The magnitude, time history and direction of the wind are the most important parameters for the prediction of the expected energy yield – i.e. the quantitative
4.4 Wind measurement and evaluation 156
assessment of a site - as well as the decision which available wind turbine is the most suitable one.
Since the wind turbine power is approx. proportional to the cube of the wind speed (until the rated wind speed is reached) the wind regime on a site has to be determined as exact as possible. An error of 10% in the wind speed measurements may produce an error in the determined power output of up to 33%. An exact knowledge of the wind regime is furthermore important for the determination of the mechanical loads and stress.
There are high requirements on the wind measuring devices, the sensors and the instrumentation for data recording. Apart from the accuracy of the wind speed sensor the devices have to be extremely robust in order to record data mainte- nance-free for long periods.
Moreover, faulty measurements due to wrong installation and sensor icing have to be prevented. In general, mechanical wind speed sensors like the cup anemome- ter are well proven and reliable sensors. Their limits of application and possible sources of errors are mostly known, so among other, they should be calibrated before and after a measuring campaign.
Sensors without moving parts, like e.g. the ultrasonic anemometer are up to now seldom applied for long-term measurements since firstly they are more expensive than the mechanical sensors, and secondly they are more susceptible to faults because they are generally more complex devices. The ultrasonic anemo- meter may measure, independent from the current wind direction, instantaneously all the three components vx, vy und vz of the wind vector
) ( ) ( ) ( )
( 2 2 2
Vector t v t v t v t
v x y z . (4.23)
Since a wind turbine is able to extract power only from horizontal wind speed component vhoriz which is perpendicular to the rotor swept
) ( ) ( )
(t v2 t v2 t
vhoriz x y , (4.24)
a vertical component vz, e.g. due to a slope (cf. Fig. 4-34), is of no use. The cup anemometer measures directly the horizontal wind speed component vhoriz but can- not resolve the vertical component vz .
Propeller anemometers (also called windmill anemometers) measure as well only the horizontal wind speed component vhoriz and have an integrated wind vane.
It is favourable that the wind direction is measured with the same sensor unit, whereas using a cup anemometer it has to be determined by a separate sensor. But unfortunately, the integrated wind vane of the propeller anemometer causes a severe problem: the propeller is wiggling in the wind which distorts the wind measurement.
4.4.1 Cup anemometer
International standards following IEC [4] and IEA [3] refer to the cup anemometer as the most suitable sensor type for wind speed measurements, Fig. 2-22 and 4-40.
The cup anemometer is a small drag driven wind mill with a vertical axis of rota- tion. Cup-like drag bodies are fixed with a lever to the vertical shaft. Instead of hemispherical cups there are more and more conical shells applied since they have a more distinct edge for defined flow separation.
Anemometers produce either an analogue or a digital signal output. The ana- logue signal is the voltage, which is proportional to the rotational speed and the wind speed (cf. equation 2-12), produced by the rotation of a small generator. The digital signal, a defined number of impulses per revolution produced by reed con- tacts or an incremental encoder, is counted over in a certain interval and offers this way a measure proportional to the wind speed.
High quality precision bearing for minimization of mechanical losses
Long shaft to minimize the influence of flow disturbances of the rotor, induced by the nacelle
Thoroughly designed rotor geometry for well defined vertical sensibility and dynamic response behaviour
Small and complete sym- metric housing without sticking out parts, with a soft outline to minimize flow disturbances
Fig. 4-40 Cup anemometer [3]
A special attribute of the cup anemometer is the so-called response length (also called distance constant, [13]) which results from signal delay occurring at rapid changes of the wind speed due to the inertia of the rotating cup star. If a virtual wind speed increases with a step function from v0 to v0 + 'v , the signal of the anemometer responds with an exponential curve (i.e. 1st order time delay).
4.4 Wind measurement and evaluation 158
There are several methods to determine the response length [13]. In the wind tunnel method, a constant wind speed is adjusted, and the acceleration of the anemometer from standstill to load-free idling is measured. The rotational speed increases rapidly, therefore, the required measurement instrumentation has to work fast (i.e. high sampling rate) and with a high resolution. Another method to determine the response length is the comparison of the cup anemometer with a high-resolution ultrasonic anemometer.
