Environmental Engineering Reference
In-Depth Information
Methods are available for calculating the attenuations provided by natural barriers such
as rolling terrain, which may interrupt the line of sight between the source and the receiver
[Piercy and Embleton 1979]. However, very little definitive information is available regar-
ding the effectiveness of natural barriers in the presence of strong, vertical wind gradients.
Piercy and Embleton [1970] postulate that the effectiveness of natural barriers in attenuating
noise is not reduced under conditions of upward-curving ray paths (as would apply in the
upwind direction) or under normal temperature-lapse conditions. However, under conditions
of downward-curving ray paths, as in downwind propagation or during temperature
inversions (which are common at night), the barrier attenuations may be reduced
significantly, particularly at large distances.
Predicting Noise from Multiple Wind Turbines
Methods are needed to predict noise from wind power stations made up of large
numbers of machines, as well as for a variety of configurations and operating conditions.
This section reviews the physical factors involved in making such predictions and presents
the results of calculations that illustrate the sensitivity of radiated noise to various geometric
and propagation parameters. A number of valid, pertinent, simplifying assumptions are pre-
sented. A logarithmic wind gradient is assumed, with a wind speed of 9 m/s at hub height.
Flat, homogeneous terrain, devoid of large vegetation, is also assumed. Noises from
multiple wind turbines are assumed to add together incoherently, that is, in random phase.
Noise Sources and Propagation
Reference Spectrum for a Single Wind Turbine
The most basic information needed to predict noise from a wind power station is the
noise output of a single turbine. Its noise spectrum can be predicted from knowledge of
the geometry and operating conditions of the machine [Viterna 1981; Glegg
et al
. 1987;
Grosveld 1985], or its spectrum can be measured at a reference distance. Figures 7-9 and
7-10 are examples of spectral data for HAWTs. The solid line shown in Figure 7-10 is a
hypothetical spectrum used in subsequent example calculations to represent a HAWT with
a 15-m rotor diameter and a rated power of approximately 100 kW. The example spectrum
has a decrease of 10 dB per decade in sound pressure level with increasing frequency. This
spectral shape is generally representative of the aerodynamic noise radiated by wind turbines.
However, predictions for a specific wind power station should be based, if possible, on data
for the particular types of turbines in the station.
Directivity of the Source
Measurement of aerodynamic noise for a number of large HAWTs [
e.g.
,
Kelley
et al.
1985, Hubbard and Shepherd 1982] indicate that the source directivity depends on specific
noise-generating mechanisms. For broadband noise sources, such as inflow turbulence and
interactions between the blade boundary layer and the blade trailing edge, sound pressure
level contours are approximately circular. Lower-frequency, impulsive noise, which results
when the blades interact with the tower or central column wake, radiates most strongly in
the upwind and downwind directions. Furthermore, while there is one prevailing wind
direction at most wind turbine sites, it is not uncommon for the prevailing wind vector to
Search WWH ::
Custom Search