UTFacultiesETEventsPhD Defence Andrea Bresciani | Physical and perceptual prediction of wind turbine noise

PhD Defence Andrea Bresciani | Physical and perceptual prediction of wind turbine noise

Physical and Perceptual Prediction of Wind Turbine Noise

The PhD defence of Andrea Bresciani will take place in the Waaier building of the University of Twente and can be followed by a live stream.
Live Stream

Andrea Bresciani is a PhD student in the department Engineering Fluid Dynamics. (Co)Promotors are prof.dr.ir. C.H. Venner from the faculty of Engineering Technology and dr. S. Le Bras, Siemens and dr. J. Maillard, CSTB Research.

The increasing demand for renewable energy has led to a surge in wind turbine installations. However, the noise emitted by wind turbines is hampering their deployment close to inhabited places because of the annoyance caused to the residents. Despite lower sound levels than other common sources, wind turbine noise ranks as the most annoying sound source compared to wind, road, and rail noises. Therefore, it is essential to understand the characteristics of the noise sources, their propagation in the surrounding environment, and to predict the annoyance perceived by the inhabitants before the installation of the wind farm.

Furthermore, wind turbines have to meet specific noise regulations. The norms for onshore horizontal-axis turbines are based on factors like distance from dwellings and the difference between the total noise, comprising of both the emitted and background noise, and the background noise alone. Such regulations are justified by the noise annoyance induced by wind turbines in specific conditions.  The challenge for the manufacturers is to enhance energy output, reduce noise annoyance, and ensure regulatory compliance. This necessitates careful optimizations in design, location, and operation.

Wind turbines emit mechanical and aerodynamic noise. Sources of mechanical noise are the drivetrain and the gearbox. Aerodynamic noise is caused by the interaction of the moving wind turbine blades and the air. The dominant noise sources in the audible frequency range (20 Hz to 20 kHz) are aerodynamic, mainly trailing- and leading-edge noise. Trailing-edge noise is caused by the interaction of the turbulent boundary layer on the blade with the trailing edge of the blade itself, while leading-edge noise is generated by the interaction of the turbulence in the inflow with the leading edge of the blade. This research focuses on developing a wind turbine noise model, accounting for trailing- and leading-edge noise sources, atmospheric propagation, and creating audible sound files (auralization) from a numerical workflow. Furthermore, the impact of terrain and atmospheric conditions on noise emission, propagation, and perception is investigated. Using analytical and empirical methods, the study prioritizes fast computational turn-around methods that can be used in a real environment and practical applications. The methods should be robust, delivering accurate results for various conditions and turbine geometries.

The proposed numerical model comprises the noise source model, the atmospheric propagation method, and the auralization technique. It is based on a RANS-informed Amiet's theory for the prediction of broadband trailing- and leading-edge noise combined with an engineering ray-based model for atmospheric sound propagation. It starts by dividing the blades of a 3D CAD model into segments, enabling precise airfoil shape representation. Trailing-edge boundary layer parameters, crucial for empirical wall pressure spectra models, are calculated using 2D RANS simulations. Amiet's theory is then used to predict broadband airfoil noise, while the modeling of sound propagation with atmospheric effects is considered with the ray-based engineering model Harmonoise. The resulting sound spectrum is converted into an actual audible sound (auralized) using the spectral shaping synthesis technique and the realism is enhanced with turbulence-induced amplitude fluctuation.

The methodology is applied to three horizontal-axis wind turbines in various conditions, validating and exploring the noise prediction capabilities for single turbines and a wind farm. Auralization realism is confirmed through listening tests.

The proposed methodology surpasses industry standards and enables auralization. It offers reliable noise predictions with available or simplified turbine geometries and incorporates weather effects. The methodology, while enhancing environmental noise assessment, also opens avenues for noise annoyance study and wind turbine acceptance promotion.