A generalized operational modal analysis framework for challenging no-next engineering applications

Operational Modal Analysis (OMA) allows systems structural identification in working conditions, moving from output data only. Although OMA appears as a revolutionary technique, its main drawback is that, to be applicable, one must assume that the excitations are modelled as uncorrelated white noises. These assumptions, in fact, are heavily binding in all those cases in which the present environmental actions cannot be described as white noises: systems having rotating parts (machine tools, engines or wind turbines) or characterised by speed and/or time correlated inputs (road and rail vehicles). In this paper we perform the modal estimation applying a frequency domain procedure based on the description of a generalized modal model of output PSDs (Power Spectral Densities). Specifically, the proposed OMA technique requires a limited knowledge about features related to the source of excitation and, therefore, it is applicable to systems for which something about the inputs is somehow known.