METHOD FOR DETECTING EXTREME LOAD ON A WIND POWER PLANT

Description

[0001] The present invention relates to a method for detecting an asymmetric extreme load caused by a gust of wind acting on a wind turbine. The present invention also relates to a method for controlling a wind turbine to reduce extreme loads. The invention also relates to a wind turbine for implementing at least one of the two methods.

[0002] Wind turbines generate electrical power from wind and are typically designed as so-called horizontal-axis wind turbines, each with an aerodynamic rotor with several, usually three, rotor blades. During operation, these rotor blades sweep across a rotor field as the aerodynamic rotor rotates. In modern wind turbines, such rotors have diameters well over 100m, sometimes over 150m. This allows a lot of power to be generated from the wind, but also results in large loads. Particularly during gusts, high extreme loads can occur briefly, which can also only occur locally within the large rotor field. This can result in asymmetric loads, which can place additional strain on the wind turbine.

[0003] To reduce such loads on the wind turbine, the wind speed and / or the load on the rotor blades can be continuously monitored. If excessive load on the aerodynamic rotor or rotor blades is detected, the rotor blades can be rotated slightly out of the wind in order to reduce the load or prevent a further increase in the load.

[0004] However, the problem can arise that the already high load on the aerodynamic rotor has led to a corresponding bending of the wind turbine tower. The tower is thus – to put it simply – pushed slightly backward or deflected backward. By initiating the load reduction, the pressure on the aerodynamic rotor is also reduced, causing the wind turbine to swing back, at least in the area of ​​the tower head. This, in turn, places a strain on the tower that should also be avoided.

[0005] To detect such loads early, wind speed measurement devices can also be used. These can measure the wind speed at a certain distance in front of the aerodynamic rotor. This allows an approaching gust of wind to be detected. Such advance detection can be achieved using a so-called LiDAR measurement system, but such LiDAR systems can be expensive and prone to errors in unfavorable air conditions, especially in heavily polluted air, including heavily rained air.

[0006] US 2021 / 115896 A1 relates to a method for reducing loads on a rotor blade. The method involves reconstructing one or more load signals from the rotor blade. US 2014 / 178197 A1 concerns a single-blade pitch control system in which the rotor plane is divided into several sectors. The individual sectors are determined for each blade during rotation using a corresponding azimuth angle sensor.

[0007] The present invention is therefore based on the object of addressing at least one of the above-mentioned problems. In particular, a solution for detecting and / or avoiding or reducing an asymmetric extreme load caused by a gust of wind is to be proposed. At the very least, an alternative solution to previously known solutions is to be proposed.

[0008] According to the invention, a method according to claim 1 is proposed. The method thus relates to the detection of an asymmetric extreme load caused by a gust of wind acting on a wind turbine. An extreme load is a load that exceeds a predeterminable load limit. An asymmetric extreme load is one that does not act uniformly on the rotor of the wind turbine, but only selectively, i.e., only in a partial area of ​​the rotor field.

[0009] The underlying wind turbine has a rotor with at least three rotor blades, whereby the rotor blades are adjustable in their blade angle and the rotor sweeps over a rotor field with its rotor blades.

[0010] The method comprises continuously measuring a blade load for each rotor blade. The blade load can be measured continuously or at least quasi-continuously. The measurement can also be carried out, for example, at predetermined time intervals, which should, however, be short enough so that, at the rated rotor speed, a blade load is measured at least three times per revolution. Preferably, however, the blade load should be measured much more frequently, in particular at least ten or at least twenty times per rotor revolution. The blade load is preferably measured at least every 60° of a rotor revolution, in particular at least every 30° of a rotor revolution.

[0011] The blade load can be measured in particular by measuring a load, e.g. by means of a strain gauge, at a blade root of each rotor blade or at the rotor blade near the blade root, in particular at a maximum distance of 10 m from the blade root on the rotor blade.

[0012] It is further proposed that for at least one sector of the rotor field, at least one temporal sector load profile be determined from recorded blade loads of different rotor blades at the same azimuth position. Such a sector load profile describes a temporal profile of a load on the rotor blades in the sector and includes a profile extrapolated for a future period. Thus, for at least one sector, a blade load of the rotor blade located in the sector is recorded. The rotor continues to rotate accordingly, so that after some time, another rotor blade enters the area of ​​the sector under consideration, and the blade load for this rotor blade is then recorded and included in the sector load profile of this sector.

[0013] These two blade loads can already represent a temporal progression. This temporal progression can be extrapolated over time, for example, using linear regression. For example, if the blade load of a first rotor blade in the sector under consideration is 40% of a reference blade load at the first point in time, and the blade load of the next rotor blade is 50% of the reference blade load at the next point in time, 60% of the reference blade load can be assumed as the extrapolated value for a third point in time, and thus also for a third rotor blade. The progression from the second to the third point in time is thus the extrapolated progression of the sector load progression.

[0014] The rotation of the rotor is referred to as rotation in the azimuth direction and accordingly the positions of the rotor blades with respect to this rotation in the azimuth direction are referred to as azimuth positions.

[0015] To determine a temporal sector load profile, it is proposed that the blade loads be recorded or taken into account at successive recording times that are separated by a partial time period in which the rotor rotates one rotor blade further, so that successive blade loads are recorded or taken into account for the respective sector. For a respective sector, a blade load is therefore always recorded when a rotor blade is in the corresponding position and the next recording takes place when the next rotor blade is in the corresponding position in the sector due to the rotation of the rotor. For this purpose, the blade loads can also be recorded continuously or quasi-continuously, with the blade loads then being assigned to successive sectors. This means that several measurements can be carried out per sector and, for example, averaged in order to record a blade load for the respective sector.This allows consecutive blade loads to be recorded for each sector, namely several blade loads per revolution.

[0016] It was particularly recognized that by recording or taking into account the loads on different rotor blades at the recording times, they can be assigned to the same sector. Therefore, there is no need to wait for the rotor to complete a full rotation.

[0017] This procedure is preferably carried out not just for one sector, but for many sectors. For example, the rotor field can be divided into twelve equal sectors, each with a size of 30°. A sector load profile is then determined for each of these twelve sectors, and separate, successive recording times can be provided for each sector.

[0018] The recording or consideration of these successive recording times can, of course, also be implemented in such a way that the absolute azimuth positions of the rotor are taken into account, and the blade loads are always recorded or considered at the same positions. The partial time period is fixed in the case of uniform rotation of the rotor at a constant speed. At a different speed, it assumes a different value, and the partial time period can also change accordingly if the rotor or the rotor speed accelerates or decelerates.

