Wind turbine mitigation in radar systems

The invention relates to a radar system, in particular an Air Traffic Control (ATC) radar system, and means by means of which it is better able to operate near one or more wind turbines and, in particular, a wind farm comprising a greater number of wind turbines .

BACKGROUND OF THE INVENTION [0002] The supply of wind farms to generate electricity is increasingly frequent. Wind farms can be located on the sea and on land. The presence of such wind farms, or even a single wind turbine, may be a problem for the operation of nearby radar systems that can be delivered at civil or military airports or air bases, which will be collectively referred to collectively as aerodromes.

[0003] The wind turbine echo can be relatively large compared to the echo from the desired target, which means that it is potentially very difficult, if not impossible, to recognize the desired target from the echo of the wind turbines or the wind farm.

[0004] Reference will be made to the problems caused by the wind turbine in the present description, but one of ordinary skill in the art will appreciate the fact that the problem has been exacerbated by the presence of more than one wind turbine in a typical wind farm scenario.

[0005] Attempts to solve the problem according to the state of the art of wind turbines contained ignoring the echo from the direction of the wind farm. This creates a problem because the approaching flying vehicle is not visible, which is rarely desirable.

[0006] Other attempts to solve the problem included covering the turbine blades with an echo absorbing material. It has been shown that this is not particularly effective and is quite expensive.

[0007] In practice, the problem is solved by the prohibition of erecting wind turbines in the vicinity of ATC radar systems. However, due to the increased need to generate electricity from environmentally friendly sources, this imposes an undesirable limitation on the choice of the place where wind farms can be constructed. This is especially true for densely populated countries, where there are relatively many airports that have overlapping ranges of ATC radars and where the construction of wind farms on large areas would be excluded.

[0008] US5311188A discloses a radar equipped with a coherent interference map.

[0009] WO2010 / 067057 discloses a method of suppressing interference and interference in a radar system.

[0010] There is therefore a need to suppress the impact of wind turbines so that they can be erected near airports, and thus ATC radar systems, without adversely affecting the operation of ATC radar systems.

SUMMARY OF THE INVENTION [0011] According to the invention there is provided a device and method laid out in the appended claims. Other features of the invention will be apparent from the dependent claims and the following description.

BRIEF DESCRIPTION OF THE FIGURES [0012] For a better understanding of the invention and to show how its embodiments can be used in practice, reference will be made, by way of example, to the accompanying schematic drawings in which:

Figure 1 shows the representation of the shoreline, wind farm and ATC radar installation;

Figure 2 is a graphical representation of a noise map according to one embodiment of the invention; and Figure 3 illustrates the use of radar equipment in accordance with one embodiment of the invention.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT [0013] Fig. 1 shows a typical ATC 1 radar installation located at the aerodrome (not shown). ATC 1 radar is located near shoreline 20, and at sea, at a distance, there is a wind farm 10, covering a larger number of individual wind turbines.

[0014] According to the state of the art, a high noise level is detected in the echoes coming from the wind farms 10, so the radar 1 would have to ignore the echo from the corresponding wind farm azimuthal range. In practice, however, it is unlikely that a wind farm will be built at the location.

[0015] There are two main sources of wind farm echoes: caused by a pylon or support tower; and caused by rotating blades. Pylons are static, so the use of Mobile Object Detection (MTD), Target Pointing

Moving (MTI) or Doppler Signal Processing (DP), each of which is well known in the art, can remove their effect on the entire echo. However, large structures, such as pylons, can "trickle" into MTD / MTI / DP channels, so it is not possible to rely solely on Doppler signal processing to filter out the impact of large static objects.

[0016] The rotating blades create a different problem. The dimensions of the blades are often the same size as those of some flying vehicles, and they often move at similar speeds, meaning that MTD / MTI / DP processing can not be used to account for and effectively remove them from the echo. For example, the radar cross-section (RCS) of a static tower may be in the order of 100,000 m²and the blades can be in the order of 1000 m^2.. In turn, the RCS of a large passenger aircraft vehicle may be in the order of 100 m²and the military jet fighter can be as little as 1 m². It can be seen that the level of echoes from "desired" targets is much lower than that generated by disturbances (eg turbines).

[0017] The rotation of the blade generates a significant echo, as can be seen in the results of the Doppler signal processing carried out in the MTD / MTI / DP process. Other measures are therefore required to overcome the echoes of the rotating blades.

[0018] Embodiments of the invention make use of complex signals or I and Q signals (in phase and quadrature) echoes to build a comprehensive interference map of the area around the radar. In this context, "complex" refers to the components I and Q of the echo. Distortion maps according to the state of art are known, which record power data or so-called log mod data, instead of comprehensive data (I and Q).

