Substructural identification of dynamic properties of composite structures

The paper presents the finite element model updating algorithm based on a substructuring method. The proposed algorithm enables model order reduction of selected substructures in order to reduce the necessary computational time. The model updating procedure consist of two levels: (i) local updating, based on updating the selected substructures and (ii) global updating based on updating the global coupled model. The minimization criterion is based on inverse variance weighting. The decisive variables can be material parameters such as Young modulus, density, and loss factor, which vary in the uncertainty limits. Additionally, the presented algorithm enables the estimation of uncertainty levels of identified model parameters. The application of algorithm was presented on the steel-polymer concrete beam example. Achieving the decrease of maximum relative error for natural frequencies from 13.4 % to 6.2 %, and its average value from 6.9 % to 2.0 %. Moreover, a substantial improvement was achieved in mapping the frequency response functions in both cases.