On Application of Bayesian Model Updating for Buildings and Bridges
Please login to view abstract download link
We present findings and lessons learned from applying Bayesain finite element model updating on several taller timber buildings and on a steel footbridge, by using measured vibration data. The features of such an updating procedure, which may produce a pretty reliable finite element model of a large civil engineering structure under service loadings, are illustrated and discussed. The procedure was effectively applied to explore influential parameters of a numerical model, as well as to identify sources of modelling error. The considered buildings are three taller multi-storey timber buildings, with two of them, which are made entirely of cross-laminated timber, being seven and eight storeys high, while the third one is a five-storey hybrid timber-concrete building also using cross-laminated timber. The considered bridge is a steel footbridge with an arc that is spanning over a highway. The updating procedure for these structures was based on detailed finite element models, modal testing results (obtained by either forced vibration or ambiental vibration tests), sensitivity analysis, classical and Bayesian model updating, and uncertainty quantification. Additionally, Bayesian inference was used for maintaining computational digital twin of a steel frame-like laboratory structure, which will be reported as well.
