Frequency domain decomposition method: A comparative study on signal processing for unbiased damping ratio estimates

Abstract

Frequency domain decomposition (FDD) is one of OMA methods in the frequency domain and this method has become well-known among engineering community engaged in the system modal identification due to its capability as a user-friendly and fast processing algorithm. Though, this method has problems in offering an accurate estimation of modal damping ratios, even though natural frequencies and mode shapes can be accurately estimated. The accurate estimation of modal damping is still an open problem and often leads to biased estimates since the errors are stemming from each step in FDD procedures and primarily caused by signal processing. Therefore, the identification of modal damping ratio turns out to be immensely essential in structural dynamics since damping is one of the crucial parameters of resonance. This study is to determine the appropriate signal processing for FDD because signal processing such as the time window, correlation function (CF) and the spectral density (SD) are the main contributors to the bias estimate. The goal of this paper is to provide necessary information on modal damping for reliable estimation.