Maximum Likelihood Finite-Element Model Updating of Civil Engineering Structures Using Nature-Inspired Computational Algorithms

Abstract

In finite-element model updating, numerical models are calibrated in order to better mimic the real behaviour of structures. Such updating process is usually performed under the maximum likelihood method in practical engineering applications. According to this, the updating problem is transformed into an optimization problem. The objective function of this problem is usually defined in terms of the relative differences between the numerical and the experimental modal properties of the structure. To this aim, either (1) a single-objective or (2) a multi-objective approach may be adopted. Due to the complexity of the problem, global optimizers are usually considered for its solution. Among these algorithms, nature-inspired computational algorithms have been widely employed. Nevertheless, such model updating approach presents two main limitations: (1) a clear dependence between the updated model and the objective function considered; and (2) a high computational cost. In order to overcome these drawbacks, a detailed study has been performed herein both to establish the most adequate objective function to tackle the problem and to further assist in the selection of the most efficient computational algorithm among several well-known ones. For this purpose, a laboratory footbridge has been considered as benchmark to conduct the updating process under different scenarios.