[0019] It is now further proposed that an extreme load be expected depending on at least one sector load profile. In the simplest case, an extreme load is to be expected if the extrapolated profile of a sector load profile reaches or exceeds a load limit. If the wind turbine then continues to operate unchanged, it is to be expected that this extreme load will actually occur and act in particular on one rotor blade. If rotation continues, the extreme load can then also act on another rotor blade.

[0020] However, it is proposed to identify such expected extreme loads using the sector load profile and then take countermeasures to reduce the load. This includes, in particular, adjusting the angle of attack of the rotor blades, or at least the angle of attack of the rotor blade that will next reach the sector for which an extreme load is expected based on the test.

[0021] Preferably, sector load profiles are recorded for each sector and evaluated accordingly. For each sector load profile, and thus for each sector, a continuous check is made to determine whether an extreme load is to be expected in the respective sector. Each sector load profile is thus checked to determine whether it reaches or exceeds a load limit.

[0022] However, testing for the expectation of an extreme load does not necessarily have to be done by comparing the sector load curve with a load limit. Additional tests may also be performed, such as determining how much such a load limit is exceeded or how quickly the load changes, i.e., how steep the sector load curve is.

[0023] Furthermore, a more complex extrapolation may also be provided, for example, using not just a straight line but a higher-order function for extrapolation. In particular, it may be possible to use more than just two blade loads recorded successively for a sector for extrapolation.

[0024] According to one aspect, it is proposed that the sector load profile be designed as a first-order or higher-order temporal polynomial function. For this purpose, the use of a straight line has already been described, which is thus a first-order temporal polynomial function, i.e., a straight line that can also have a gradient. For a higher order, a time-dependent quadratic polynomial function, for example, can be assumed.

[0025] Additionally or alternatively, it is proposed that an expected extreme load be assumed if the sector load profile for a future point in time reaches or exceeds a predetermined load limit. Such a future point in time is, in particular, one that is offset from the current point in time by the partial time period. The current point in time can, in particular, be the point in time at which the blade load for the sector under consideration was last recorded. In this respect, the future point in time is, in particular, the point in time at which the next rotor blade will reach the sector under consideration.

[0026] However, it is also possible to choose a different future point in time, especially one that is even further in the future, e.g., two part-time periods away from the current point in time. The further in the future the point in time is chosen, the earlier an expected extreme load could be identified, but the less reliable such an expectation can also become.

[0027] According to one aspect, it is proposed that the sector load profile be determined from at least two consecutive blade loads of a sector and at least one associated time period, and that a check be performed for a next successive acquisition time, which is still in the future, to determine whether the sector load profile reaches or exceeds the predetermined blade load limit. Using two consecutive blade loads of the same sector and the associated partial time period, a temporal gradient of the sector load profile can also be determined, and assuming a constant gradient, the expected blade load of the same sector can be determined for the next acquisition time.

[0028] In this respect, a prediction of the blade load is made for the next recording time. The sector load profile thus determined, and thus also the predicted blade load value for the next recording time, can be compared with the predetermined blade load limit. This allows for advance detection of whether an extreme load is to be expected. An extreme load is assumed when the predetermined blade load limit is reached or exceeded.

[0029] According to one aspect, it is proposed that an expected blade load be determined using the sector load profile for a successive recording time to be tested. The current blade load is recorded at the successive recording time to be tested and compared with the expected blade load to determine an expected deviation. The sector load profile is then adjusted depending on the determined expected deviation.

[0030] The expected curve, which was calculated in advance, is compared with the curve that actually occurred. The deviation at this successive recording time to be checked can form the determined expected deviation. The sector load curve can be adjusted accordingly. In particular, this involves a prediction for the next recording time, which follows the currently checked recording time. The expected deviation can be used, for example, to determine if the blade load in the sector under consideration is increasing faster than predicted by the sector load curve. This can then be detected if the blade load limit may be reached faster than expected.

[0031] The sector load curve can be adjusted in particular by changing the sector load curve from a straight line to a curved line, or by otherwise increasing its order. However, the sector load curve can also be supplemented by an additive component. For example, a straight line with a higher gradient can be appended to a sector load curve that is defined as a straight line with a gradient. It is also possible to simply adjust the parameterization of the sector load curve. If the sector load curve is defined as a straight line with a gradient, this gradient can be changed accordingly. If the sector load curve is a second or higher order polynomial function, namely dependent on time, several parameters can be adjusted accordingly.

[0032] According to one aspect, it is proposed that the blade loads of a rotor blade having a blade root be measured on the rotor blade in the area of ​​the blade root, in particular as blade bending or blade bending moment. It has been particularly recognized here that the entire load acting on the rotor blade acts at the blade root or in the area of ​​the blade root. Such a blade load can be measured in particular by one or more strain gauges in this area.

[0033] Ideally, such a blade load is absorbed directly at the blade root. However, the blade root can often be made of metal, or at least transition into a metallic connection area for connection to the rotor hub, while the rotor blade is otherwise made of a different material, particularly glass-fiber reinforced plastic. To avoid corresponding measurement inaccuracies, which can occur particularly in the aforementioned transition area between two materials, a small distance can be provided between the measuring sensor on the rotor blade and the blade root, particularly to the connection flange. The blade root area is understood to be the inner 10% of the rotor blade. This is the first 10% from a blade connection adapter of the rotor blade to a rotor blade tip. Bending can occur here, which can be easily detected and essentially reflect the entire load on the rotor blade.

[0034] For this purpose, bending sensors, particularly strain gauges, can be arranged on the rotor blade in the area of ​​the blade root, which are also distributed circumferentially around a rotor blade's longitudinal axis. At least two bending sensors, particularly strain gauges, should then be arranged, distributed circumferentially at 90° around the rotor blade's longitudinal axis. This is based on the realization that the blade load to be measured can also have a direction. This direction depends in particular, but not exclusively, on the set blade angle. By distributing the bending sensors in this way, the direction of the respective blade load can be detected and thus the overall amplitude of the blade load can be correctly measured.

[0035] According to one aspect, it is proposed that several sectors of the rotor field be monitored for extreme loads, and that the blade loads be recorded for each observed sector at successive recording times, namely such that successive blade loads are recorded for each sector, and at least one change in the blade loads of the respective sector is determined from this. To this end, it is further proposed that an expected change in the blade loads of a second sector be inferred from the at least one change in the blade loads of a first sector. This is done, in particular, such that the sector load profile of the second sector is adapted depending on the sector load profile of the first sector.

