Finite-element-model updating of civil engineering structures using a hybrid UKF-HS algorithm

Abstract

Finite-element-model updating allows reducing the discrepancies between the numerical and the experimental dynamic behaviour of civil engineering structures. Among the different methods to tackle the updating problem, the maximum likelihood method has been widely used for practical engineering applications. In this method, the updating problem is transformed into an optimization problem where the relative differences between the numerical and experimental modal properties of the structure are reduced via the modification of the most relevant physical parameters of the model. However, this method often presents the drawback of requiring high simulation times in order to perform the updating process when dealing with complex structures. To overcome this limitation, in this paper a novel hybrid Unscented Kalman Filter – Harmony Search (UKF-HS) algorithm is proposed and its implementation details are discussed. In order to validate such hybrid algorithm and further illustrate its performance, the finite-element-model updating of a benchmark footbridge is performed using two different approaches (single-objective and multi-objective) and three different computational algorithms, namely: (i) genetic algorithms; (ii) harmony search; and (iii) the novel UKF-HS hybrid algorithm. The obtained results reveal that the proposed hybrid algorithm may be considered as an adequate alternative tool to efficiently perform the finite-element-m