The dominant aeroacoustic noise source of wind turbines in the audible range is the trailing edge noise. State of the art in research for its prediction at rotors are semi-analytical models fed by flow statistics from RANS simulations. In the present work, the TNO model by Parchen is enhanced to predict the trailing edge noise at airfoils at high lift coefficients, or even beginning trailing edge separation. Different parameters of the model are modified, which have been identified as relevant for this improvement. Among them, the modification with the most significant impact is the consideration of the Turbulence-Turbulence Interaction in the model equation. Further proposed modifications include a new model for the convection velocity for the Wall Pressure Fluctuations (WPF), a model for the moving axis spectrum which considers the decay of turbulence scales and an extension of Kamruzzaman’s anisotropy model by the effect of the pressure gradient. In sum, all modifications yield the enhanced model developed in this work. An extensive test matrix is available for validation. It contains measurements of WPF and far-field sound spectra on several airfoils and from different wind tunnels, as well as selected hot-wire measurements of the underlying boundary layer. It can be shown that the enhanced model significantly improves the prediction quality of WPF and far-field sound for high lift coefficients, while maintaining a similar good prediction quality as the old models for small lift coefficients. Moreover, it is able to correctly reproduce the shape of the WPF spectra close to the maximum lift coefficient of the polar. As a last step, the new model is transferred to the 3D case of a rotor and validated with measurements of WPF and far field noise on a multi-megawatt wind turbine. The comparison shows a considerable improvement of the spectral levels with the newly developed model.weiterlesen