POD analysis of the wake dynamics of an offshore floating wind turbine model

Abstract

The wake dynamics of a floating wind turbine model experiencing realistic surge motion and immersed within a properly scaled atmospheric boundary layer is studied through wind tunnel experiments. The turbine is modelled by a porous disk representing the floating 2MW wind turbine located at the offshore test site in Le Croisic (France). Experiments were conducted in the LHEEA’s atmospheric wind tunnel. A surge motion is imposed on the model, using a linear actuator, to replicate realistic behaviors under wave swell. Realistic frequencies of actuation are considered to study their effects on the wake properties. The wake is characterized using Stereoscopic Particle Image Velocimetry (SPIV) measurements in a y - z plane normal to the flow, at two different streamwise locations x = 4.6D and 8.1D. In addition to the documentation of the main wake statistics, the velocity fields are analyzed using Proper Orthogonal Decomposition (POD). The velocity field is decomposed into a set of spatial and temporal modes. The eigenvalues convergence is shown to be relatively slow, due to the high Reynolds number turbulent boundary layer within which the model is immersed. When varying the surge motion frequency, the spatial modes do not show any significant change in shape and amplitude. However, the spectral analysis performed on the temporal modes shows the emergence of peaks at the surge motion frequency and the overall increase of the low-frequency energy content in the Power Spectral Density, in particular for the highest frequencies of motion tested.