Identification of finite element models in structural dynamics

Abstract Structural identification is a technique for restricting uncertainties in structural modelling by making use of available experimental data. For large and complicated systems these data are frequently very limited, thus interpretative models which retain most of the a priori information, such as finite element models, are preferable. Those aspects of identification procedures which use FE models in the frequency domain and a Bayesian context are examined. Attention is given to the optimal choice of parameters and observed quantities, identifiability, expected accuracy and to features of the numerical procedure. A quantitative evaluation of these aspects is undertaken to identify a FE model of a spatial framed structure using pseudo-experimental data.