Ajuste com quantificação de incertezas de modelos em elementos finitos através de função objetivo baseada na energia de deformação modal

Abstract

This work presents a new methodology for model updating with uncertainty quantification. The methodology is developed in an iterative and non-probabilistic way, considering the numerical and experimental uncertainties of the model updating process. Thus, the structural identification is performed by the SSI method and the uncertainties of the identification through the bootstrap resampling are calculated. For the optimization, the genetic algorithms are used, having as objective function the modal deformation energy and the natural frequency of the system. For the uncertainty quantification, the structural identification, modeling and optimization fluctuations are considered through the adjustments of the "n" samples generated by the bootstrap, performing the calculation of the statistics and then defining the confidence intervals of the adjustment made. Numerical simulations are performed for the evaluation of the implemented methodology, followed by vibration experimental tests conducted in a metal beam and a space frame. The results were satisfactory with the implemented algorithm being able to carry out the structural identification and adjustment of models considering the uncertainties in all the simulations and experimental tests performed.