Conditioning of FRF measurements for use with frequency based substructuring

Frequency based substructuring approaches have been used for the generation of system models from component data. While numerical models show successful results, there have been many difficulties with actual measurements in many instances. Previous work has identified some of these typical problems using simulated data to incorporate specific measurement difficulties commonly observed along with approaches to overcome some of these difficulties.This paper presents the results using actual measured data for a laboratory structure subjected to both analytical and experimental studies. Various commonly used approaches are shown to illustrate some of the difficulties with measured data. A new approach to better condition the measured functions and purge commonly found data measurement contaminants is utilized to provide dramatically improved results.Several cases are explored to show the difficulties commonly observed as well as the improved conditioning of the measured data to obtain acceptable results.