MODAL PARAMETER IDENTIFICATION OF A RECTANGULAR BEAM USING EXPERIMENTAL AND OUTPUT-ONLY MODAL ANALYSES

Resumo

In order to obtain the modal parameters of a structure, different approaches can be used: (i)
Experimental Modal Analysis (EMA), which uses excitation (input) and response (output) data; (ii)
Output-Only Modal Analysis (O-OMA), which uses only the output. Both techniques are widely used
and their advantages and disadvantages are known. The EMA allows performing a full numerical
validation, because the estimated modal matrix is already mass normalized. In the O-OMA case, since
the excitation force is unmeasured, the estimated modal matrix can only be normalized to unity. This
paper aims to identify the modal parameters of a rectangular beam, simulating “free-free” boundary
condition in the excitation’s direction. The beam was instrumented with 9 accelerometers equidistantly
arranged along the length and the excitation was through a multiple reference impact testing. The
modal parameters were identified using the Polyreference Least Square Complex Frequency Domain
method, considering two approaches: (i) EMA and (ii) O-OMA methods. Results from the EMA and
O-OMA showed that the discrepancies between the natural frequencies were at most 0.5%, and all the
vibration modes presented MAC values higher than 0.9. Experimental results were also compared with
those obtained through a finite element model of the rectangular beam, showing small differences both
in the natural frequencies and vibration modes. It is therefore concluded that the two approaches lead
to good estimates of natural frequencies. As for the vibration modes, the modal matrices obtained by
the two methods present a small difference, since the mass of the beam is small.