[0019] Fig. 2 shows the structure, in principle, of such a map. It includes a series of concentric circles, separated from each other by a distance equal to the span of the radar system. A greater number of lines radially protrude from the radar position, separated from each other by the pulse rate of the radar system. The pulse frequency is determined by the Repetition Repetition Interval (PRI).

[0020] Each one of the echoes obtained from the radar transmission can be assigned one of the cells 100 of the comprehensive interference map. It is worth noting that the cell of the comprehensive interference map is not the same as the radar span cell. The cell of the comprehensive interference map is related to both the range (distance from the radar) and the azimuth, and determines the region. Three-dimensional radar (3D) can also memorize the height information. Depending on whether 2D or 3D data is stored, a comprehensive interference map stores the spatial information for a specific 2D or 3D region.

[0021] For a long time, the echo (and in particular a pair of I / Q values) received by each of the cells is inclined towards a constant value. By storing these constant values, it is possible, when receiving a real-time echo, to detect changes in the long-term constant value. This difference in the received echo (amplitude and phase) means the target near wind turbines.

[0022] In fact, two comprehensive interference maps are used for a given radar frequency. If the radar uses two transmission frequencies (eg band E and F, which are common), we will get a total of four comprehensive interference maps - two for each frequency.

[0023] The first comprehensive interference map is created and updated in real time and "live" reflects the I and Q data received by the radar system. For example, if a flying vehicle passes over a wind farm, the values ​​stored in the cells of the comprehensive interference map from which echoes are picked up, they will differ between subsequent readings as the flying vehicle moves. If the changes in I and Q are observed in real time, it will be possible to deduce the route of the flying vehicle from changes in the corresponding values.

[0024] The second comprehensive interference map is intended to represent a static environment situation and will not be updated in real time. In other words, the values ​​stored in the second comprehensive interference map will be the one that radar should receive in the absence of targets and will represent an echo only from the constant surroundings, including, of course, the wind farm's pylons. Over time, the values ​​of the second map will stabilize on a substantially constant value. However, there must be an option to change the data stored in the second map to take into account changes in the environment. For example, a new pylon may be created that requires consideration, or a service unit may be moored near the wind farm, which must also be selectively ignored or included, as is the case for pylons.

[0025] In order to provide this flexibility, a certain relationship exists between the first and the second map. In the above-mentioned scenario, when a service unit is moored near the wind farm, the mooring movement of the unit will be returned by changing values ​​in one or more cells of the first map. However, once the unit is stationary, the respective cells will return to substantially constant values.

[0026] It is desirable that the entity can be selectively ignored by updating the data in the second map to return the new static object.

This can be done by detecting changes in the cell of the first map and then detecting the end of changes. Initial changes must have a specific threshold or be higher, and you can also set a threshold for the following stabilization. Once the value has stabilized, the new static value, which is stored in the first map, is copied to the corresponding cell in the second map. In order to ensure that changes are not made too often, and to ensure that the second map actually reflects a static situation, the data from the first map can be copied to the second map only after a certain number of readings without showing changes. This specific number can be determined as the case may be, but for starters, if two consecutive readings do not show changes from the previously changed value,

[0027] Raw or unprocessed data I and Q, which are used to populate the first map, form the basis of all further radar operations. However, in contrast to the system according to the state of art, now on the second map there are details of the static environment in which the radar operates. This data can also take the form of I and Q data and can be subtracted from raw I and Q data to obtain balanced data that only relate to moving objects or potential targets.

In a similar process, if the target detection process has identified the target, and it was also visible on the previous reading, in fact static data from the second map is copied to the first map. This ensures that any changes to the raw I and Q data that would otherwise be used to update the first map can be easily detected, which eliminates the problem of slow return to the fixed value during the movement of the target. [0029] Comprehensive interference maps are frequency dependent, thus in a typical ATC radar system that uses two frequencies in a given band, there is a pair of complex interference maps provided to each of the frequency bands, and the respective maps are used to appropriate frequencies.

[0030] This technique allows the use of interpolation and extrapolation techniques, which means that it is possible to estimate with relatively high accuracy the echoes that would be received between the two actual pulse transmissions. In practice, it allows you to fill in "gaps" in the echo by interpolating the results, and thus estimating the echoes of these gaps.

[0031] Interpolation is needed because the pulse transmission time is not synchronized with the azimuthal position of the antenna. Therefore, it is not possible to guarantee that a particular pylon will be completely displayed on a particular reading. Interpolation techniques are therefore needed to fill in the gaps in received signals.

