COMPUTER-IMPLEMENTED METHOD FOR CALIBRATING WIND TURBINES IN A WIND FARM

Abstract

A computer-implemented method for calibrating a nacelle position of at least one wind turbine in a wind farm is provided having a plurality of spatially distributed wind turbines, in which a first calibration step of recalibrating a precalibrated nacelle position of at least one wind turbine is followed by a second calibration step, in which a wind direction sector specific further recalibrating of the nacelle position of the at least one wind turbine to be calibrated is performed by applying a wind direction sector specific correction value for an identified wind direction sector to the previously obtained calibrated nacelle position.

Claims

1. A method for calibrating a nacelle position of at least one wind turbine in a wind farm having a plurality of spatially distributed wind turbines, the method being a computer-implemented method, wherein the method comprises: first calibration step A: recalibrating a precalibrated nacelle position of at least one wind turbine to be calibrated based on an evaluation of computed power ratios and predetermined power ratios between at least one downstream wind turbine and at least one upstream wind turbine for different wind directions, wherein the at least one upstream wind turbine having a wake effect on the at least one downstream wind turbine, thereby obtaining a calibrated nacelle position of the at least one wind turbine to be calibrated; and second calibration step B: wind direction sector specific further recalibrating of the nacelle position of the at least one wind turbine to be calibrated by applying a wind direction sector specific correction value for an identified wind direction sector to the previously obtained calibrated nacelle position.

2. The method according to claim 1, wherein the method, for implementing the first calibration step A, comprises: determining at least one upstream wind turbine and at least one downstream wind turbine of the wind farm, the at least one upstream wind turbine having a wake effect on the at least one downstream wind turbine for different wind directions of a wind impinging the at least one upstream wind turbine, computing at least one computed set of power ratios between the at least one downstream wind turbine and the at least one upstream wind turbine for different wind directions, identifying at least one wind direction mismatch by comparing at least one power ratio minimum and/or at least one power ratio maximum in the at least one computed set of power ratios with at least one power ratio minimum and/or at least one power ratio maximum in at least one predetermined set of power ratios for the at least one upstream wind turbine and the at least one downstream wind turbine, and recalibrating a precalibrated nacelle position of at least one wind turbine to be calibrated, the at least one wind turbine to be calibrated being selected from the at least one downstream wind turbine and the at least one upstream wind turbine, wherein the recalibration is based on the identified at least one wind direction mismatch to obtain a calibrated nacelle position of the at least one wind turbine to be calibrated; and wherein the method, for implementing the second calibration step B, implements the wind direction sector specific further recalibrating of the nacelle position of the at least one wind turbine to be calibrated with the further steps of: identifying a wind direction sector with a remaining wind direction mismatch after the previous recalibration, the remaining wind direction mismatch is identified for a specific wind direction sector of a plurality of wind direction sectors by comparing a power ratio at a given wind direction sector of the at least one computed set of power ratios with at least one power ratio at the given wind direction sector in the at least one predetermined set of power ratios, and applying a wind direction sector specific correction value for the identified wind direction sector to the previously obtained calibrated nacelle position of the at least one wind turbine to be calibrated for the identified wind direction sector, thereby obtaining a wind direction sector specific nacelle position of the at least one wind turbine to be calibrated.

3. The method according to claim 1, wherein the at least one computed set of power ratios and/or the at least one predetermined set of power ratios is determined as a function of a range of the different wind directions.

4. The method according to claim 1, wherein the at least one predetermined set of power ratios is based on a model or simulation, an upfront simulation.

5. The method according to claim 1, wherein several wind direction mismatches are identified.

6. The method according to claim 5, wherein a measure of central tendency is calculated from the several wind direction mismatches and the precalibrated nacelle position of the at least one wind turbine to be calibrated is recalibrated based on the computed measure of central tendency to obtain the calibrated nacelle position of the at least one wind turbine to be calibrated.

7. The method according to claim 6, wherein a median value or mean value is computed as measure of central tendency.

8. The method according to claim 1, wherein, in the step of identifying a wind direction sector with a remaining wind direction mismatch after the previous recalibration, the at least one power ratio for the given wind direction sector is located between a power ratio minimum and a power ratio maximum of the computed set of power ratios, and an interpolation is performed for identifying the further wind direction mismatch.

9. The method according to claim 1, wherein further comprises precalibrating the nacelle position of the at least one wind turbine to be calibrated based on a precalibrated nacelle position of at least one precalibrated wind turbine of the wind farm.

