A Perspective on Model Updating Performance

This paper describes a study of the effects of discretisation on model updating analysis. Minimising the discrepancy between FE model predictions and experimental modal test data is the primary objective of model updating. Numerical case studies presented in this paper show the influence on the final values of the updating parameters for an initial FE model with and without discretisation errors and/or parameter errors. It is concluded that the final values of the updating parameters are composed of two components: those that represent the true parameter errors of the model and those that compensate for the approximations in the discretisation of the model. These two components may will be of the same order of magnitude.