Wind Turbines and Radar
Figure 1: Wind turbine
Figure 1: Wind turbine
Renewable energy sources such as wind turbines will increasingly shape the appearance of the landscape. However, near radar installations they are regarded as undesirable sources of interference, and planning permission is often refused because of their proximity to a radar unit.
It is claimed that these wind turbines reduce the range of radar units. This is often difficult to imagine. The following explanation shows why:
Shadowing
A direct effect, for example through shadowing, cannot be measured because it would be too small. Just as the human eye can see through a rotating fan, radar can simply see through the circle swept by the rotor blades. Furthermore, shadowing behind an obstacle only 2 to 5 metres wide, such as a tower, can be detected directly only up to approximately 1000 metres behind that obstacle. Beyond this distance, the wavefront closes again because of the diffraction of the electromagnetic waves around the obstacle. Naturally, a small proportion of the transmitted energy is reflected and is no longer available for targets behind it, but this does not explain the reduction in range. In this case, the wind turbine is merely an obstacle like any house, tree or molehill and, when the rotor is not turning, it does not stand out at all in the mixture of signals from fixed targets.
Actual studies of shadowing were published in 2003 by Alenia-Marconi Systems following extensive computer simulations. According to these studies, the reduction in the echo signal from an aircraft flying directly in the optical shadow of a wind turbine tower is less than 2 dB. This value is completely lost in the normal fluctuation of the echo signal caused by the complex shape of its effective reflecting area.
It must also be considered that, with a horizontal antenna beamwidth of typically 1.6° and a wind turbine tower diameter of approximately 3 m, the turbine would have to be located within a radius of 3m/[2·sin(1.6°/2)] = 107 m (!) in order to cover the antenna pattern completely, that is, its half-power beamwidth. In practice, wind turbines are located much farther away and can therefore shadow only a small part of the energy of a transmitted pulse.
All affected radar installations operate in frequency bands in which electromagnetic waves have quasi-optical properties. A simple illustrative experiment can therefore be carried out using a powerful flashlight, preferably one whose light beam can be focused by moving the reflector. Aim the flashlight at a wall approximately 6 to 10 metres away and rotate the reflector to focus the beam into the smallest possible illuminated spot. Then hold a pencil at any point in the light beam (note: only the pencil, not your hand as well!) and observe the shadow cast on the wall. If no shadow is visible, observe the effect of the pencil on the brightness of the illuminated spot. You will notice
- that beyond a certain minimum distance from the flashlight, no effect on the brightness can be observed;
- that up to a certain distance between the pencil and the illuminated wall, no shadow can be observed either.
The ratio of the pencil diameter to the diameter of the light beam approximately corresponds to the ratio of a wind turbine at a distance of 5 to 10 km to the antenna-pattern width of an air traffic control radar.
Relying solely on shadowing is therefore an unacceptable simplification of the effect of a wind turbine on a radar unit and has little to do with actual conditions. A shadow can be measured only directly and at a relatively short distance behind the wind turbine. (That is, at an altitude and distance which a pilot should avoid anyway for understandable reasons.) No one has yet been able to explain why the shadowing caused by a wind turbine has a negative effect on radar, while shadowing of the same magnitude, for example by a group of trees, can be tolerated.
The attenuation behind a group of trees or a wind farm is approximately 2 decibels. However, the natural variation of an aircraft echo signal, known as fluctuation loss, can exceed 30 decibels and may therefore be more than one thousand times greater. The radar must also be able to cope with this.
Effect on Radar Range
The radar unit can distinguish echo signals from stationary objects from those of a moving object, such as an aircraft, by evaluating the Doppler frequency. The display of echo signals from fixed targets, such as buildings, trees and hills, is suppressed by a technical circuit. A wind turbine is the exception. Its rotating blades also generate a Doppler frequency, making the echo signal from a wind turbine confusingly similar to the echo signal from a hovering helicopter. This is precisely what makes the wind turbine an obstacle that can no longer be tolerated. The radar unit generates a false alarm for this echo signal. This false alarm must initially be processed in exactly the same way as a genuine alarm. Determining that it is a false alarm requires time. For digital radar units this means processing time, and for analogue radar units it requires the attention of the air traffic controller or radar operator. It is often necessary to wait for another antenna rotation before a decision can be made about the target. A single false alarm is not yet a problem, but wind turbines are usually located close together. Each wind turbine generates a false alarm, and the false alarm rate very quickly rises to levels that can no longer be tolerated. A light aircraft flying above the wind farm could therefore be missed.