10. The method according to claim 9, the precalibration of the nacelle position of the at least one wind turbine to be calibrated having the steps of: determining a raw nacelle position of the at least one wind turbine to be calibrated, determining a precalibrated nacelle position of at least one precalibrated wind turbine of the wind farm, computing a correction for the raw nacelle position of the at least one wind turbine to be calibrated based on the precalibrated nacelle position of the at least one precalibrated wind turbine, and applying the computed correction to the raw nacelle position of the at least one wind turbine to be calibrated such that the precalibrated nacelle position of the at least one wind turbine to be calibrated is obtained.

11. The method according to claim 1, wherein the at least one power ratio maximum and/or the at least one power ratio minimum in the at least one computed set of power ratios and/or the at least one predetermined set of power ratios is of a magnitude of at least 0.1, or of at least 0.2, from an adjacent minimum and/or maximum of the at least one computed set of power ratios and/or the at least one predetermined set of power ratios.

12. A system for carrying out the method of claim 1.

13. The system according to claim 12, wherein the system is a wind farm having a plurality of spatially distributed wind turbines.

14. A computer program product, comprising a computer readable hardware storage device having computer readable program code stored therein, said program code executable by a processor of a computer system to implement a method comprising instructions, which, when the computer program is executed by a computer, cause the computer to carry out the method of claim 1.

15. A computer-readable medium comprising instructions which, when executed by a computer, cause the computer to carry out the method of claim 1.

Description

BRIEF DESCRIPTION

[0058] Some of the embodiments will be described in detail, with reference to the following figures, wherein like designations denote like members, wherein:

[0059] FIG. 1 shows a side view on a wind turbine for use in a wind farm;

[0060] FIG. 2 shows a schematic illustration of a wind farm with two wind turbines, a downstream wind turbine experiencing wake from an upstream wind turbine;

[0061] FIG. 3 shows a schematic illustration of a wind farm with the upstream wind turbine executing a wake steering control operation;

[0062] FIG. 4 shows a schematic illustration of the wind farm of FIG. 2 and a power ratio graph;

[0063] FIG. 5 shows a schematic illustration of a further wind farm having five wind turbines;

[0064] FIG. 6 shows power ratio graphs of three wind turbines of the further wind farm of FIG. 5 being in wake interaction with each other;

[0065] FIG. 7 shows a schematic illustration of a method or calibrating a nacelle position of a wind turbine in a wind farm; and

[0066] FIG. 8 shows a schematic illustration of a wind farm having a system or means for carrying out the method of FIG. 7.

DETAILED DESCRIPTION

[0067] FIG. 1 shows a wind turbine 1. The wind turbine 1 comprises a supporting tower 6 and a nacelle 7, wherein the nacelle 7 is attached to the supporting tower 6. The wind turbine 1 further comprises a rotor 8 and a number of wind turbine blades 9 attached thereto. The wind turbine 1 also comprises a yaw system (not shown) for yawing the nacelle 7 relative to the supporting tower 6. By the yawing of the nacelle 7, the rotor 8 and wind turbine blades 9 may be positioned for best performance given the wind directions of a current wind. Also, the wind turbine 1 may comprise an anemometer (not shown) for measuring the wind direction of the wind.

[0068] FIG. 2 shows a schematic illustration of a wind farm 10 with two wind turbines 1, 2 having the design as described with reference to FIG. 1. Although the wind farm 10 of this example is only shown with two wind turbines 1, 2, the number of wind turbines 1, 2 of the wind farm 10 may also be much higher, e.g., at least 5 (see FIG. 4, for example), 10, 20 or more.

[0069] Given the wind 11 indicated with its direction by the arrows illustrating the wind 11, the wind turbine 1 is located upstream of the wind turbine 2. Accordingly, the wind turbine 1 may also be referred to as the upstream wind turbine 1 as it is in front of the wind 11. The wind turbine 2, on the other hand, may be referred to as a downstream wind turbine 2 because it is located behind the upstream wind turbine 1 given the current wind direction of the wind 11 represented by the arrows. This designation of upstream and downstream of the wind turbines 1, 2 may change with a shift of direction of the wind 11. In particular, other wind turbines not shown in FIG. 2 may then become upstream and/or downstream of the wind 11 located wind turbines 1, 2.