[0032] The complex interference map shown in Fig. 2 was determined relative to fixed azimuth positions and these fixed positions may not coincide with transmissions from the radar antenna on a given readout. It is assumed that the changes between adjacent cells of the comprehensive interference map are gradual and gentle, which allows the interpolation to be used to account for the failure of comprehensive noise maps to be covered with the azimuthal position of the transmission.

[0033] The large static components identified with pylons are removed by interpolating between the I and Q values ​​of the map in each of the azimuthal cell spanning cells adjacent to the current PRI azimuth to determine the I and Q values ​​that would be expected in a given cell. range for this azimuth and this transmitted frequency, and by subtracting these I and Q values ​​from the I and Q values ​​of each of the range cells, respectively. This is a change from the interference map according to the state of the art (ie, non-complex), which basically only included power or log power values, not comprehensive values.

[0034] Therefore, when receiving an echo from the direction and range of a wind farm whose data is stored in complex interference maps, a comparison is made with a second map that represents the long-term average echo and, with this comparison, it is possible to determine whether the received echo represents a moving the target among the echoes of the wind farm pylons.

[0035] As the effect of the pylons can be eliminated at levels I and Q, it represents low loss measures to reduce the static impact of the wind farms in a manner that also prevents the passage of the pylons echo by processing the MTD / MTI / DP. due to the spectral spread caused by rotation of the antenna.

In addition to the above-mentioned complex interference maps, the radar system 5 according to this embodiment may use conventional (non-complex) high resolution residual energy interference maps that can be provided for MTI / MTD / DP channels and for normal radar channels (NR). They can generate multi-filtered output, such as in MTD processing, and such maps can be used by the filter and through the combined output data in a summation or selection process (e.g., in GOF (Fine filter) processing.

[0037] The use of a much higher azimuth and range resolution in the ATC radar system means that the echo of the turbine from the rotating blades can be suppressed without suppressing the echo of the target from between the blades.

[0038] In order to further increase the efficiency of the technique, the radar range cell is smaller than normal for ATC radar systems. According to the state of ATC radar art, the size of the span cell would be in the field of 60 by 240 meters (or in some cases larger). In most situations this is acceptable, but it does not allow to easily suppress the impact of wind turbines. A typical turbine blade has a length of 60 meters, and they are separated from each other by typically 600 meters, so the resolution of the system is not sufficient to distinguish an object of this size using spatial cells of size according to the state of the art.

[0039] The ATC systems generally need not have range cells smaller than 60 or 240 meters, because the resolution they should have is related to the objects or targets to be followed, e.g. flying vehicles. However, marine radar systems often have smaller span cells, because they require more resolution for objects that they have to track, such as impending threats, rockets or similar.

[0040] It has thus far not been necessary to equip ATC radar systems with smaller span cells, as there was no need. However, to help reduce the impact of wind farms, a smaller span cell will allow the ATC radar system to effectively view wind farm turbines and detect targets among turbines, or at least near turbines. It has been shown that a span cell with dimensions of the order of 30 meters is suitable, although smaller range cells will give better resolution.

[0041] A disadvantage of using a smaller span cell is the need for increased process power in the radar system to process a stronger received echo.

In systems according to the state of art, the process power was high and any increase in the required process power would be problematic from the technical point of view and too expensive. However, progress in the availability of powerful processing systems at a more reasonable cost makes it possible to use smaller range cells in the embodiments of the invention.

[0042] In most radar systems, there is a process called the Fixed Indicator

False Alarms (CFAR), which guarantees that the level of false alarms or detect will be constant or close to a permanent one. In practice, CFAR averages echo and scales them so that echoes above the threshold defined by the average "noise" will be classified as a significant echo so that the echo from local high density disturbances produces only a small number of domains. However, if there is a small target in such area, it may also be undesirably suppressed. This is particularly true for the wind farm scenario, where local disturbances typically consist of a substantially small background echo from between the turbines at a relatively large echo of the turbines themselves, even after suppression of this echo by various previously mentioned interference maps.

[0043] These echoes of the turbines, if included in the background level calculations, result in CFAR suppression of targets that are in or near the wind farm. This problem can be overcome using the so-called "censored CFAR", which does not take into account the echoes of the turbines in calculating the average background level. Other CFAR censored techniques are known in the art, but they have been used to prevent suppression by detecting other nearby targets, e.g. two flying vehicles flying in a formation, or two or more ships traveling in a convoy that would mutually suppress self detection.

[0044] However, the censorship used in the embodiments of the invention differs from that according to art techniques in that the number of censored cells per unit surface is selected to match the expected maximum number of turbine contaminated echoes. This results in minimizing the suppression of targets by the presence of a wind farm. In other words, when calculating CFARs for a particular cell, adjacent cells are examined in the usual way. If at least one echo from the one or more adjacent cells is significantly higher than the others, it may distort the calculated mean and negatively affect the CFAR calculation for a particular cell.

