Open AccessBayesian model updating with consideration of modeling error
Author(s) -
Erliang Zhang,
Pierre Feissel,
Jérôme Antoni
Publication year - 2010
Publication title -
european journal of computational mechanics
Language(s) - English
Resource type - Journals
eISSN - 2642-2085
pISSN - 2642-2050
DOI - 10.13052/ejcm.19.255-266
Subject(s) - modal , bayesian inference , computer science , bayesian probability , polynomial chaos , inference , algorithm , posterior probability , bayesian linear regression , cantilever , statistical model , mathematics , artificial intelligence , monte carlo method , engineering , statistics , structural engineering , chemistry , polymer chemistry
On account of measurement and modeling errors, structural identification is better tackled within the statistical framework. In this work, a complete process of Bayesian inference for the characterization of the dynamic behavior of a linear structure is presented in the frequency domain. The polynomial chaos expansion is adopted as a surrogate model to propagate the parameter uncertainty and thus accelerate the evaluation of their posterior probability distribution. Moreover, one hybrid modal model is proposed by introducing some additional variables so as to deal with the modeling errors. Bayesian updating is validated experimentally on a steel square plate and the proposed hybrid modal model is illustrated numerically on a cantilever beam.