Multi-objective optimization framework for finite element model updating and response prediction variability.

A multi-objective optimization framework based on modal data is presented for finite element model updating in structural dynamics. The framework results in multiple Pareto optimal structural models that are consistent with the measured data and the norms used for reconciling finite element models with data. Computationally efficient methods for estimating the gradients and Hessians of the objective functions with respect to the model parameters are proposed and shown to significantly reduce the computational effort for solving the single or multi-objective optimization problems. Theoretical and computational developments are addressed and demonstrated by updating the finite element model of a concrete bridge structure using modal data identified from ambient acceleration time history measurements. The results clearly indicate that there is wide variety of Pareto optimal structural models that trade off the fit in various measured quantities. The variability in Pareto models affect the variability in response predictions.