Damage Localization and Quantification Assessment in Structures Using Probabilistic Model Updating

Please enable javascript in your browser.

Zhiyanpour, Zahra | 2020

380 Viewed

Type of Document: M.Sc. Thesis
Language: Farsi
Document No: 52937 (09)
University: Sharif University of Technology
Department: Civil Engineering
Advisor(s): Bakhshi, Ali
Abstract:
This research focuses on probabilistic damage detection in the finite element model by using an application of Bayes' inference rule. Density and elastic modulus are sources of uncertainties in structures, therefore the probabilistic method is chosen instead of deterministic. According to the test's data, the distribution of density and elastic modulus have been updated. Then through Bayes' rule the probability of each damage state updates by modal data. Damage states are defined through a coefficient between 0-1 by multiplying to the inertia of elements of the structure. After each observation, the posterior probabilities are updated and this process happens through the multiplication of the likelihood function and the prior probabilities. In the end, the Bayesian algorithm verified by a simulated three-story frame with a different damage scenario and a real three-story frame in an intact state
Keywords:
Elastisity Modulus ; Bayesian Updating ; Likelihood Ratio ; Probabilistic Damage Localization ; Probabilistic Damage Detection ; Density Modulus Distribution

No TOC