Logo
fct Faculdade de
Ciências e Tecnologia
da Universidade Nova de lisboa

UNIDEMI

Research and Development Unit for Mechanical and Industrial Engineering

Project

Projects > Adaptive methods for reliability analysis of complex structures – ADAPTREL
Adaptive methods for reliability analysis of complex structures – ADAPTREL

The main objective of the present proposal was to develop efficient methods for reliability analysis of complex and realistic structural systems using advanced non-linear Finite Element (FE) methods. Although several methods have been proposed and applied in the reliability analysis of simple structural elements or simplified reliability formulations, the generalized use of reliability approaches in large-sized structural systems has not been simple and straight-forward. The increase in the accuracy of structural performance prediction of large systems has been met with a large increase in computational cost and, therefore, the safety assessment of complex and realistic structural systems, although possible, is still costly to be practical. This fact has limited the use of the structural reliability methods mainly to evaluate the implicit safety levels of relatively simple structures and to calibrate design formats where a consistent reliability level is required. Although this has been an important area of application, efficient methods for reliability analysis of complex and realistic structural systems are required so that probabilistic tools can be successively applied to unsolved problems and new challenges in the structural engineering where risk and reliability methodologies are significant contributors.

 

Partners: FCT, IST, UNIDEMI, CENTEC

 

Funding organization: Science and Technology Foundation (PT)

 

Main achievements: The project addressed several methodologies used in the safety assessment of complex structures, allowing the extension of the current knowledge in several areas: - Monte Carlo methods with importance sampling - Adaptive response surface techniques - Surrogate models (Artificial Neural Network, Kriging interpolation) - FORM method combined with design sensitivity analysis Three software packages were developed during project: ProRel, PROAES, EFFECT Two Phd thesis and 5 Master thesis were concluded in the context of the project. From a total of 24 papers produced, 19 are already published or accepted (11 in international journals, 6 at international conference proceedings, and 2 at national conference proceedings).

Reference:PTDC/ECM/115932/2009
Start date:01 Mar 2011
End date:01 Sep 2014
Main researcher:Celeste Jacinto, Assistant Professor
Research group:Systems Engineering and Management
Associated Attachments:
  • Not available
• Associated Researchers
No result found
• Associated Papers
No result found
© 2017 UNIDEMI
Projeto Estratégico PEst-OE/EME/UI0667/2014