The use of antiresonances for robust model updating

In this paper, an updating technique that includes antiresonances in the definition of the output residual is considered. Antiresonances are not a global system property, but are typical of each frequency response function (FRF), thus allowing the residual vector to be enlarged with data identified from additional FRFs. However, antiresonance information is not independent of mode shape information; it is rather an alternative, which is preferable for several reasons. Antiresonances can be identified from experimental FRFs with much less error than mode shapes; furthermore, correlation between test and analysis antiresonances is a good index of the correlation between test and analysis FRFs. In the implementation of the technique, matching problems arise whenever antiresonances identified from transfer FRFs are used; unlike the situation for point FRFs, the distribution of antiresonances may be significantly altered by small changes in the structural model. Such problems may be circumvented by restricting the experimental database to point FRFs; in this case, the procedure is quite robust and excellent results are obtained, although it is necessary to plan experimental testing differently from the usual modal testing, with possible impact on related costs. For this reason, a procedure to deal with transfer FRFs by establishing a correlation between test and analysis FRFs at antiresonances using frequency domain assurance criterion (FDAC), is also evaluated. The procedure is not very robust and requires special attention to give acceptable results. (C) 2000 Academic Press.