Eventos Anais de eventos
COBEM 2021
26th International Congress of Mechanical Engineering
Bayesian probabilistic approach for railway bridge model updating
Submission Author:
Thiago Moreno Fernandes , SC
Co-Authors:
Thiago Moreno Fernandes, Rafael Holdorf Lopez
Presenter: Thiago Moreno Fernandes
doi://10.26678/ABCM.COBEM2021.COB2021-1669
Abstract
In structural engineering, there are many cases which design data is insufficient or even nonexistent. Thus, there is a demand to assess the structural integrity of these constructions, usually depending on experimental field data. In order to provide information for helping on assessing the structural safety, researches have given special attention to model updating techniques, as strategies based on finite element models to simulate the behavior of the structure. The main idea of the model updating is to improve the match between theoretical model and experimental results. In this context, Bayesian approach can be an efficient tool in estimating parameters of interest. A Bayesian methodology summarize the degree of information regarding these parameters by means of probability distributions. This technique also takes into account errors and uncertainties inherent to the update process. In this work, it is proposed to estimate unknown design variables by means of the Bayesian numerical model updating based on experimental modal responses. A real railway bridge structure is studied. One numerical model of the railway bridge is updating in order to estimate two parameters of interest of structure: vertical stiffness of the supports and modulus of elasticity of concrete. In addition, the standard deviation of errors between experimental and numerical measurements is estimated, in order to understand the potential of Bayesian inference. Finally, it is presented that in the creation of numerical models to assess real structures, one must be very sparing in order not to make a problem overly complex and, as result of, high computational cost.
Keywords
bayesian model updating, Finite Element Model, Parameters Identification, Structural Safety, Railway Bridges
