Wind turbine control for limiting rotor thrust

Abstract

Aspects of the invention relate to a method of controlling a wind turbine comprising a rotor and a plurality of rotor blades. The method includes: receiving a plurality of blade pat load signals indicative of a measured pat load on a respective rotor blade, each blade pat load signal received from a blade pat load sensor of a respective rotor blade; determining an average blade tap load signal based at least in part on the plurality of blade tap load signals; and determining a maximum thrust level of the thrust limit controller based on the average blade pat load signal and a reference load value indicative of a maximum allowable change in the average blade pat load signal relative to a normal value.

Description

Technical Field

[0001] The present invention relates to a method of controlling a wind turbine. In particular, the present invention relates to a method of controlling a wind turbine to limit rotor thrust during variable wind conditions, such as those caused by gusty winds. Aspects of the present invention relate to a method, a computer-readable storage medium, a controller, and a wind turbine. Background Art

[0002] Wind turbines typically include one or more controllers for controlling various components of the wind turbine. For example, a wind turbine controller may be used to control the pitch angle of the wind turbine's rotor blades and / or the speed of the wind turbine's generator. The wind turbine controller may control the components with the goal of maximizing the power production / energy captured by the wind turbine from the wind and / or minimizing the loads experienced by the various wind turbine components during operation.

[0003] Wind turbine controllers typically determine control actions or set points for components based on various operating parameters associated with the operation of the wind turbine. For example, a wind turbine controller may use rotor blade loads or current wind speed near the wind turbine to determine appropriate control actions.

[0004] One problem is that some wind turbine controllers need to react to relatively rapid changes in wind speed, i.e., provide appropriate control actions to mitigate the loads and stresses experienced by the wind turbine tower. For example, a controller used to limit the thrust loads experienced by a wind turbine rotor needs to react to rapid load changes (e.g., due to wind gusts that produce a sudden increase in blade loads). However, filters used to remove high-frequency noise content from the measured / estimated wind speed or blade loads may have the undesirable effect of also filtering out these faster (i.e., shorter timescale) changes. This can result in suboptimal controller performance during changes on shorter timescales.

[0005] It is against this background that the present invention has been devised. Summary of the Invention

[0006] According to one aspect of the present invention, a method of controlling a wind turbine including a rotor and a plurality of rotor blades is provided. The method includes receiving a plurality of blade flap load signals indicating measured flap loads on respective rotor blades, each blade flap load signal being received from a blade flap load sensor of a respective rotor blade; determining an average blade flap load signal based at least in part on the plurality of blade flap load signals; and determining a maximum thrust level of a thrust limit controller based on the average blade flap load signal and a reference load value, the reference load value indicating a maximum allowable variation of the average blade flap load signal from a normal value.

[0007] Thus, in the context of the present invention, the normal value is the load value expected for the current conditions, and the maximum allowed variation represents the extent to which the load is reasonably expected to vary during these conditions. The average blade flap load signal is a signal that varies with the blade load signal, and the reference signal determines how much the average blade flap load signal can vary before the maximum thrust level is modified (typically reduced). To give an example, the reference load value may be x1 times the standard deviation of the varying average blade flap load signal relative to the mean of that signal, where the mean represents the normal value, i.e., in this example, the normal value is set to the mean of the varying average blade flap load signal, and x1 times the standard deviation represents the maximum allowed variation before the maximum thrust level is modified.

[0008] In this way, the cyclic variation of the flapping load at each individual blade is substantially neutralized (cancelled) in the average blade flapping load signal, so that the average blade flapping load signal can be used to detect whether the average blade flapping load varies outside the acceptable range for normal operation (by exceeding a reference load value). The maximum thrust level can then be adjusted accordingly to protect the wind turbine, thereby allowing the wind turbine to quickly respond to relatively sudden load changes that may be caused by gusts and alleviate excessive loads.

[0009] In an example, the method further includes determining a reference load value based on the corresponding interval of the average blade flap load signal. Optionally, the reference load value is determined at a prescribed frequency. For example, the reference load value may be determined at each time step of the sample load. In this manner, the reference load value indicates a maximum allowable variation of the average blade flap load signal relative to a normal value for the corresponding time period or corresponding operating condition.

