On the use of consistent and significant information to reduce ill-conditioning in dynamic model updating

A dynamic model updating technique, force residual updating-interactive technique (FRU-IT), is described. The technique is based on minimisation of the force (input) residual. Like most parametric identification problems, dynamic model updating is affected strongly by measurement errors due to the ill-conditioning of the problem. The proposed method is used to define and solve the set of updating equations in order to minimise the influence of measurement errors. This includes the selection of experimental degrees of freedom to ensure data consistency, the choice of the working frequencies to pursue data significance, and the use of regularisation techniques; specifically, an interactive use is made of the truncated singular value decomposition. The performance of the method is demonstrated by test cases.