Operational modal analysis of wind turbine drivetrain with focus on damping extraction.

Due to changes in generator topology, wind turbines are operating in much wider speed ranges and thus at more varying
operating conditions. This has a positive influence on the energy production, but results in much wider gear mesh excitation
ranges that can lead to tonalities. As such, a high-quality characterization of the modal model of the gearbox housing and
its interaction with the turbine nacelle is essential to pinpoint the specific resonances responsible for these tonalities. Not
only the knowledge of the resonance frequencies is essential, also the damping should be well characterized. This allows
to rank the modes with regard to importance for design modifications. This paper investigates the performance of different
modal estimation methods to assess these damping values. An SSI, p-LSCF and Polymax Plus estimator are benchmarked
with one another. A sensitivity study is done to optimize the configuration parameters of the different modal estimators
specifically for damping extraction. A simulation example is first used to perform this optimization. Acceleration signals
embedding the modal content of a rotating machine are simulated to serve as inputs for the modal estimators. Once detailed
insights are gained, data of a wind turbine is used. Acceleration measurements along the gearbox housing together with
encoder measurements serve as inputs. Damping estimates from rotating data sets are compared. The impact on the
accuracy of the damping estimates is illustrated for the different modal estimators.

Keywords:Modal analysis, Harmonics