[0070] The upstream wind turbine 1 generates electricity based on the energy of the wind 11 impinging on the upstream wind turbine 1 and causes a wake 12 impinging on the downstream wind turbine 2. In the situation shown in FIG. 2, the downstream wind turbine 2 is fully affected by the wake 12. The wake 12 results in a significantly decreased energy production of the downstream wind turbine 2 compared to the upstream wind turbine 1.

[0071] It may be possible to deflect the wake 12 from the downstream wind turbine 2 by a wake steering control operation executed by the upstream wind turbine 1. Such situation is shown in FIG. 3.

[0072] In the situation shown in FIG. 3, the yaw system of the upstream wind turbine 1 is misaligned with respect to an optimal positioning into the wind 11 such that the wake 12 may be entirely deflected from the downstream wind turbine 2. Thereby, it may be possible to increase the overall annual energy production of the wind farm 10. However, for best performance, the wake steering control operation requires precise nacelle position data. In other words, a calibration of the nacelle position of the wind turbine 1 must be as precise as possible to increase the performance as much as possible.

[0073] The calibration method, as explained later with reference to FIG. 7 in more detail, uses the power ratios between the downstream wind turbine 2 and the upstream wind turbine 1 for calibrating any one or both of the wind turbines 1, 2 of the wind farm 10 as illustrated in a graph in FIG. 4. FIG. 4 shows the wind farm 10 of FIG. 3 with the wake 12 from the upstream wind turbine 1 impinging on the downstream wind turbine 2, or in other words, impacting the downstream wind turbine 2. The wind 11 in this example has a wind direction of 110? when measured in clockwise direction relative to a north direction defined as 0? wind direction as seen in FIG. 4. The calibration performed by the methods presented below may therefore also be referred to as a north calibration.

[0074] A method for calibrating the wind turbine 1 in the wind farm 10 is exemplary explained in the following. This method is merely an example based on only two wind turbines 1, 2 of the wind farm 10. However, as explained above, there may be a much higher number of wind turbines and wake effects to consider. Also, more than only one power ratio minimum or power ratio maximum as explained in the following may be considered for recalibrating a precalibrated nacelle position of a wind turbine 1, 2 in a wind farm 10.

[0075] When two turbines 1, 2 as shown in FIG. 4 are aligned into the prevailing wind, the wake 12 generated by the upstream or front wind turbine 1 causes a loss of power in the downstream or rear wind turbine 2. In embodiments, the method proposed in the following is based on detecting these situations between nearby wind turbines 1, 2 using operational wind farm data to detect the power losses caused by wakes 11 and compare it with wind farm 10 model predictions.

[0076] The power ratio graph beneath the schematic illustration of the wind farm 10 shows two curves corresponding to sets of power ratios as functions of the wind direction from 0? to 360?. The power ratios Pw.sub.2/Pw.sub.1 are the ratio of the power Pw.sub.2 generated by the downstream wind turbine 2 to the power Pw.sub.1 generated by the upstream wind turbine 1. The power ratios may alternatively be the other way around, i.e., Pw.sub.1/Pw.sub.2.

[0077] One of the curves shows a set of power ratios Pw.sub.2/Pw.sub.1 from field data and the other one shows a set of power ratios Pw.sub.2/Pw.sub.1 expected by a model or simulation. The field data power ratios may also be referred to as the computed set of power ratios because these are being computed based on the power that is generated in the field or, in other terms, in operation of the wind turbines 1, 2. On the other hand, the simulation or model power ratios may be referred to as predetermined power ratios because these are based on an upfront simulation of a model of the wind farm 10.

[0078] The computed set of power ratios, as may be seen from its respective curve, shows a drop or, in other words, power ratio minimum at a wind direction of approximately 140?. However, the predetermined set of power ratios, as may be seen from its respective curve, shows the power ratio minimum at a wind direction of approximately 110?. This power ratio minimum occurs due to the wake 12 impinging on the downstream wind turbine 2 when the wind 11 impinges on the upstream wind turbine 1 in a direction of 110? as seen in the schematic illustration of the wind farm 10 in FIG. 4.

[0079] The field data is false with respect to its wind direction in this case and thereby the computed set of power ratios indicates the respective power ratio minimum falsely at 140? instead of 110?, which is assumed to be correct based on the simulation. Accordingly, there is a wind direction mismatch of ?30?. This means that the nacelle position of the upstream wind turbine 1 is wrong by ?30? and needs to be corrected by ?30?. A recalibration of a precalibrated nacelle position of the wind turbine 1 may be undertaken by adding the wind direction mismatch of ?30? to the precalibrated nacelle position and thereby obtain a calibrated nacelle position.

