Abstract

Up to now, existent FRF based model updating methods use the differences between measured and analytical FRFs at a fixed frequency, as residual to minimize. This approach does not take into account that FRFs between a reference model (experimental) and a perturbed one (a finite elements model not yet updated), displace in two axes : amplitude and frequency. A more physical correlation, then, uses also the frequency shift . The problem is how to find it.Taking as base the well known Modal Assurance Criterion (MAC), the Modal Scale Factor (MSF), and the concept of frequency shift, two correlation techniques, the Frequency Domain Assurance Criterion (FDAC) and the Frequency Response Scale Factor (FRSF), are presented. They help to quantify the level of correlation between responses coming from the two models, determine the frequency shift at all measured frequencies, and establish a suitable set of frequencies to use during the updating procedure.