Application of operational modal analysis and blind source separation / Independent component analysis techniques to wind turbines (bn0562)

Wind turbines are complex structures and their dynamics vary significantly in operation, in comparison to that under stand still (parked) conditions. Presence of rotational loads and considerable aeroelastic effects makes it difficult to apply conventional system identification techniques to wind turbines. Operational Modal Analysis (OMA) is a technique of determining dynamic characteristics of a structure based on the output responses only (Input excitation is assumed to be stochastic and its information is not required). Recently it was shown by means of analytical and experimental studies that Blind Source Separation and Independent Component Analysis (BSS/ICA) techniques, which are more common in bio-imaging, wireless communications etc, can also be utilized for system identification purposes in typical output only scenarios. This paper explores the possibility of applying these techniques (OMA and BSS/ICA) to wind turbines. These techniques lend themselves very well to this situation as wind turbine is a huge structure and also, stochastic excitation of turbine from the wind can be used as input. However, time variant nature of operating wind turbines and presence of harmonic components in excitation, still pose several challenge to application of OMA and BSS. To tackle these issues one at a time, this paper focuses in understanding only the aeroelastic effects, thus avoiding effect of rotational loads (and hence presence of harmonic components in excitation force). Aeroelastic effects are significant even in case of parked wind turbines, thus studies are conducted on simulated data obtained for a parked turbine. Ability of OMA and BSS/ICA techniques in determining damping, which is a combination of both structural and aeroelastic damping, is assessed by means of these studies. An OMA algorithm SSI-DATA (Data based Stochastic Subspace Identification) and a BSS technique called Second Order Blind Identification (SOBI) are used for this purpose. It is shown that these techniques work satisfactorily on parked turbine data, paving way to conduct further studies on more complex case of operational wind turbine.