[0080] FIGS. 5 and 6 show an example with more wind turbines 1, 2, 3, 4, 5 of the wind farm 10 being spatially distributed from one another in a terrain, which may be a complex terrain. In embodiments, the method described above with reference to FIG. 4 may also be applied to this wind farm 10.

[0081] FIG. 5 indicates wind 11.sub.2, 11.sub.3 at two different exemplary wind directions, namely 229,7? for wind 11.sub.3 and 266,6? for wind 11.sub.2. It is known from simulation that these winds 11.sub.2, 11.sub.3 will cause wakes 12.sub.2 and 12.sub.3 at the downwardly of the wind turbine 1 located wind turbines 2 and 3.

[0082] FIG. 6 shows the predetermined (model) set of power ratios and the computed (field data) set of power ratios as curves in two separate graphs similar to the illustration of the graph of FIG. 4. The top graph shows the ratio of power generated by the wind turbine 2 to the power generated by the wind turbine 1 as a function of the wind direction measured at the wind turbine 1 (WindDir.sub.1). The bottom graph shows the ratio of power generated by the wind turbine 3 to the power generated by the wind turbine 1 as function of the wind direction measured at the wind turbine 1 (WindDir.sub.1). The computed set of power ratios in both cases has a limited scale of wind directions from approximately 200? to 310? because these are the wind directions of wind 11 occurring in the wind farm 10 of FIG. 5 and thereby measurable in terms of their power generation.

[0083] FIG. 7 schematically shows the steps 101, 102, 103, 104 of a computer-implemented method 100 for calibrating a nacelle position in any one or all of the wind turbines 1, 2, 3, 4, 5 of the wind farm 10 in FIG. 5. In embodiments, the method is based on the power ratio principle as explained with reference to FIG. 4, includes the elements of the calibration method explained with reference to FIG. 4 and will be explained in the following with respect to the wind turbines 1, 2, 3 of FIG. 5 and FIG. 6.

[0084] In a first step 101 of the computer-implemented method 100, at least one of the wind turbines 1, 2, 3, 4, 5 is precalibrated. This may be done by trigonometry, using a compass, using data from the wind farm 10 to detect a specific wake 12 or similar. This precalibration does not need to be very accurate.

[0085] In a second step 102, a rough calibration of the other wind turbines 1, 2, 3, 4, 5 is performed based on the at least one precalibrated wind turbine 1, 2, 3, 4, 5 of the wind farm 10. In this second step 102, raw nacelle positions of the wind turbines 1, 2, 3, 4, 5 to be calibrated are determined. Then, a correction for each of the determined raw nacelle positions of the wind turbines 1, 2, 3, 4, 5 to be calibrated based on the precalibrated nacelle position of the at least one precalibrated wind turbine 1, 2, 3, 4, 5 is calculated. And, finally, in the second step 102 the computed corrections for each one of the wind turbines 1, 2, 3, 4, 5 to be calibrated are applied, e.g., added or subtracted, to their raw nacelle positions such that the precalibrated nacelle position of the wind turbines 1, 2, 3, 4, 5 to be calibrated are obtained. The rough calibration thereby delivers a roughly calibrated or precalibrated nacelle position for each of the wind turbines 1, 2, 3, 4, 5 in the wind farm 10 if only one or some of the wind turbines 1, 2, 3, 4, 5 are precalibrated.

[0086] Once each wind turbine 1, 2, 3, 4, 5 has the precalibration performed, which is supposed to be close to the real nacelle position of each one of the wind turbines 1, 2, 3, 4, 5, a more accurate calibration will be calculated using the power ratio technique explained before. Each wind turbine 1, 2, 3, 4, 5 will be compared with other closer wind turbines 1, 2, 3, 4, 5, and all the relevant maxima and minima of the power ratio curve in function of wind direction will be compared with the one predicted by a model of the wind farm 10. FIGS. 5 and 6 show an example where the wind turbine 1 is being calibrated using wind turbines 2, 3. In this example, six different corrections are needed to match the field data and the model. The more accurate calibration is carried out in the third step 103 of the method 100.

[0087] The more accurate calibration performed in step 103 may also be referred to as a thin calibration. For the thin calibration, upstream wind turbines 1, 2, 3, 4, 5 and downstream wind turbines 1, 2, 3, 4, 5 of the wind farm 10 for different wind directions of a wind 11 impinging the upstream wind turbines 1, 2, 3, 4, 5 are determined. For the sake of simplicity, as seen in FIG. 5, only the wind turbine 1 will be considered as upstream wind turbine 1 and the wind turbines 2, 3 will be considered as downstream wind turbines 2, 3 to explain the method 100 by example in the following.

