Bayesian operational modal analysis of closely spaced modes for monitoring wind turbines

Abstract

In this study, the applicability of Bayesian operational modal analysis (BAYOMA) to an operating onshore concrete-steel hybrid wind turbine tower is investigated. The results of the identification then provide reliable parameters for the structural health monitoring (SHM) of the tower. In the context of wind turbines, typical assumptions of linear time-invariant OMA methods are violated, so the validity of the identification uncertainties of BAYOMA is not necessarily given. In addition, closely spaced modes occur, for which the mode shape in particular is subject to high uncertainty. It can be stated, that the main part of the mode shape uncertainty corresponds to the alignment of these in the mode subspace. Due of these challenges, mode shapes are generally not taken into account when monitoring wind turbine towers. In order to include the mode shape in SHM scheme, the second-order modal assurance criterion (S2MAC) is applied in this study. This metric is able to eliminate the alignment uncertainty by comparing the mode shape with a mode subspace. Besides mode shapes, the reliability of natural frequencies and damping can also be better quantified by knowing the identification uncertainty. This finally enables a well-founded selection of suitable monitoring parameters for the future application of SHM for wind turbines. Preprint submitted to Engineering Structures.