On the Optimal Correction of Structural Dynamic Models

Abstract It is well known that a discrete analytical model cannot accurately represent the dynamics of a continuous physical system over the frequency range of the model. Methods have been presented with the intention of improving analytical models (finite element models in particular) by the use of either (a) measured modal data or (b) measured frequency responses. In this paper we consider in particular the selection of weighting terms and the physical significance of the estimates. In both cases the discussion is illustrated by means of example problems.