Review and application of Artificial Neural Networks models in reliability analysis of steel structures

Abstract

This paper presents a survey on the development and use of Artificial Neural Network (ANN) models in structural reliability analysis. The survey identifies the different types of ANNs, the methods of structural reliability assessment that are typically used, the techniques proposed for ANN training set improvement and also some applications of ANN approximations to structural design and optimization problems. ANN models are then used in the reliability analysis of a ship stiffened panel subjected to uniaxial compression loads induced by hull girder vertical bending moment, for which the collapse strength is obtained by means of nonlinear finite element analysis (FEA). The approaches adopted combine the use of adaptive ANN models to approximate directly the limit state function with Monte Carlo simulation (MCS), first order reliability methods (FORM) and MCS with importance sampling (IS), for reliability assessment. A comprehensive comparison of the predictions of the different reliability methods with ANN based LSFs and classical LSF evaluation linked to the FEA is provided.

Citation

Chojaczyk, A., Teixeira, A., Neves, L. C., Cardosa, J., & Soares, C. G. (2015). Review and application of Artificial Neural Networks models in reliability analysis of steel structures. Structural Safety, 52(A), https://doi.org/10.1016/j.strusafe.2014.09.002

• Journal Article Type: Article
• Acceptance Date: Sep 14, 2014
• Online Publication Date: Oct 11, 2014
• Publication Date: Jan 1, 2015
• Deposit Date: May 22, 2017
• Publicly Available Date: May 22, 2017
• Journal: Structural Safety
• Print ISSN: 0167-4730
• Electronic ISSN: 0167-4730
• Publisher: Elsevier
• Peer Reviewed: Peer Reviewed
• Volume: 52
• Issue: A
• DOI: https://doi.org/10.1016/j.strusafe.2014.09.002
• Keywords: Artificial Neural Networks; Structural reliability; Monte Carlo simulation; Importance sampling; First-order reliability methods; Finite element analysis; Ultimate strength; Stiffened plates
• Public URL: https://nottingham-repository.worktribe.com/output/985332
• Publisher URL: http://www.sciencedirect.com/science/article/pii/S016747301400085X