUP; António Abel Henriques (Autor) FEUP

Ata de Conferência Internacional

Título: Geotechnical Safety and Risk IV - Proceedings of the 4th International Symposium on Geotechnical Safety and Risk, ISGSR 2013

Páginas: 283-289

4th International Symposium on Geotechnical Safety and Risk, ISGSR 2013

Hong Kong, 4 December 2013 through 6 December 2013

Indexação

Scopus - 0 Citações

COMPENDEX

Classificação Científica

FOS: Ciências da engenharia e tecnologias > Engenharia civil

CORDIS: Ciências Tecnológicas > Engenharia > Engenharia civil ; Ciências Tecnológicas > Engenharia > Engenharia civil > Geotecnia

Outras Informações

ID Authenticus: P-00G-2DD

DOI: 10.1201/b16058-41

Abstract (EN): Metamodelling is actually a very interesting research area for the replacement of simulation models when the trade-off between efficiency and accuracy is appropriate. Several approaches have been proposed, and localised metamodelling techniques applied in reliability-based analysis may involve the search for the most probable failure point. Response surface methodology has been suggested for the approximation of functions, namely implicit limit state functions. The response surface is typically constructed from a polynomial function fitted at a number of points, but no clear guidance has been given for the effect of varying their location. Results for a geotechnical design example with correlated nonnormal random variables are presented in the light of the quality of the design point approximation. In particular, interpolation methodologies in conjunction with appropriate experimental design are suggested for extracting nonlinear response surfaces whose roughness varies substantially over the input domain, so that sequential algorithms may adapt to nonlinear features © 2014 Taylor & Francis Group, London.

Idioma: Inglês

Tipo (Avaliação Docente): Científica