[0088] Further, for the thin calibration, the computed (field data) sets of power ratios seen in FIG. 6 are computed for the determined wind turbines 1, 2, 3. In this exemplary case, 1-min (or 10-min) field data samples are binned to create the continuous plot. The computed (field data) set of power ratios in the top graph illustrates the power ratio Pw.sub.2/Pw.sub.1 showing two power ratio minima and one power ratio maximum having a magnitude of at least 0.2 in terms of the power ratio Pw.sub.2/Pw.sub.1 compared to their closest minimum or maximum in the curve or, in other words, computed set of power ratios. In other words, the prominence of the maxima and minima is used, which is the distance between the highest point in a maxima (or lowest in a minima) and the surrounding local extrema. For instance, in FIG. 4 at 110? the prominence is around 0.4 (1-0.6).

[0089] Further, in the thin calibration of the third step 103, these power ratio minima and maxima are compared with the power ratio minima and maxima of the predetermined (model) set of power ratios being located in terms of wind direction closest to the wind directions of the power ratio maxima and minima of the computed set of power ratio. A wind direction mismatch between each respective one of the power ratio minima and maxima is identified. This is done by subtracting the respective wind directions of the power ratio minima and maxima of the computed set of power ratios with the respective wind directions of the corresponding power ratio minima and maxima of the predetermined set of power ratios.

[0090] As seen in the top graph of FIG. 6, this delivers a wind direction mismatch of ?1? for the power ratio maximum, a wind direction mismatch of ?3? for the first power ratio minimum and a wind direction mismatch of ?1? for the second power ratio minimum.

[0091] Similarly, in the bottom graph of FIG. 6, such identification of wind direction mismatches delivers wind direction mismatches of ?1?, ?2? and ?2? for the relevant power ratio minima and maxima.

[0092] The identified wind direction mismatches or, in other terms, their values may now be used for recalibrating the precalibrated nacelle position of the wind turbines 1, 2, 3 in the third step 103. For example, a median of the wind direction mismatches identified as described above may be computed as a value for the thin calibration and the median may be added to the roughly calibrated or precalibrated nacelle position of the upstream wind turbine 1 acquired by the rough calibration in step 102. Thereby, a recalibrated or thin calibrated nacelle position of the upstream wind turbine 1 is acquired in step 103.

[0093] So far, the discussed embodiment was focusing on the first calibration step A as defined in the initial part of this document.

[0094] The recalibrated or thin calibrated nacelle position of the upstream wind turbine 1 is more precise than the precalibrated or roughly calibrated nacelle position but it still may be optimized. For this purpose, a wind direction dependent (or wind direction sector dependent) nacelle position of the wind turbine 1 may be calibrated by a wind direction dependent calibration in step 104.

[0095] This step 104 corresponds to the second calibration step B as defined in the initial part of this document.

[0096] In step 104, the wind turbine 1 is calibrated for a given wind direction by identifying a further or remaining wind direction mismatch by comparing a power ratio at the given wind direction (or wind direction sector) in the computed set of power ratios with a power ratio at the given wind direction (or wind direction sector) in the predetermined set of power ratio. This further wind direction mismatch is then added to the calibrated nacelle position of the wind turbine 1 to be calibrated, thereby obtaining the wind direction dependent (or wind direction sector dependent) nacelle position of the wind turbine 1.

[0097] For example, the wake 12 generated by wind turbine 1 is expected to affect wind turbine 2 at a wind direction of 229.7?, and wind turbine 3 at a wind direction of 266.6?. It may now happen that wind turbine 2 found the wake 12 at 228?, and wind turbine 3 at 268? due to terrain effects. So, a wind direction sector with a remaining wind direction mismatch after the previous recalibration is identified. In order to artificially match the wind farm 10 data with the model expectations, an additional correction in function of wind direction will be needed with +1.7? applied at 228?, and ?1.4? at 268? to get these wind direction dependent nacelle positions. Thus, a wind direction sector specific correction value is applied. Accordingly, the (absolute) wind direction based on the measured wind direction itself is corrected. An aspect of this final correction is to fix the fact that the wake is not where it is expected that it should be.