[0036] It is planned to record asymmetric extreme loads, so that the blade loads of the sectors are expected to differ from one another. Nevertheless, it was recognized that an increase, in particular a strong increase, in a blade load in one sector also allows conclusions to be drawn about a corresponding increase in a blade load in another sector, in particular in a neighboring sector. It is therefore proposed that findings from one sector's sector load profile be taken into account for another sector, namely in particular for its sector load profile. In particular, the sector load profile of the second sector can be adapted from the sector load profile of the first sector if the first sector is located in front of the second sector in the direction of rotation of the rotor.

[0037] However, other scenarios are also considered. In particular, it was recognized that an increase in blade loading at one position in the rotor field, especially a significant increase in such a blade load, generally allows conclusions to be drawn about an increase in blade loads in the entire rotor field. In other words, while the absolute blade loads of one sector cannot be directly determined from those of another sector, a scenario of an increase in wind speed and thus in blade load can apply to the entire rotor field or at least to a region of the rotor field extending across several sectors.

[0038] Additionally or alternatively, the at least one change in the blade loads of the first sector can be used to determine the expected change in the blade loads of the second sector by determining a first sector load profile for the first sector and a second sector load profile for the second sector. A first expected deviation is determined for the first sector load profile, and the second sector load profile is adjusted depending on the first expected deviation. Thus, an expected deviation is also determined here, but is not used, or not only used, to improve the sector load profile of the same sector, but is used to improve the sector load profile of another sector. As a result, an extreme load expected in a sector can be detected even earlier.

[0039] According to the invention, it is proposed that an extreme load time be determined at which an extreme load is expected to occur. In particular, it is proposed that the extreme load time be determined from the at least one sector load profile. It is therefore proposed to check when an extreme load occurs or is expected. Depending on this, the wind turbine control system can take appropriate precautions.

[0040] One possible precaution is to adjust the blade pitch of the rotor blades, or at least one of the rotor blades, at an early stage in order to mitigate the expected extreme load. It was particularly recognized that taking such an early measure prevents the blade load, and thus the thrust on the rotor blade and thus the rotor, from being reduced until the wind turbine's tower head has already been pushed significantly backward. If the load is only reduced then, the tower head may swing back forward, which in turn may lead to tower loads that are to be avoided. By reducing the blade load at an early stage, this tower load can be avoided, or at least reduced.

[0041] According to one aspect, it is proposed that the blade angles of the rotor blades are adjustable independently of one another.

[0042] Because the rotor blade pitches are independently adjustable, the blade loads can be individually influenced. The proposed method's early detection of specific loads also allows for targeted independent adjustment, i.e., individual blade pitch adjustment. It was also recognized that individual blade pitch adjustment can place a significant strain on the pitch drives. However, early detection of the expected blade load allows for timely initiation of the adjustment, allowing for a lower pitch speed and thus reducing the load.

[0043] Additionally or alternatively, it is proposed that the blade angle of the respective rotor blade be taken into account to determine a sector load profile for each detected blade load. In particular, it is proposed that each detected blade load be converted, depending on the associated blade angle, into an equivalent blade load that corresponds to a blade load at a predetermined reference blade angle.

[0044] It was particularly recognized that the blade angle has a significant influence on the blade load and can therefore be taken into account when determining the sector load profile. It was also recognized that the wind turbine may have individual blade pitch, in which the rotor blades can assume different blade angles independently of each other. This can lead to different blade loads for different rotor blades in the same sector under the same wind. It is proposed that this be taken into account and, in particular, calculated out.

[0045] Even if the blade angles of all rotor blades are adjusted synchronously, different blade angles can result in the same sector for different blades because the rotor blades are in the same sector at different times.

[0046] Each sector load curve should reflect the load of the respective sector, regardless of the specific rotor blade currently experiencing such a load. To ensure consistent results, it is proposed to convert each recorded blade load into an equivalent blade load. The equivalent blade load then indicates the blade load that would occur if the rotor blade were at the predetermined reference blade angle. The reference blade angle thus creates a uniform reference value, making blade loads recorded at different blade angles easier to compare.

[0047] In addition, the sector load curve also includes a curve extrapolated for a future period. If the blade load is related to a reference blade angle, this extrapolation can be carried out without further ado, without having to know the blade angle of the corresponding rotor blade for the extrapolated period. In addition, the sector load curve can be designed as a continuous curve, although a blade load in the corresponding sector can actually only occur if a rotor blade is located in the sector. Thus, time periods between two points in time would be undefined with regard to the rotor blade angle, since there is no rotor blade in the sector at that time. By referring to the reference blade angle, the dependence on the blade angle is essentially calculated out, thus eliminating the aforementioned problem of uncertainty.

[0048] Particularly preferably, a part-load blade angle is used as the predetermined reference blade angle, which has been determined and is used as the aerodynamically optimal blade angle for the part-load range or part-load operation.

[0049] In particular, this is based on a wind turbine that is controlled in such a way that a fixed blade angle is used for all rotor blades in the partial load range.

[0050] The partial load range or partial load operation is the range in which the prevailing wind speed has not yet reached the nominal wind speed and / or the wind turbine is still operating below nominal speed, nominal power and / or nominal torque without being artificially throttled.

[0051] At such a rotor blade angle, a maximum blade load can be expected compared to other blade angles. If the sector load curve is related to this blade angle, the sector load curve also reflects the maximum load curve. The maximum load can be evaluated accordingly, and depending on this, if it exceeds the predetermined blade load limit, a blade angle with a lower blade load can be set or specified. Ultimately, the wind turbine or individual rotor blades can already be operated at a blade angle different from the predetermined reference blade angle.

[0052] Preferably, a minimum blade angle can be determined from an analysis of a sector load profile, which must not be undercut. If the blade angle is smaller, it is then adjusted to this specified minimum blade angle. However, if the current blade angle is larger or the same, no adjustment is required. This is based on a common definition of blade angles, whereby they reach their maximum value towards a feathered position, i.e. their maximum value when they are turned completely out of the wind. The part-load blade angle mentioned is accordingly a small angle. The part-load blade angle can be in the range from -5° to 10°, in particular in the range from 0° to 5°, whereas the blade angle in feathered position is in the range from 80° to 110°, in particular in the range from around 90° to 100°.

[0053] According to one aspect, it is proposed that each recorded blade load be converted into a local wind value depending on the associated blade angle, wherein it is particularly provided that a wind field is created from the wind values ​​of some or all sectors. Furthermore or alternatively, it is proposed that each sector load profile be converted into a wind profile within the sector, so that each wind profile contains a profile extrapolated for a future period. In particular, it is provided that a wind field profile is created from the wind profiles of some or all sectors.

