Wind Energy Handbook. Michael Barton Graham
frequency. Self‐noise that is broadband through the auditory range and usually dominates is found to depend on the relative wind speed over the blade at between the fifth and sixth power but closer to the former. Because the tip speed ratio is normally high, this relative flow speed is effectively the local blade speed, and the noise sources are therefore mainly significant over the outer 25% of the blade and when the rotor is operating. Because of the high power dependence of the radiated sound intensity, therefore, on rotor tip speed, the ‘simplest’ way to reduce aerodynamic noise radiation from a wind turbine is to run the rotor at a lower tip speed ratio. This has always been known and is a major reason why turbines with fewer than three blades are not favoured for onshore designs, since optimum tip speed ratio increases as rotor solidity decreases. But as there are efficiency constraints also on lowering tip speed ratio, a major effort continues aimed at reducing the aerodynamic blade noise sources themselves.
3.19.2 Inflow turbulence‐induced blade noise
Inflow turbulence interacts with the blades and generates noise due to the unsteady blade forces that arise as a result. This noise source is generally found to be a less strong source than the blade's self‐induced noise, although this is not always so. The only obvious method of alleviating the inflow noise is through a control system to mitigate the unsteady loading of individual blades due to the turbulence. Distributed control capable of operation at frequencies high enough to affect the audible noise spectrum is not yet a feature of wind turbines. Inflow noise intensity is mainly only at a level to be of concern in high, gusty winds where, because there is so much wind noise from other sources in the environment, the additional turbine noise is less significant. There is a long history and a great deal of theory developed for the prediction of rotor blade noise due to turbulent inflow because of its importance in noise radiation from aircraft turbojet engines and rotorcraft. The method originated by Amiet (1975) based on prediction of unsteady blade loading taking into account compressibility was originally developed for aero‐engines. It has been further developed and is still current as a prediction method for wind turbine rotors but involves fairly extensive computational effort. Moriarty et al. (2005) have produced a simpler model based on parameterised results for standard blade geometries.
3.19.3 Self‐induced blade noise
Self‐induced aerodynamic noise arises from a number of causes: (i) interaction of the blade turbulent boundary layers with the trailing edge, (ii) noise due to locally separated flow, (iii) noise due to the vortex wake, usually due to and from a blunt trailing edge but at low Reynolds numbers can alternatively be laminar wake instability, and (iv) noise due to the blade‐tip vortex.
3.19.4 Interaction between turbulent boundary layers on the blade and the trailing edge
Interaction of turbulent eddies in the blade boundary layer with the trailing edge [i.e. (i) in the previous section] is usually regarded as the most important noise source, and efforts are continuing to design blades to minimise it. Modelling techniques have been developed to predict the aero‐acoustic radiation from this source; see Brooks et al. (1989) and Zhu et al. (2005). Two main methods of reducing the intensity have been considered:
1 Blade profile design to reduce the thickness of the suction surface boundary layer (which, being the thicker of the two, therefore has the larger turbulent eddy scales as well as the greater source layer thickness) at the trailing edge. Some progress has been achieved in reducing this boundary layer's thickness by reducing the strength of the suction pressure on the suction surface while compensating to maintain overall lift and particularly lift/drag ratio by increasing the positive pressure downstream of stagnation on the pressure surface. This method, perhaps because of the constraints involved, has yielded moderate noise reductions of up to about 2 dB. Families of low noise aerofoils have been designed, such as the DTU‐LNxxx series shown in Figure 3.78a–c; see also Wang et al. (2015).
2 Making the trailing edge serrated (see Figure 3.79) or by adding flexible ‘brushes’ to it. This concept is based on Howe's (1991) analysis of the reduction in radiation efficiency of a trailing edge as a result of making it serrated in plan. Although Howe's theory doesn't give a very good prediction of the actual sound power reduction that is achieved, nonetheless the technique has been shown to give useful noise reductions of more than 3 db. Serrations of this type appear to be possible without significantly affecting the section lift or drag. They may be part of the outer blade design or have been sometimes in the form of an add‐on to existing blades. A good description is given in Zhu et al. (2016).
3.19.5 Other blade noise sources
The remaining three sources of aerodynamic blade self‐noise are usually less significant than i) above involving interaction between the turbulent boundary layers and a sharp trailing edge:
ii) Noise due to locally separated flow is more usual from the inner blade where intensities are limited by low relative velocities. Significant separation is unusual on the outer blade under low to moderate wind conditions for which blade noise may be a concern.
iii) Noise due vortex shedding from the trailing edge of the blade can be of concern. It should be considered only if the outer blade aerofoil section has a particularly blunt trailing edge. If it occurs, it can be more irritating than purely broadband noise because of the strong tonal content.
iv) Tip vortex noise does not seem to be particularly well understood but can be minimised by a well‐designed tip with appropriate rounding.
Figure 3.78 Low noise aerofoil family: (a) DTU‐LN1xx, (b) DTU‐LN2xx, and (c) DTU‐LN3xx.
Source: From Zhu, Shen, and Soerensen (2016).
Figure 3.79 Diagram of serrated trailing edge for reduction of TE noise.
3.19.6 Summary
The effect of noise on adjacent populations is the reason for concern about noise. Mainly in the case of wind turbines this concerns human populations. However, underwater propagation of sound from offshore wind turbines may need to be kept in mind with respect to marine animals but is very unlikely to be as significant as it is for tidal stream turbines. Noise effect on adjacent populations is normally defined by a geographical noise footprint based on contours of perceived noise (PNdB). In drawing these up account has to be taken of the different efficiencies of propagation of noise at different frequencies, in particular that low frequency noise travels much farther than high frequencies, and of the non‐uniform sensitivity of the ear over the audible frequency range. Because of the major issues surrounding aircraft noise and the siting of runways, there is a great deal of research published on this.
This Section 3.19 on aerodynamic noise has only attempted to summarise the main issues and research into the subject where it concerns noise arising from wind turbine rotor blades. In practice this is the most important source of noise from a wind turbine, and because noise has become one of the major planning constraints for siting wind turbines, it is likely that the industry will continue considerable effort into the development