[0098] Thus, the following logic is gained:

CalibratedNacellePosition=RawNacellePosition+InitialCorrection+CorrectionBySector,

wherein the RawNacellePosition and the InitialCorrection are determined for the whole range of wind direction angles (360?), and wherein CorrectionBySector is a correction value for a wind direction sector, thereby having as a result a CalibratedNacellePosition which is also specific on the wind direction sector.

[0099] Thus, the CalibratedNacellePosition could be considered a vector of values, each vector element defining a calibration value for a given wind direction sector. The dimension of the vector is the number of wind direction sectors that are evaluated.

[0100] FIG. 8 schematically shows a system 10 comprising a controller 20 for carrying out the method 100 according to FIG. 7.

[0101] The controller 20 in this example is located in one of the wind turbines 1, 2, 3, 4, 5 and has a computer 21 or computing unit for executing the method 100. In embodiments, the method 100 is carried out when a computer program stored in a computer-readable storage medium 22 of the controller 20 is executed. The computer-readable storage medium 22 may further comprise the simulation data of the predetermined set of power ratios.

[0102] The controller 20 may alternatively be located at a distance from the wind turbines 1, 2, 3, 4, 5 within the wind farm 10 or away from the wind farm 10. However, the controller may be physically or wirelessly connected to the wind turbines 1, 2, 3, 4, 5 as seen by the connections 232, 233, 234, 235.

[0103] The controller 20 may by the connections 232, 233, 234, 235 receive the respective raw nacelle position, values of generated power of the other wind turbines 2, 3, 4, 5 in the field or operation and if necessary other data as described above for recalibrating the wind turbines 2, 3, 4, 5.

[0104] Alternatively, each one of the other wind turbines 2, 3, 4, 5 may have such a controller 20 configured as described above for recalibrating its own nacelle position according to the method 100.

[0105] Already installed controllers 20 in wind turbines 1, 2, 3, 4, 5 may be used, which may also serve other functions of the wind turbines 1, 2, 3, 4, 5, such as controlling the yaw system for the wake steering control operations, may be used.

[0106] Table 1 shows an example of results of the method 100 as explained above but carried out for all wind turbines 1, 2, 3, 4, 5 of a further wind farm 10. Note that the numbers of the nacelle positions are chosen randomly and table 1 is intended to merely explain the method 100 and its results by way of example in more detail.

TABLE-US-00001 TABLE 1 results of the calibration method 100 carried out for an exemplary wind farm 10 having five wind turbines (WT) 1, 2, 3, 4, 5. step nacelle position WT 1 WT 2 WT 3 WT 4 WT 5 101 raw 125 69 340 2 102 102 precalibrated ?23 33 ?238 100 0 103 recalibration 1 ?2 0 1 2 recalibrated ?22 31 ?238 101 2 104 wind direction dependent

[0107] In step 101, the raw nacelle position of each one of the wind turbines 1, 2, 3, 4 has been determined.

[0108] The wind turbine 5 has been precalibrated, and therefore the precalibration in step 102 is carried out based on the precalibrated nacelle position of wind turbine 5. Each raw nacelle position of the wind turbines 1, 2, 3, 4 is subtracted from the precalibrated nacelle position of wind turbine 5 of 1020 noted in the row raw nacelle position in table 1.

[0109] Then, for the recalibration step 103, the above-described technique of identifying the wind direction mismatches based on the power ratio minima and maxima in the computed and predetermined set of power ratios is carried out for all wake interactions of the wind turbines 1, 2, 3, 4, 5. The median of the identified wind direction mismatches is computed as noted in the row of step 103 of table 1 for each one of the wind turbines 1, 2, 3, 4, 5.

[0110] Then, the respectively computed median values are added to their respective precalibrated nacelle positions to get the recalibrated nacelle positions of each wind turbine 1, 2, 3, 4, 5. These recalibrated nacelle positions are at least on average expected to be more precise than the precalibrated nacelle positions because the simulation of the predetermined set of power ratios is generally expected to be very precise with respect to the location of the wakes, i.e., in terms of the wind directions of the power minima and maxima.

[0111] In the further step 104, as explained above, the wind direction dependent nacelle position may be computed for every single wind direction. This row is intentionally left open in table 1 due to its dependance on the specific wind direction of interest.

[0112] Although the present invention has been disclosed in the form of embodiments and variations thereon, it will be understood that numerous additional modifications and variations could be made thereto without departing from the scope of the invention.

[0113] For the sake of clarity, it is to be understood that the use of a or an throughout this application does not exclude a plurality, and comprising does not exclude other steps or elements.