WIT Press

Neural Networks And Fuzzy Sets: An Alternative For Risk Analysis And Damage Assessment Of Structures

Price

Free (open access)

Paper DOI

10.2495/ERES990241

Volume

41

Pages

10

Published

1999

Size

934 kb

Author(s)

E. Villanueva & M. Sanchez-Silva

Abstract

Appropriate damage assessment procedures are fundamental for making decisions regarding existing structures. For instance, to decide about corrective measures after an earthquake or to define risk management strategies. Damage assessment is one of the most challenging problems to solve due to difficulties in defining, assessing and modelling the variables involved and those associated with handling uncertainty. The proposed methodology is a combination of system theory, neural networks, fuzzy logic and interval theory. A three-layer feedforward neural network is used to classify structural damage. Tests of the model have demonstrated the reliability and value of this tool for making decisions under uncertainty. 1 Introduction One of the most diff

Keywords