On the Initial Assessment of a New Complex Frequency Domain Decomposition (C-FDD) Technique Formulated in Modal Model

Regardless of the parametric system identification technique used when performing Operational Modal Analysis (OMA), it is always helpful to use a non-parametric identification technique based on modal decomposition such as the well-known Frequency Domain Decomposition (FDD) to cross check the identification results. Because the parametric techniques tend to provide both physical and non-physical (or noisy) modes, a comparison with the decomposition of the spectral densities gives the analysts the possibility to distinguish former from latter and, therefore, increase the robustness and accuracy of the OMA. In this paper, a new modal decomposition approach formulated in frequency domain modal model is introduced. The underlying idea in this formulation is to write the modal model Half Spectrum for two different frequency lines. In order to assess its performance, the introduced decomposition approach is applied to a practical application example, and the obtained decomposition is compared to that of the classic FDD.