[0054] The recorded blade load is based on a specific wind situation, which can therefore also be derived from the recorded blade load. For this purpose, for example, relationships between the recorded blade load and the wind situation can be recorded in a measuring system with additional complex wind sensors, particularly LiDAR. Based on these relationships, which can be recorded in advance, wind values ​​and the wind field can then be deduced from the loads during ongoing operation. The wind field thus describes the distribution of the wind values ​​across the sectors of the rotor field. Relationships between the recorded blade load and the wind situation can also be determined through simulations, e.g., using the blade element method (BEM).

[0055] Such a wind situation can include not only an average wind force but also a wind direction for the relevant sector. A wind value can thus be formed as a pair of values ​​consisting of wind speed and wind direction. If the wind directions differ from one sector to the next, a wind shear can also be derived from this. This can also be part of the wind value. Wind directions and shear can be derived from variations in blade load from one sector to the next. For example, a wind direction could be derived from two different load values, one representing a bending moment and the other a torsional moment of the same blade.

[0056] From the sector load profiles, a wind pattern for the respective sector can also be derived, i.e., a temporal change in the wind value of the sector in question. Here, too, changes in wind direction can be part of such a wind pattern. Wind shear can also be derived from this, especially by further considering the wind patterns of other, particularly neighboring, sectors.

[0057] It was thus recognized that wind patterns can be derived from the sector load patterns, and these can be considered as a wind field and thus as a wind field pattern. The underlying sector load patterns also contain extrapolated patterns, and in the same way, the wind patterns and thus the wind field pattern can also contain an extrapolated area. This makes it possible to predict the wind field in the rotor field.

[0058] However, it is also possible to create a wind field using only the wind values ​​from some or all sectors. Therefore, it is not necessarily necessary to create a wind field profile with a forecast; simply creating an actual state can also be useful.

[0059] According to one aspect, it is proposed that, upon detection of an expected extreme load in at least one sector, plant operation is modified in order to reduce or limit a load on the wind turbine. In particular, it is proposed that the plant operation be modified by adjusting the blade angle of at least one of the rotor blades in order to reduce or limit a blade load in the at least one rotor blade. Thus, it is proposed to use the detection of an expected extreme load in such a way that measures to reduce the load are taken in a timely manner by changing the operation of the wind turbine. In particular, it is proposed to turn the rotor blades, or at least one, out of the wind in a timely manner. The effects of this have already been described above.

[0060] In particular, it is proposed that, once an extreme load point has been identified at which the extreme load is expected, the turbine operation be modified before the extreme load point is reached. This not only prevents the extreme load from occurring, but also prevents an overreaction when changing the wind turbine operation, which would occur if the reaction were to occur only after the extreme load has occurred.

[0061] Additionally or alternatively, it is proposed that a sector be identified in which the extreme load is expected, and that the blade angle of a rotor blade be changed before it reaches the sector for which the extreme load is expected. In particular, individual blade adjustment is proposed here, which specifically adjusts only the respective rotor blade to reduce the load in order to avoid the expected extreme load in precisely the localized sector. Otherwise, the wind turbine can continue to operate normally, and in particular, blade adjustment is not required in other areas of the rotor field, or not to the same extent, so that as much energy as possible can continue to be generated there.

[0062] In addition, or alternatively, it is proposed that the blade angles of all rotor blades be adjusted. In particular, it is suggested that this be done before the extreme load is expected and that the angles be reduced again once the loads have decreased. However, it is also conceivable in principle that all rotor blades be adjusted simultaneously when one approaches the sector in which the extreme load is expected.

[0063] According to one aspect, it is proposed that an adjustment angle for adjusting at least one rotor blade is determined from the at least one determined sector load profile, and / or a target time is determined by which the adjustment angle is to be adjusted, and in particular an adjustment speed is determined and specified from the target time and the adjustment angle, and / or a minimum blade angle to be set is determined below which a rotor blade is not adjusted in its blade angle.

[0064] Thus, it is specifically intended that the sector load profile not only determines whether a blade adjustment is required, but also determines its magnitude. A blade angle can be directly derived and set from the sector load profile, and a blade angle profile can also be considered. Such a blade angle profile can be based on the sector load profile or selected so that the blade angle adjustment begins early, so that it reaches a desired final value in good time.

[0065] The sector load curve not only provides early insight into when adjustment is advisable, but also into how best to perform it. The load curve indicates when maximum load can be reached and how large this would be. From this, the target time at which blade adjustment should be completed can be derived, as well as the blade angle that should be achieved to sufficiently reduce the expected load. This yields the adjustment angle. Using the adjustment time to be observed, i.e., the time until the target time, the adjustment speed can be determined and specified. Each rotor blade can therefore be adjusted to the new position at an adjusted adjustment speed.

[0066] The greatest load is to be expected at a low blade angle, especially at a part-load blade angle, which was already explained above. To relieve the load or to avoid large loads, a rotor blade is turned out of the wind, which means its blade angle increases. A minimum blade angle setting therefore represents a lower limit for the blade angle setting and thus an upper limit for the load. It may happen that the blade angle is already greater than such a minimum blade angle setting, and in this case, an expected extreme load, which is derived from at least one sector load profile, should not result in a reduction in the blade angle, as this would further increase the current load.

[0067] According to one aspect, a method for controlling a wind turbine is proposed, wherein the wind turbine is controlled depending on the detection of an asymmetric extreme load caused by a gust of wind acting on a wind turbine, and wherein the control method implements at least one method for detecting an asymmetric extreme load caused by a gust of wind according to one of the preceding aspects. The control method is thus based on the detection method. The wind turbine is controlled depending on the detection of the extreme loads.

[0068] To the extent that the method for detecting an asymmetric extreme load caused by a gust of wind also includes steps or features for controlling or other steps or features, these steps or features should also be included in the method for controlling.

[0069] The invention also relates to a wind turbine, prepared to to carry out a method for detecting an asymmetric extreme load caused by a gust of wind acting on the wind turbine, and / or to carry out a method for controlling the wind turbine, wherein the wind turbine is controlled as a function of detecting an asymmetric extreme load caused by a gust of wind acting on a wind turbine, wherein the wind turbine has a rotor with at least three rotor blades, the rotor blades are adjustable in their blade angle, and the rotor sweeps over a rotor field with its rotor blades, and the method comprises continuously detecting a blade load for each rotor blade, determining at least one temporal sector load profile from detected blade loads of different rotor blades of the same azimuth position for at least one sector of the rotor field,which describes a temporal progression of a load on the rotor blades in the sector and includes a progression extrapolated for a future period, wherein the blade loads are recorded or taken into account at successive recording times which are separated by a partial time period in which the rotor rotates further by one rotor blade, so that successive blade loads are recorded or taken into account for the respective sector, and checking for the expectation of an extreme load depending on the at least one sector load progression, wherein, an extreme load time is determined at which an occurrence of an extreme load is to be expected, in particular that the extreme load time is determined from the at least one sector load profile (208).

