A CRE model updating method in structural dynamics with uncertain measurements

Abstract

This chapter deals with the problem of model validation and updating in the presence of measurement uncertainties. The interpretation of the validation method based on the “Constitutive Relation Error” (CRE) in the framework of the general theory of inverse problems proposed by Tarantola is presented. The extension of this method to the case of uncertain experimental data leads to an appropriate choice of the norm to be used for the distance between the model and the experimental results, to minimize the effect of noise on the process of model validation and updating. This approach does not provide a satisfactory tool for model validation because it does not enable one to quantify the quality of a model with respect to uncertain measurements.