Measures must be taken to reduce the false alarm rate in the radar unit. A common practice is to enter the locations of these wind turbines into a clutter map – a type of electronic map stored in the radar signal processor – and to store the measured average echo-signal amplitudes for these locations over the long term. During the next antenna rotation, only the difference between the currently measured value and the stored echo-signal amplitude is processed further. This method is highly effective when the echo signal from a stationary obstacle remains constant. Unfortunately, this is rarely the case with wind turbines because the wind direction and rotor positions are constantly changing. This means that a higher threshold must be defined for this point in the terrain, above which the radar recognises a significant target. At this point, the radar signal processor reaches its limits because all these calculations must be performed in real time. Processors are often used in which these calculations are not controlled by software programs, but are performed by extremely fast hard-wired logic. If this logic cannot decide whether a target is present or absent, the decision is passed to the next processing stage, which can compare several antenna rotations, but whose result is no longer available in real time.
All these measures for reducing the false alarm rate simultaneously reduce the probability of detection for genuine targets. Because different thresholds are used in signal processing, the false alarm rate and the probability of detection are interdependent and inversely proportional. This is the crucial point: the range of a radar unit is specified for a constant probability of detection. If that probability deteriorates, the effective range of the radar unit also decreases. This also explains why a smaller number of wind turbines near the radar unit may still have been tolerated, whereas the construction of additional turbines would significantly affect the radar range.
Possible Solutions
Figure 2: The Doppler spectra of an aircraft and a wind turbine differ considerably.
In addition to restrictive construction limitations, several technical solutions are also available. A possible improvement must reduce the time required to identify a potential false alarm. The engineering consultancy Cambridge Consultants developed a radar method that scans the relevant direction ten times per second instead of the previous rate of once every 4 seconds. It can therefore clearly identify in real time that the target in question is a wind turbine. Cambridge Consultants calls this method holographic radar™
This is possible with a radar unit that uses a phased-array antenna to perform digital beamforming. This technology, previously developed only for highly advanced military radar units, has now become affordable for civilian applications as well. Because no mechanically rotating assemblies are used, the method is maintenance-free and completely independent of weather conditions such as wind speed. For air traffic control applications, it is also a major advantage to determine the third coordinate, namely flight altitude, independently of secondary surveillance radar.
By using a modern statistical method to determine thresholds for CFAR, the previous losses in probability of detection near wind turbines can be almost completely eliminated.
Figure 2 shows the differences in Doppler frequency during successive pulse periods for a twin-engine propeller aircraft and a wind turbine observed by a radar in the I-band. The aircraft produces a Doppler frequency with a greater amplitude from the fuselage, wings and tail surfaces. This Doppler frequency results from the radial velocity of the aircraft. The propellers generate additional peaks at other Doppler frequencies, known as JEM lines (from “Jet Engine Modulation”).
The strongest echo signal from a wind turbine results from the reflecting surface of the stationary tower and has no Doppler frequency (fD = 0). The turbine blades also generate Doppler frequencies whose maximum magnitude is comparable to that of a propeller aircraft. However, every small part of the large blades has a different radial velocity. Consequently, no pulse-shaped amplitudes arise as they do from aircraft propellers; instead, there is a continuous transition from the negative to the positive maximum. These spectral differences make it possible to distinguish a wind turbine unambiguously from an aircraft.
Further trials are being conducted [1] to reduce the radar cross section of wind-turbine rotors by using stealth technology. Apart from being very expensive, this technology is unfortunately designed for only a specific frequency band and is therefore only moderately effective. It must also not make the wind turbine completely invisible, because otherwise the risk of collision in poor visibility would become too high despite the use of navigation radar. Curiously, these stealth wind turbines therefore have a corner reflector, so that navigation radar units can still detect them. This is not a contradiction, because the echo signal reduced by stealth technology fluctuates strongly and contains a Doppler frequency. By contrast, the corner reflector produces a constant echo signal in the direction of the searching radar unit, which the radar can suppress or display as required.
As another possible solution, every wind turbine could in future be equipped with a transponder for secondary surveillance radar that informs the radar unit: This echo signal comes from wind turbine xyz! Ideally, the wind direction and rotor speed could also be transmitted. The radar signal processor would treat this echo signal as a target, correlate it with the transponder signal, but could omit it from the display or prevent the signal from being sent to remote displays. The technical effort required on the wind-turbine side would be relatively small. On the radar side, only the software would need to be adapted for this solution, and newer radar units might also require increased processing power in their radar signal processors because of the now much larger number of relevant targets. Instead of secondary surveillance radar, a connection using a dedicated IP address for each wind turbine would also be possible. Administering this solution would require considerably more effort, because it would need standardisation that might first have to be recognised internationally by ICAO. Incidentally, this solution would have another advantage: the known locations of the wind turbines and their transponder signals would make the existing test transponders unnecessary, and the radar units could use the distributed wind turbines for internal calibration of their receiver channels.
[1] Stealth technology: http://www.hitek-ltd.co.uk/