A METHOD FOR INFERENCE OF THE UNCERTAINTIES OF THE MODAL PARAMETERS THROUGH THE BOOTSTRAP AND CLUSTERING TECHNIQUES ASSOCIATION

Resumo

The evaluation of structural integrity is a complex task that has increased the development of
techniques for monitoring the physical responses of structural systems. Monitoring can be accomplished
by acquiring different types of response from the structure to the loads that act on it through continuous
signals in time. These signals can be studied through modal analysis, obtaining estimates of the modal
parameters and, after them, global information of the structural integrity. In this context, it is of
paramount importance to quantify the uncertainties of the estimated modal parameters. An inaccurate
estimate can compromise the entire monitoring of a structure, and thus avoid a possible interruption of
its operation, or even its collapse. This work aims to develop a modal analysis methodology capable of
obtaining more robust estimates and with lower levels of uncertainty of the modal parameters of a
structure through the association of bootstrap, clustering and Data-driven Stochastic Subspace
Identification (SSI-DATA) techniques. To achieve this goal, simulated data of a structural system
subjected to uncorrelated random loading introducing noise are used. Finally, we discuss the results
found for the uncertainties of the estimated parameters.