[0010] Optionally, the reference load value is based on the standard deviation of the average blade flap load signal. Optionally, the normal value of the average blade flap load is the mean of the average blade flap load signal. This provides a convenient measure of acceptable load variation adapted to conditions.

[0011] In an example, the maximum thrust level is determined as: a first thrust level if the average blade flapping load signal is less than or equal to a reference load value; and a second thrust level if the average blade flapping load signal is greater than the reference load value, wherein the second thrust level is less than the first thrust level. In this manner, as the average blade flapping load signal increases above the reference load value, the maximum thrust level decreases. For example, the first thrust level and the second thrust level may be respective values ​​of the maximum allowable thrust.

[0012] Optionally, determining the maximum thrust level further comprises: comparing the average blade flapping load signal to a further reference load value indicating a significant variation in the average blade flapping load signal relative to a normal value; and determining the maximum thrust level to a second thrust level if the average blade flapping load signal is greater than the further reference load value. In this manner, the maximum thrust level can be maintained at the second thrust level while the average blade flapping load signal is greater than the further reference level. Thus, the second thrust level can be set to correspond to a minimum acceptable rate of power production.

[0013] In an example, the method further comprises determining a maximum thrust level based on a function for transitioning from the first thrust level to the second thrust level if the average blade flapping load signal is greater than a reference load value but less than or equal to another reference load value. Optionally, the function comprises a proportional decrease from the first thrust level toward the second thrust level based on a difference between the average blade flapping load signal and the reference load value. The proportional decrease may allow for a rapid change in the maximum thrust level, thereby ramping down from the first thrust level to the second thrust level (or vice versa).

[0014] Optionally, the additional reference load value is also based on the standard deviation of the average blade flap load signal. For example, the reference load value and the additional reference load value may be equal to X1 times and X2 times, respectively, the standard deviation of the average blade flap load signal relative to the mean of the average blade flap load signal. X2 may be greater than X1. The multiple may be a tunable value set based on the wind turbine design, which may be adjusted to set how much the average blade flap load signal can vary before the maximum thrust level is modified.

[0015] Optionally, determining at least one of the reference load value and / or the further reference load value comprises applying a low pass filter to the standard deviation of the average blade flap load signal.

[0016] Optionally, each of the plurality of blade flap load signals is high-pass filtered before determining the average blade flap load signal based thereon. In this manner, the average blade flap load signal can be effectively centered around the mean, i.e., the high-pass filter can produce an average blade flap load signal having a zero mean. In this manner, there is no explicit need to determine a normal value as the mean, and the normal value (relative to which the average blade flap load variation is assessed) can be considered to be zero.

[0017] Optionally, a notch filter is applied to the average blade flap load signal before determining the maximum thrust level based thereon. The notch filter may be configured to remove cyclic variations of each rotor blade passing through the wind turbine tower and may therefore have a frequency of approximately three times the rotor frequency.

[0018] Optionally, a hold function is applied to the determined maximum thrust level. The hold function can be used to improve stability, for example, by preventing cyclic pitching of the rotor blades. The hold function can be implemented as an asymmetric hold function.

[0019] In an example, the method may further include outputting the determined maximum thrust level to a thrust limit controller to control the pitch angles of the plurality of rotor blades based on the determined maximum thrust level. In this manner, the rotor blades may be pitched to limit the maximum rotor thrust experienced by the wind turbine.

[0020] According to another aspect of the present invention, there is provided a computer program comprising instructions which, when executed by a computer, cause the computer to perform the method of the first aspect.

[0021] According to another aspect of the present invention, there is provided a non-transitory computer-readable storage medium having stored thereon instructions which, when executed by one or more processors, cause the one or more processors to perform the method as described in the previous aspect of the present invention.

[0022] According to another aspect of the present invention, a control system for a wind turbine comprising a rotor and a plurality of rotor blades is provided. The control system comprises one or more controllers configured to: receive a plurality of blade flap load signals indicating measured flap loads on respective rotor blades, each blade flap load signal being received from a blade flap load sensor of a respective rotor blade; determine an average blade flap load signal based at least in part on the plurality of blade flap load signals; and determine a maximum thrust level for a thrust limit controller based on the average blade flap load signal and a reference load value indicating a maximum allowable variation of the average blade flap load signal from a normal value.