[0045] Any echo that is significantly higher than a certain threshold must be treated as an insulated value or a statistical anomaly and can be omitted for CFAR calculations. It is a concept that is the basis for what is described here as censored CFAR. The threshold above which a specific echo must be placed to qualify as an isolated value can be determined in advance or determined as the case may be.

[0046] The problem when using maps for an area with a wind farm is precisely that when the turbine blades are spinning, they are not always in positions displayed by the radar. This means that sometimes the echo is missing or small from a given blade or turbine, but if the blade is in a certain position in which it sends a large echo, it can absorb any other signals. The net result is that the echo from the wind farm area is very "pointed" and there is a large change in the echo amplitude, depending on whether the shovel is fully displayed or not. A similar phenomenon occurs in the case of helicopters, where a certain combination of position / orientation of the blade and radar display may result in bladeflash. In this context, bladeflash was used to refer to the large echo caused by a similar phenomenon in turbines.

[0047] Typically, such bladeflash phenomena are hundreds of times stronger than typical large radar echoes, but they occur rarely and last shorter than a single PRI for a typical E / F band radar. Therefore, they generate a strong response in only one coherent signal, such as that used in MTD / MTI / DP radars. It is therefore possible to compensate for bladeflash by not processing a pulse that contains bladeflash or impulse processing to remove a disturbed pulse or to convert it to an echo calculated so as to fix the pulse, i.e. it would appear that bladeflash did not occur. All processing performed to compensate bladeflash shall be carried out before using the impulse data to update the interference map before filtering in the MTD / MTI / DP process.

[0048] As it is impossible to accurately predict echoes from wind turbine blades because it is not known in advance when bladeflash occurs, it is impossible to use the echo to update the residual interference map in a simple manner using a smoothed alpha map, as it is used according to the state the art of ATC radar systems. Such a map would contain an average echo from the turbine, which would be largely distorted by the occasional occurrence of bladeflash. In order to solve this problem near wind farms, and in particular at the location of individual turbines, identified by their locally increasing long-lasting echo (or map values). map usage may change. In the vicinity of the wind farm, precise rules can be defined to further suppress the echoes of the turbines, based on the variability of the ruby ​​echoes. E.g, these rules can be relatively simple, eg if the echo is above a predetermined threshold more often than N times in the M readings window, reduce the signal strength in this location by a value based on the signal levels previously observed in the same location. You can specify more complex rules, based on statistical and time variables, echoes from this location are measured at the specified M readings. The M value can be selected based on test data or can be calculated in real time based on previous observations. You can specify more complex rules, based on statistical and time variables, echoes from this location are measured at the specified M readings. The M value can be selected based on test data or can be calculated in real time based on previous observations. You can specify more complex rules, based on statistical and time variables, echoes from this location are measured at the specified M readings. The M value can be selected based on test data or can be calculated in real time based on previous observations.

[0049] Fig. 3 shows a schematic embodiment of the equipment according to one embodiment of the invention. The system includes an antenna 200 that is identical to the antennas known in the art. The antenna is connected to an RF unit 210, which includes a transmitting device and a receiver for respectively transmitting and receiving signals. The input data to the RF 210 unit is in digital form and the output from the RF unit is in digital form. The RF unit 210 is connected to a DSP unit 220. The DSP unit 220 includes a processor that is operatively connected to the program memory 240 and the operating memory 250. The program memory stores a program for performing the method according to one embodiment of the invention as described above. Operational memory is used to store comprehensive interference maps and other data,

[0050] The DSP unit 220 is connected to a display 260 that is used by the ATC operator to monitor the targets in the usual manner.

[0051] By using one or more high resolution interference maps in combination with smaller than normal span cells, bladeflash suppression and censored CFAR, embodiments of the invention can effectively reduce the wind farms impact on the echo received by the radar.

[0052] At least some embodiments of the invention can be constructed , partly or entirely, using dedicated equipment. As used herein, terms such as "component", "module" or "unit" may include, but are not limited to, hardware such as Direct Programmable Gate Matrix (FPGA) or Dedicated

Integrated Circuit (ASIC) that performs specific tasks. Alternatively, elements of the invention may be adapted to exist on an addressable data medium. The functional elements of the invention may thus, in some embodiments, include, for example, components such as software components, dedicated software components, class components and task components, processes, functions, attributes, routines, subprograms, software code segments, drivers, embedded software, microcode , circuits, data, databases, data structures, tables, circuits and variables. Although the demonstration embodiments are described with reference to the components, modules and units discussed below, such functional elements can be combined into a smaller number of elements or divided into a larger number of elements.