[0070] In particular, such a method for controlling the wind turbine is implemented in the wind turbine, in particular in a process computer, control computer, or other control device. A detection device is provided for detecting a blade load for each rotor blade. This can, for example, comprise strain gauges on the rotor blades and at least one evaluation device for evaluating corresponding signals received from the strain gauges.

[0071] In particular, the wind turbine is prepared to execute a method according to one of the aspects described above. In particular, the wind turbine comprises a control device for executing a control method, and the control method is implemented in the control device. For this purpose, the control device may comprise a process computer in which the method is implemented. However, the control device may also comprise several units among which the method is distributed.

[0072] Preferably, the control device comprises a control unit for controlling the wind turbine and a detection unit for detecting an extreme load.

[0073] The invention is explained in more detail below using exemplary embodiments with reference to the accompanying figures. Figure 1 shows a wind turbine in a perspective view. Figure 2 shows a timing diagram to illustrate one aspect. Figure 3 shows a front view of a rotor field. Figure 4 shows a schematic side view of a wind turbine and thus also of a Figure 3 Figures 5 and 6 each show a front view of a rotor field at different rotor positions. Figure 7 shows a timing diagram illustrating a forecast improvement.

[0074] Fig. 1shows a schematic representation of a wind turbine according to the invention. The wind turbine 100 has a tower 102 and a nacelle 104 on the tower 102. An aerodynamic rotor 106 with three rotor blades 108 and a spinner 110 is provided on the nacelle 104. During operation of the wind turbine, the aerodynamic rotor 106 is set in rotation by the wind and thus also rotates an electrodynamic rotor or rotor of a generator, which is directly or indirectly coupled to the aerodynamic rotor 106. The electrical generator is arranged in the nacelle 104 and generates electrical energy. The pitch angles of the rotor blades 108 can be changed by pitch motors at the rotor blade roots 109 of the respective rotor blades 108.

[0075] The wind turbine 100 has an electrical generator 101, which is indicated in the nacelle 104. Electrical power can be generated by means of the generator 101. A feed-in unit 105, which can be designed particularly as an inverter, is provided for feeding in electrical power. This allows a three-phase feed-in current and / or a three-phase feed-in voltage to be generated according to amplitude, frequency, and phase for feeding into a grid connection point PCC. This can be done directly or jointly with other wind turbines in a wind farm. A system controller 103 is provided for controlling the wind turbine 100 and the feed-in unit 105. The system controller 103 can also receive default values ​​from external sources, in particular from a central farm computer.

[0076] Figure 2shows a time diagram in which an exemplary load curve of an equivalent blade load is shown for three rotor blades, which is thus converted to a reference angle, i.e. standardized. These load curves, which can also be synonymously referred to as blade load curves, are shown in the Figure 2 referred to as B1, B2 and B3, i.e. as load curves for a first, second and third rotor blade.

[0077] In addition, a blade angle curve 202 is shown. The blade angles can be adjusted independently of each other, which was also the basis here, but the deviations between the blades are for the general explanation according to Figure 2 not important, so for the sake of simplicity only this single blade angle curve 202 is shown in this Figure 2 is shown.

[0078] The time diagram thus shows time in seconds on the abscissa, load in kilonewton meters on the left ordinate, and the blade angle in degrees on the right ordinate. The temporal zero point is chosen at the point where a load curve would have reached a limit, namely the permissible maximum value, if the blade angle had not been changed. A suboptimal blade load curve B2' is shown, which reaches an extreme load value of 204 at t=0 and represents a curve without blade angle adjustment.

[0079] Starting about 5 s before this zero point, the three load curves B1-B3 are subject to fluctuations, but with increasing amplitude. Previously, they were subject to similar fluctuations, but not increasing, which is not shown here. The fluctuations in the load of these three load curves B1-B3 are due, among other things, to the fact that the rotor blades sweep over different areas in the rotor field as the rotor rotates and are therefore exposed to different wind loads. The loads can be lower in the lower area in particular and higher in the upper area. However, wind fluctuations or other variations in the wind field also occur. The curves are also shown in a simplified manner, since further smaller fluctuations are not important here.

[0080] The time diagram of the Figure 2is based on the operation of the wind turbine in which the rotor rotates at approximately 20 rpm. This means that the rotor rotates approximately one rotor blade every 1 s. This results in a type of load peak for each blade changing every 1 s. The rotor blades cover a wind range from which a high load emanates. This will be particularly discussed in the following sections. Figures 3 and 4 explained.

[0081] In the time diagram according to Figure 2 The extreme load value of 204 is also shown as a horizontal line. This depicted load should not be reached, or at least not exceeded.

[0082] Starting from the initially low load, it can be seen that this load is increasing. To prevent the extreme load value 204 from being reached, a corresponding check can be performed. The loads shown can be recorded, for example, using strain gauges at or near the blade root and compared with a limit value, such as the extreme load value 204.

[0083] However, this has the disadvantage that such detection and comparison occurs very late. Alternatively, a comparison could be made with a reduced limit value to trigger a countermeasure, namely a blade adjustment, earlier. However, this could lead to an undesired triggering if the extreme load value 204 is not reached at all, but such a reduced limit value is only reached once.

[0084] Instead, it is proposed to use the Figure 2to evaluate the increase to be detected in order to anticipate the expected reaching of the extreme load value 204.

[0085] Such a load increase can be detected, for example, by observing only one blade load curve, e.g., the first blade load curve B1. For this purpose, an individual increase line 206 is shown as a solid line, which is essentially formed as a connection between the two maximum values ​​of the first blade load curve B1. However, such an individual load line 206 has the disadvantage that it can be inaccurate and / or requires at least one rotor revolution to be recorded at all.

[0086] Instead, a sector load curve 208 is proposed, which is shown there as a long dashed straight line. The sector load curve 208 is shown here as a straight line and is determined from two values ​​of the two blade load curves B1 and B3. The two load values ​​used for this are shown as the first and second load values ​​W1 and W2. From these two values, which are only 1 second apart in the example shown, this sector load curve 208 can be determined, and it can be seen that the load in the sector would reach its maximum value very soon.

[0087] At the moment the second load value W2 is reached, this sector load curve 208 can be calculated, and the imminent reaching of the extreme load value 204 can be detected. Thus, at this point in time—here, one second before the extreme load value 204 is reached—pitching is initiated. The blade angle of all blades is increased, thus turning the rotor blades out of the wind. This is shown by the blade angle curve 202.

