A new method for finite element model updating in structural dynamics

A method is propounded in this paper to update finite element models in the field of structural dynamics. It is especially intended to solve cases for which other methods either cannot be applied or are inefficient. It is posed as the minimization of an error function defined in the time domain. The minimization is carried out by a novel adaptive sampling algorithm, which constitutes the main contribution in this work. In this algorithm, the solution is searched by sampling the parameters to be updated within a bounded space in an iterative form. The Beta distribution, which is consistent with the bounded character of the search space, is chosen for the sampling. The characteristics of the distributions are changed in each step of the process depending on the results of the previous ones. The method has been tested through a simulated dynamic model and an experimental case. In these particular cases, the performance of the proposed adaptive algorithm was better than those of other related stochastic algorithms it was compared with.