4.4.2 Ultrasonic anemometer
Ultrasonic anemometers, Fig. 4-41, were developed for the investigation of the turbulent field in the surface boundary layer. There are up to three pairs of sonotrodes (speaker-microphone combination) arranged in a way that the three spatial components of the flow are resolved, equation (4.23). The ultrasonic im- pulses, emitted at a rate of 100 Hz, travel over the distance s between speaker and microphone with the speed of sound c. The wind speed component parallel to the sound travelling direction of the sonotrode superimposes the emitted signal and leads to different run times for the way with (t1) and against the wind (t2), which is called the Doppler Effect.
v c t s
1 and
v c t s
2 (4.25)
Rearranging the equations yields the wind speed component v parallel to the sound travelling direction of this sonotrode, simply given by the travelling distance and the two run times.
ááạ
ăă ã
©
§
2 1
1 1
2 t t
v s (4.26)
It is favourable that the wind speed component is determined independently from the speed of sound which varies with air density and humidity. As mentioned above, the combination of three sonotrodes gives a sensor which allows the instan- taneous measurement of all three wind speed components, Fig. 4-41.
A problem inherent of the ultrasonic anemometer is the deflection of the flow by the sonotrodes heads, and the sensor arms as well. Therefore, the measurement in the direction of the arms is inaccurate. This is a potential source of measure- ment uncertainties of long-term measurements with theses sensors. Moreover, there are effects from the aging of the piezo-electric sonotrodes. At least, the sensor behaviour is temperature dependent.
Fig. 4-41 Ultrasonic anemometer
4.4.3 SODAR
The abbreviation SODAR stands for SOnic Detecting And Ranging. It is a remote sensing method for the exploration of the atmospheric boundary layer. The principle is similar to this of Radar and Sonar measurements. The SODAR technique is mainly applied for recording the vertical wind speed profile. The SODAR device emits small sharply bundled sound signals in the audible range. These are reflected at the boundaries between atmospheric layers of different refractive numbers which comes from differences in temperature and humidity. These boundaries between atmospheric layers move with the local speed of the wind.
The frequency of the back-scattered signal is shifted with respect to the emitted signal due to the Doppler Effect. If the sender and the receptor are at the same location the system is called a monostatic SODAR. The measured Doppler shift is proportional to the wind speed component parallel to the beam direction of the sending antenna.
If several antennas or senders are applied which point in different directions the three-dimensional vector of the wind can be measured. The senders are mostly arranged in a way to receive a phase-shifted signal (phased array) which allows simulating the behaviour of several antennas pointing into different directions.
Using the different run times between sender and scattering volume receptor parallel measurements of the wind speed in different heights are possible to determine the vertical wind speed profile.
The SODAR measures within a certain air volume, in contrast to the cup anemo- meter which is limited to the measurement at one single point of the wind field.
4.4 Wind measurement and evaluation 160
In order to measure in heights between 20 and 150 m above ground, so-called Mini-SODAR systems are applied. They operate in the frequency range between 4 and 6 kHz and reach a resolution of height between 5 and 10 m.
The biggest advantage of the SODAR systems is that they allow measurements of wind speed and wind profiles in large heights where a measuring mast would be too expensive. Especially in very complex terrain, e.g. where a forest is found in the main wind direction, the SODAR systems allow a comparison of measured and simulated wind profile. But it has to be considered that under stable atmos- pheric conditions there is only a small scattering, so there will be no valid measur- ing signals. Moreover, the signal-to-noise ratio decreases with increasing height, so the allocation of the signals to the related height above ground becomes more and more difficult. Thus, the installation and adjustment of the SODAR systems as well as the evaluation and interpretation of the SODAR measurements require extensive experience and routine.
Fig. 4-42 Scheme of a SODAR measurement
Since the SODAR has a significantly higher power consumption than a cup anemometer the power supply may by a problem at remote sites. Moreover, the measurement quality is strongly dependent on measurement disturbances by obstacles close to the antennas which may disturb and reflect the signals. Since the emitted signals are in the audible frequency range the measurements are sensitive to background noise e.g. from cowbell, croaking frogs, the roar of the surf and more. It should be remembered that the wind profile measured with a SODAR system in a shorter measuring period is representative only for the specific prevail- ing atmospheric stratification. All in all, using a SODAR the determination of the mean wind speed with a desired precision is difficult, therefore it should be
applied only in combination with conventional wind sensors installed on a measuring mast for additional investigations of the wind regime.