[0088] By adjusting the blade angle(s) early, a moderate adjustment speed is possible, which in this case is 2° per second (2° / s). However, it could also be lower, e.g., 1° / s. In the example shown, the blade angle has thus changed by 2° in the first second from the start of adjustment. This already means that the load curve B2 of the second blade no longer reaches the extreme load value 204. After two seconds, the blade angle has been adjusted by 4°, which leads to a significant reduction in load. The blade angle can then be reduced again, at least somewhat. For the purpose of illustration, it is shown in the Figure 2 from 1s the constant value of 4°.

[0089] In particular, the following is proposed: At the time a blade pitch adjustment is initiated, the current load on the blades, a pitch angle, and the expected load are known or can at least be estimated. Based on this, an angle can be predicted, so that a resulting final angle is known. This allows the pitch speed to be specified accordingly and extended slowly or quickly, depending on the situation. The pitch speed, i.e., the speed at which the rotor blades are adjusted, can thus be specifically specified and controlled.

[0090] This sector load curve 208 can alternatively also consider the load values ​​of all three blade load curves B1-B3, for example, by additionally considering the third load value W3 from blade load curve B2. The sector load curve 208 would then be somewhat different. It could also be formed as a straight line, but also as a second-order curve, i.e., described by a time-dependent second-order polynomial.

[0091] The load values ​​W1 to W3 can, for example, each be the average of several measured values ​​for the respective blade in the relevant sector. The sector load curve 208 does not have to be a tangent of the two blade load curves B1 and B3 or all three load curves. However, the load values ​​W1 to W3 are in the range of the maximum values ​​of the blade load curves, because this is where the sector with the maximum load lies. Sectors with lower loads are also examined, but do not trigger the blade adjustment. Therefore, their analysis is not presented here.

[0092] For example, the rotor field can be divided into 36 sectors of 10° each, and in each of these 36 sectors, such a sector load profile 208 is recorded. This then also results in sector load profiles that may, for example, also be in the range of the lower values, but then do not lead to the detection of a high load.

[0093] In the example of 36 sectors, this results in 36 sector load profiles, and of these, the most critical is considered for load minimization, i.e., the one that tends to reach the extreme load value 204. Sector load profile 208 is one such and thus is also located approximately at the load peaks of the three blade load profiles B1-B3.

[0094] By using all rotor blades, i.e. the load values ​​of all rotor blades of a sector, the sector load curve 208 can, for example, be constructed early on as a straight line based on two load values, i.e. one load value each for two rotor blades. However, it is also possible to record multiple values, e.g. three values, as already described above. This still makes it possible to determine the blade load curve before the rotor has completed one revolution. Nevertheless, three values ​​are already available, which also make it possible to display the sector load curve 208 not only as a straight line, but also as a curve. When using three load values, a second-order curve, i.e. one that can be described by a second-order polynomial, can be used.However, it's also possible to use a straight line that's essentially overdetermined by evaluating three values, but this can compensate for inaccuracies in the values. The result can thus be more accurate.

[0095] It is also possible to record even more values. If four values ​​are recorded, which would then require one rotation of the rotor, a straight line can still be used as the sector load curve 208, which is even more overdetermined and thus allows inaccuracies to be compensated for even better. However, it is also possible to use a second-order curve, which is also overdetermined and thus allows inaccuracies to be compensated for. If necessary, however, it would of course also be possible to use an even higher-order curve.

[0096] Simulating sector load curves using a first- or second-order curve is the preferred variant.

[0097] The evaluation of such a sector load curve 208 can be carried out by checking whether the sector load curve 208 reaches the extreme load value 204 within a predetermined test period TP. In Figure 2 This test time TP is approximately 2 s. However, the current value of the sector load curve, and thus also the current load, has only reached a reduced extreme load value of 210. Detection is therefore very early, so that the measure, adjusting the rotor blades, results in only the reduced extreme load value of 210 being reached. This test time TP can also be selected depending on the adjustment speed of the blade adjustment.

[0098] Figure 3shows a rotor field 320 in a front view, in which three rotor blades A-C are also schematically indicated. Also shown is an extreme load-driving wind event 322, such as a gust. This wind event 322 occurs in a specific area of ​​the rotor field 320 and thus in a specific sector. The wind event 322 can also affect multiple sectors.

[0099] A rotor rotation 324 is also schematically indicated by a corresponding arrow. However, the rotation of the rotor does not change the position of the wind event 322, and thus one rotor blade after another will reach this wind event as long as it persists.

[0100] In Figure 3Two sectors S1 and S2 are shown as examples, each of which can be represented by an average angle ϕ x or ϕ y , respectively. Rotor blade B is currently located in sector S2. It is possible to record how long one or two rotor blades take to cover this angle. From this, a load increase rate can also be derived.

[0101] The situation of the Figure 3 is in Figure 4 shown in a side view. Wind event 322 is simplified as a cylinder moving toward rotor field 320. This is indicated by event movement 326. It is thus evident that wind event 322 only occurs for a limited time. This indicates the time duration TE, which thus illustrates the timescale of the wind event driving the extreme load.

[0102] Figures 5 and 6illustrate how the load situation can be handled, in particular how a sector load profile can be recorded. Figure 5 shows an example of a sector at approximately the 8 o'clock position. Figure 5 Thus, a rotor field 520 is also shown, which has three rotor blades A, B, and C. Rotor blade C is in the 8 o'clock position. There, a current load is recorded, for example, by appropriate strain gauges at the blade root or near the blade root. A measurement angle α m is also recorded. The measurement angle α m is the current blade angle of rotor blade C. It can be measured, but it can also be read from the control data of the wind turbine or reused.

[0103] Depending on the rotor blade angle and the measured load, a load standardized for a specific angle value is thus obtained. In particular, a blade angle of 0° or a partial load angle can be used as a reference value.

[0104] Here too, the rotor continues to rotate according to rotor rotation 524 and one position later is in Figure 6 illustrated.

[0105] In Figure 6 The rotor blade B is therefore in the 8 o'clock position and there the load on the blade and the blade angle, illustrated by the measuring angle α m , are recorded again.

[0106] Here, the load can be calculated, normalized to a reference angle, and this then provides two values ​​from which a sector load profile can be determined. This normalization allows different blade positions to be calculated, because it is possible that a blade adjustment was performed during the rotation of rotor blade B from the 4 o'clock position to the 8 o'clock position. It is also possible that rotor blades B and C have different blade angles. However, in order to be able to make a statement for the sector under consideration, particularly regarding the load profile based on values ​​from different rotor blades, this normalization to a reference angle is preferably carried out.

