Comparing optimization methods for Finite Element model updating of a reinforced concrete structure

Different local as well as global optimization procedures can be applied to calibration of numerical models of structures based on experimental data. Gradient-based procedures, on one hand, and evolutionary algorithms, on the other hand, show different performance in terms of accuracy and computational efforts. The present paper reports a comparative assessment of different local as well as global optimization procedures when applied to Finite Element (FE) model updating of a reinforced concrete structure based on the results of Operational Modal Analysis tests. In particular, the FE model of the structure has been refined by minimizing the scatter between the experimental and numerical modal properties. The Douglas-Reid method has been applied for model updating in combination with different gradient-based as well as global optimization procedures in order to evaluate how the choice of the optimization procedure might influence the results; moreover, model updating results are compared and discussed in order to assess their physical significance and the opportunities of model updating for the indirect, non-destructive evaluation of structural properties of existing structures in operational conditions.