PARAMETER IDENTIFCATION OF FLEXIBLE STRUCTURES USING HEURISTIC OPTIMIZATION METHODS

Rennan Otávio Kanashiro¹; Renan Franco Corrêa²; Fabian Andres Lara-Molina¹; Edson Hideki Koroishi¹

¹ UTFPR-CP; ² UTFPR

doi:10.20906/CPS/COB-2015-1278

Resumo

The Finite Element models provide accurate results, however in experimental applications it is possible to note a significant difference between the analytical predictions and the experimental results. This difference may be related to uncertainties in the geometry configuration, inappropriate material parameters, and inaccurate boundary conditions. In order to cope with this drawback, several studies have dealt with the solution inverse problems embedded in the updating of FE models. Therefore, this paper presents an evaluation of two heuristic optimization methods, one method is the Differential Evolution and the other one is Particle Swarm. Both are used to fit the FE models of flexible structures. The influence of the initial population size is analyzed for each optimization methods. In the end, the results showed the efficiency of the methods.

Palavras-chave: flexible structures; model fitting; heuristic optimization methods; inverse problems