[0107] The procedure explained for the 8 o'clock position as an example is carried out equally and essentially simultaneously for all sectors into which the rotor, and thus the rotor field 520, has been divided. Of course, a load value cannot be recorded exactly simultaneously for all sectors, because the rotor blades can only be in one sector at a time. Therefore, the corresponding values ​​can be recorded for three sectors simultaneously.

[0108] Figure 7 shows a time diagram in which load measurements for two sectors are presented. The abscissa shows a time axis and the ordinate a load, which corresponds to the left ordinate in Figure 2 could correspond, at least in principle. In Figure 7 The improvement of a forecast should be explained in principle, so that concrete load values ​​are not shown. Figure 2Very schematic. Here, it is assumed that load values ​​in two sectors are considered for just under one revolution.

[0109] Measurements in sector S 1 are represented by a + and measurements in sector S 2 by an x. Accordingly, the first measurement is taken at time t 1 for the first sector, the next measurement at time t 2 for the second sector. At time t s another measurement is taken in the first sector, namely for the next leaf, and at time t 4 for the second sector, also for this next leaf. These measured values ​​are marked as measured values ​​M 1 -M 4. From the measured values ​​M 1 and M 3 for the first sector, a sector load curve can be determined, for example as a straight line. From this, a load value for time t s can be forecast, which is shown in the diagram as P 5. A circle was chosen as the symbol to make it clear that this is not a measurement but a forecast. Likewise, a forecast value P 6 has been determined for time t s for the second sector, namely from the measured values ​​M 2 and M 4 .

[0110] Now, at time ts, another measurement M 5 is taken, which turns out to be higher than the forecast value P 5 . It can therefore be deduced that, contrary to the forecast, the wind speed is increasing. This information can also be used for the second sector S 2 . From this, it can be deduced that the forecast P 6 may be too low. Similarly, for the stronger increase in wind speed derived from the development of the first sector, it can be deduced that this is also to be expected for sector S 2 . Accordingly, the forecast P 6 can be increased to the corrected forecast K 6 .

[0111] The invention thus describes a solution that allows for early detection of a sector load profile. A load profile is detected sector by sector, based on the values ​​of each rotor blade located in the sector at a given measurement time.

[0112] Particularly Figure 2illustrates a so-called load control event, which, although shown schematically, corresponds to an actual load application. The illustration is also intended to convey in principle that the loads on all blades are already continuously increasing, making early intervention possible using the blade angle.

[0113] In particular, the following assumptions were made. It was assumed that the loads increase over a characteristic time scale, typically several seconds. Furthermore, it is assumed that there is a local asymmetric wind speed increase that does not affect the entire rotor field. This is particularly important with the Figures 3 and 4can be illustrated. Therefore, the depicted event can also be considered a shear event. It is assumed that the timescale is such that each blade sweeps over the extreme load forcing of the wind event or wind condition several times. Thus, each blade sweeps over the corresponding sector or sectors in which such extreme load forcing of the wind event occurs several times.

[0114] In particular, the following measures are suggested, which will also be explained in the figures. A temporal progression of the blade bending moments of all blades will be recorded. For example, strain gauges can be installed and continuously evaluated.

[0115] It is possible to determine a difference between two blade bending moments of two consecutive blades, for example from blade C to blade B, as described in the Figures 5 and 6The load can be recorded as a function of the rotor position and a difference calculated. In this case, two load values ​​are recorded.

[0116] When determining blade bending or blade loading, and thus also when determining the difference between two blade bending moments, the influence of the blade angle is also taken into account. For this purpose, the blade angle of the leading blade can be considered, for example, or a general reference angle can be used. In either case, the influence of the blade angle is taken into account, and this can be done by determining the blade angle difference between the two measurement times, i.e., between the two rotor blades. This blade angle difference can also be taken into account when determining the difference between the two blade loading values, i.e., between the two blade bending moments, because it causes a relative change in the blade bending moment, which is then taken into account.

[0117] Based on these differences in both blade angle and time, a slope is determined, which is used to estimate an expected extreme load at certain discrete time intervals.

[0118] The slope can be determined by a linear line or by a higher-order polynomial. For this purpose, more than two values ​​can be included if necessary.

[0119] When a limit criterion is exceeded, a control is triggered to increase the blade angle of one or all blades in order to reduce the expected extreme load. The limit criterion can be formed from a combination of an extreme load limit, such as the extreme load value 204 of the Figure 2, and a time interval from the current point in time until the occurrence of a corresponding extreme load, i.e., until the extreme load limit is reached. In other words, the line, if it rises at all, can reach the limit criterion, i.e., the extreme load value 204. However, if this occurs far in the future, no action needs to be taken.

[0120] The number of sectors into which the rotor field is divided can be selected, e.g. 36 sectors, each of 10°.

[0121] This allows the blade angles to be adjusted early on by observing such a pitch, allowing for a low pitch speed. Accordingly, in addition to reducing extreme loads on the blades, more gentle operation can also be achieved. This can, in particular, help avoid peak loads on pitch motors and reduce tower deflection in extreme scenarios.

[0122] In particular, the load can be extrapolated for a third of a rotor revolution. After a third of a rotor revolution, two values ​​can be recorded, namely for one blade and the next, and extrapolated from these. For each sector considered, the quality of the forecast after 1 / X rotor revolution is thus known. Currently, it is assumed that X = 3. The focus here is on identifying the quality of the forecast or improving it, as described in Figure 7is explained.

[0123] Although a third of a revolution is required to rotate the rotor from one blade to the next in a sector, an assessment of the quality of the forecast is possible before another third of the rotor revolution has occurred, although it would also be possible to wait for a third of the rotor revolution. Even after the rotor has moved a little further, it is possible to determine whether the wind increase is generally stronger or weaker. This knowledge can be applied to the forecasts already created, thus assessing their quality. If the wind increase accelerates during this time, it will be clear that a forecast has been exceeded.

[0124] This information can then be used for the following sectors to correct the forecast, as shown in Figure 7 is explained.

[0125] The proposed solution has achieved a particular improvement. In a previous variant, which is now being improved, the pitch system reacts when an extreme load occurs in a single blade and attempts to prevent a further increase in the load by pitching out very quickly. Such a short-term reduction in the thrust absorbed by the rotor causes the tower to accelerate sharply forward, which also leads to unwanted extreme tower loads. The proposed solution, however, allows for slower pitching, i.e., slower adjustment of the rotor blades, which does not result in a rapid decrease in thrust and thus does not cause the tower to accelerate sharply forward. Previously, the tower would essentially swing back forward from a previously strong thrust. This is now avoided.

