This presentation focuses on the development of the Bayesian statistical framework in the area of structural model updating and damage detection. Unlike the deterministic approach, which targets in pinpointing a single model to represent the structure, the Bayesian approach aims in calculating the posterior (updated) probability density function (PDF) of the adjustable model parameters under a given model class and for a given set of measurement. One of the main problems in structural model updating is the selection of a class of models with “suitable” complexity. For a given set of measured data, the uncertainties associated with the identified model parameters will be unacceptably high when the model class is too complex. This presentation will cover the recent development of the use of Bayesian model class selection in addressing this problem. Several examples about the application of the Bayesian model updating and model class selection in structural damage detection will be presented.

Speaker
Dr. LAM, Heung-Fai is currently assistant professor in the Department of Building & Construction at the City University of Hong Kong (CityU). He is currently the associate editor of the International Journal of Applied Mathematics and Mechanics (IJAMM) [http://ijamm.bc.cityu.edu.hk]. Furthermore, he is a committee member of the IASC-ASCE Task Group on Structural Health Monitoring (SHM) (2003 – present), and member of the System Identification and Structural Control (Subcommittee 5) of the International Association for Structural Safety and Reliability (2005 – present). He is one of the main contributors in the development of the IASC-ASCE SHM benchmark study.