Supervisorer: Andreas Fischer, DTU Vindenergi - Franck Bertagnolio, DTU Vindenergi
Eksaminator: Peter Møller Juhl, Syddansk Universitet
Titel: Analytical Investigation of the Noise Emitted from Aerofoils with a Serrated Trailing Edge
Climate change has become an increasing concern in society throughout the years.
As a result, the demand for renewables like wind energy is increasing as well. However, the public acceptance, due to the noise and appearance of the wind turbines, is threatening the generation of onshore wind farms. Many past studies have focused on the emission of aerodynamic noise from wind turbines. The most dominant source of aerodynamic noise has shown to be the turbulent boundary layer trailing edge noise emitted from the wind turbine wings, Oerleman et al. [19]. So far, the trailing edge noise level is held below a limit, causing a non-optimal overall performance of the wind turbine. Therefore, trailing edge serrations, of which analytical noise predictions were first presented by Howe [14, 15], are developed in order to lower the noise emission without affecting the performance of the turbine.
In the work presented in this thesis the analytical derivation of the noise prediction model for serrated trailing edges, developed by Lyu et al. [18] and based on the work by Amiet [1], is reevaluated. Following the analytical revision, the 2nd order model is implemented to the existing TNO type model, first implemented by Fischer et al. [12] and later optimized by Georgakakis [13]. The implementation of the 2nd order solution is justified, since it has been shown in [12] that the model needs more accuracy. As the model is successfully implemented, it has been optimized in cases of symmetry along the trailing edge in order to decrease the computational time.
Furthermore, a truncation study is provided where a simple linear relation between the frequency and the summation mode n is implemented. Through previous work, Georgakakis [13], an error in the directivity of the model has been suspected. Therefore, this thesis provides an analysis of the directivity using the straight- (Amiet [1]) and serrated- (Lyu et al. [18]) trailing edge noise prediction models, respectively. Thus, the analysis shows that a problem is indeed appearing in the directivity, causing non-reliable estimations of the far field sound spectrum. For validation of the implemented 2nd order model, aerofoil data and measurement results from noise experiments have been used. The computational validation results are for all trailing edge setups obtained at an observer position = 90o above the midspan trailing edge, as well as by an average of 10 observer positions along the experimental microphone array. Comparing the results with the provided experimental data shows a significantly better estimation for the position at = 90o, concluding that the directivity of the model needs improvement.