[0023] According to another aspect of the present invention, there is provided a wind turbine comprising a control system as described in the previous aspects of the present invention.

[0024] It is expressly intended that within the scope of the present invention, the various aspects, embodiments, examples and alternatives set forth in the preceding paragraphs, in the claims and / or in the following description and drawings, and in particular the individual features thereof, may be employed independently or in any combination. That is, all embodiments and / or features of any embodiment may be combined in any manner and / or combination, unless such features are incompatible. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Examples of the present invention will now be described with reference to the accompanying drawings, in which:

[0026] Figure 1schematically illustrates a wind turbine according to an aspect of the present invention;

[0027] Figure 2 shows the determination according to one aspect of the present invention Figure 1 The steps of an exemplary method of determining a maximum thrust level of a wind turbine;

[0028] Figure 3 A shows a graph of wind speed during an example period;

[0029] Figure 3 B shows the Figure 3 During the example period of A for Figure 1 An exemplary graph of corresponding blade flapping loads determined for each rotor blade of a wind turbine;

[0030] Figure 3 C shows the Figure 3 A and Figure 3 During the example period of B, according to Figure 2 an exemplary graph of average blade flapping loads determined by the method of FIG. 1 ; and

[0031] Figure 4 Schematically illustrates the method for determining Figure 1 A control module device for a maximum thrust level of a wind turbine, the device being used by one or more controllers of the wind turbine and Figure 2 method to achieve this. DETAILED DESCRIPTION

[0032] Examples of the present invention advantageously provide for determining a maximum thrust level for a thrust limit controller to control a wind turbine in response to relatively rapid changes in blade loading (eg, due to wind gusts).

[0033] Examples of the present invention exploit the fact that the effects of relatively rapid changes in wind speed are initially experienced by the rotor blades of a wind turbine. Therefore, examples of the present invention use measurements of the loads experienced by the rotor blades in order to capture shorter timescale variations in rotor thrust, as described in more detail below.

[0034] However, the inventors have discovered that due to short timescale variations in wind speed (eg due to gusts), the blade load measurements for individual rotor blades exhibit cyclic variations that impair the ability to detect blade load variations.

[0035] To alleviate this problem, examples of the present invention advantageously combine blade load measurements determined for multiple rotor blades of a wind turbine by determining an average blade load based thereon. For example, a blade flapping load signal may be received from a respective blade load sensor attached to each rotor blade of the wind turbine, and an average blade flapping load signal may be determined based on the received signal. In this way, cyclic variations in the flapping load at each individual blade are substantially neutralized. Thus, the average blade flapping load signal may be used to detect when the average blade flapping load varies outside an acceptable range for normal operation of the wind turbine, i.e., to detect when the average blade flapping load varies from an expected value for the prevailing conditions (e.g., average wind speed) due to gusts. In this case, the expected value for normal operation and / or the acceptable range of variation corresponds to the current or (in most cases) most recent conditions and may therefore be continuously updated based on a sliding window. In this way, the limits of acceptable blade loads may be more tightly controlled under lower turbulence conditions.

[0036] The maximum thrust level may then be adjusted accordingly to protect the wind turbine. For example, the maximum thrust level may be dynamically adjusted to reduce the maximum thrust level in response to an increasing change in the average blade load signal, for example to produce a proportional response to the magnitude of the wind speed increase.

[0037] Thus, it is contemplated that the present invention will provide enhanced performance and responsiveness of a wind turbine to relatively rapid changes in wind speed, protecting the wind turbine from significant thrust load variations, thereby extending the life of the wind turbine. The present invention will simultaneously minimize periods of reduced power production and facilitate further optimization of wind turbine design, as the reduced power production resulting from low maximum thrust levels can be limited to periods of high variation in the average blade flapping load signal, while for periods of low variation in the average blade flapping load signal, a higher maximum thrust level can be set.

[0038] Figure 1An example of a wind turbine 10 is illustrated in a schematic diagram. