engleski

Finite element model updating of bridge structures using game theory based algorithm

In this research, an implementation of game theory has been developed as a computational tool to assist designers in solving the maximum likelihood finite element model updating problem of bridge structures. Finite element model updating is usually performed using the deterministic maximum likelihood method by defining the updating problem as an optimization problem in terms of a single or multi-objective function. In this way, a numerical model is obtained that reflects the actual structural behaviour. Despite its widespread use, the maximum likelihood method has some limitations that are directly related to the performance of the updating process. To overcome these limitations, game theory has been considered herein. For this purpose, the conventional updating problem, formulated as either a single or a multi-objective function, has been transformed into a game theory problem, considering three different game models: non-cooperative, cooperative and evolutionary. The performance of the proposal has been assessed on two case studies the simple finite element model of the laboratory bridge and a high-fidelity finite element model of the real pedestrian suspension bridge. As a result of this research, game theory has proven to be an efficient tool to improve the performance of the finite element model updating process under the maximum likelihood method, as it ensures a reduction in computational time without compromising the accuracy of the solution.

finite element model updating, game theory, maximum likelihood method, structural dynamic parameters, bridge structures

nije evidentirano