Claims

1. A method for identifying an asymmetrical extreme load which is caused by a gust of wind and acts on a wind power installation (100), wherein the wind power installation - has a rotor (106) having at least three rotor blades (108); - the rotor blades (108) are adjustable in terms of the blade angle thereof; and - the rotor by way of the rotor blades thereof sweeps a rotor field (320); and the method comprises - continuous detecting of a blade load for each rotor blade (108); - ascertaining for at least one sector of the rotor field at least one temporal sector load profile (208) from blade loads detected of different rotor blades with the same azimuth position, said sector load profile (208) describing a temporal profile of a load on the rotor blades in the sector and containing a profile extrapolated for a future temporal period, wherein - the blade loads are detected or taken into account at successive detection time points which are spaced apart by a partial period in which the rotor rotates further by one rotor blade, so that successive blade loads are detected or taken into account for the respective sector; and - checking in terms of expecting an extreme load as a function of the at least one sector load profile (208), characterized in that - an extreme load time point at which an extreme load is expected to arise is determined, in particular wherein - the extreme load time point is determined from the at least one sector load profile (208).

2. The method as claimed in claim 1, wherein - the sector load profile (208) is configured as a temporal polynomial function of the first or a higher order; and / or - an extreme load to be expected is assumed when the sector load profile (208) for a future time point reaches or exceeds a predetermined blade load limit (204).

3. The method as claimed in claim 1 or 2, wherein - the sector load profile (208) is ascertained from at least two successive blade loads of a sector and at least one associated partial period; and - it is checked whether the sector load profile (208) for a next successive detection time point, which is still in the future, reaches or exceeds a, or the predetermined, blade load limit (204), respectively.

4. The method as claimed in one of the preceding claims, wherein - a blade load to be expected is ascertained by means of the sector load profile (208) for a successive detection time point to be checked; - the current blade load is detected at the successive detection time point to be checked and is compared with the blade load to be expected so as to ascertain an expectation variance; and - the sector load profile (208) is adapted as a function of the ascertained expectation variance.

5. The method as claimed in one of the preceding claims, wherein - the blade loads of a rotor blade that has a blade root are in each case detected, in particular in each case as blade flexing or blade bending moment, on the rotor blade in the region of the blade root [for description: inner 10%].

6. The method as claimed in one of the preceding claims, wherein - a plurality of sectors of the rotor field (320) are observed for extreme loads; and - the blade loads for each observed sector are detected at the successive detection time points, so that successive blade loads are detected for each sector and at least one change in the blade loads of the respective sector is ascertained therefrom; and - a conclusion pertaining to a change to be expected in the blade loads of a second sector is drawn from the at least one change in the blade loads of a first sector, in particular such that - the sector load profile (208) of the second sector is adapted as a function of the sector load profile (208) of the first sector; and / or such that - a first sector load profile is determined for the first sector, and a second sector load profile is determined for the second sector; - a first expectation variance is ascertained for the first sector load profile; and - the second sector load profile is adapted as a function of the first expectation variance.

7. The method as claimed in one of the preceding claims, wherein - the blade angles of the rotor blades are adjustable in a mutually independent manner; and / or - the blade angle of the rotor blade is taken into account for determining a sector load profile (208) for each blade load detected, in particular wherein - each blade load detected as a function of the associated blade angle is converted into an equivalent blade load which corresponds to a blade load at a predetermined reference blade angle.

8. The method as claimed in one of the preceding claims, wherein - each blade load detected as a function of the associated blade angle is converted into a local wind value, wherein it is provided in particular that - a wind field is established from the wind values of some or all sectors; and / or that - each sector load profile is converted into a wind profile in the sector so that each wind profile contains a profile extrapolated for a future temporal period, wherein it is provided in particular that - a wind field profile is established from the wind profiles of some or all sectors.

9. The method as claimed in one of the preceding claims, wherein - when identifying an extreme load to be expected an installation operation in at least one sector is changed so as to reduce or delimit a load on the wind power installation (100), in particular wherein the installation operation is changed in that the blade angle of at least one of the rotor blades is adjusted so as to reduce or delimit a blade load on the at least one rotor blade (108), wherein in particular - when an extreme load time point at which the extreme load is to be expected has been identified, the installation operation is changed before the extreme load time point is reached; and / or - a sector in which the extreme load is expected is identified, and the blade angle of a rotor blade is changed before said rotor blade reaches the sector for which the extreme load is expected; and / or - the blade angle of all rotor blades are adjusted.

10. The method as claimed in one of the preceding claims, wherein - from the at least one ascertained sector load profile (208) - an adjustment angle for adjusting at least one rotor blade (106) is determined; and / or - a target time point until which the adjustment angle is to be adjusted is determined, and an adjustment speed is in particular determined and predefined from the target time point and the adjustment angle; and / or - a minimum blade angle to be adjusted is determined, a rotor blade in terms of the blade angle thereof not being adjusted below said minimum blade angle to be adjusted.

11. A method for controlling a wind power installation (100), wherein the wind power installation is controlled as a function of identifying an asymmetrical extreme load which is caused by a gust of wind and acts on a wind power installation (100); wherein the method for controlling comprises or uses at least one method as claimed in one of the preceding claims.

12. A wind power installation (100), prepared for - carrying out a method for identifying an asymmetrical extreme load which is caused by a gust of wind and acts on the wind power installation (100); and / or - carrying out a method for controlling the wind power installation (100), wherein the wind power installation is controlled as a function of identifying an asymmetrical extreme load which is caused by a gust of wind and acts on a wind power installation; wherein the wind power installation - has a rotor (106) having at least three rotor blades (108); - the rotor blades (108) are adjustable in terms of the blade angles thereof; and - the rotor (106) by way of the rotor blades (108) thereof sweeps a rotor field (320); and the method comprises - continuous detecting of a blade load for each rotor blade (108); - ascertaining for at least one sector of the rotor field at least one temporal sector load profile (208) from blade loads detected of different rotor blades with the same azimuth position, said sector load profile (208) describing a temporal profile of a load on the rotor blades in the sector and containing a profile extrapolated for a future temporal period; wherein - the blade loads are detected or taken into account at successive detection time points which are spaced apart by a partial period in which the rotor (106) rotates further by one rotor blade (108), so that successive blade loads are detected or taken into account for the respective sector; and - checking in terms of expecting an extreme load as a function of the at least one sector load profile (208), wherein - an extreme load time point at which an extreme load is expected to arise is determined, in particular wherein - the extreme load time point is determined from the at least one sector load profile (208).

13. The wind power installation (100) as claimed in claim 12, wherein - the wind power installation (100) is prepared for carrying out a method as claimed in one of claims 1 to 11, in particular wherein the wind power installation (100) has a control device for carrying out one of the methods, and the method is implemented in the control device.