Comparison of response transmissibility and power spectral density transmissibility on operational modal analysis

Due to the independence of input spectra, the transmissibility-based operational modal analysis (OMA) methods have been developed rapidly over the last two decades. The response transmissibility (RT) methods can extract the modal parameters from multiple different load cases while the power spectral density transmissibility (PSDT) methods just use one single load case by combining multiple transfer outputs. Their ability to identify modal parameters under non-white excitations has been claimed, verified and applied by many researches. However, some problems, such as the robustness under different load conditions, the extent of noise influence and the application conditions, still arise when they are applied in the real field. Moreover, the comparison between different transmissibility-based methods on the identification accuracy and robustness is rare, which is significant for users to select the very method for their tasks. Four representative transmissibility-based OMA methods are considered and compared in this paper. The influence factors of RT and PSDT functions are analyzed first and the results reveal the requirements of the load case for four OMA methods. A numerical example is used to assess and compare the robustness of four methods to measurement noises, damping ratio of non-white noise excitations, non-fully-excited load cases and harmonic excitations, and their identification ability is further evaluated through a real field example. The results show that the transmissibility-based OMA methods are independent of input spectra only under certain conditions and some user-defined parameters play important roles in their identification ability.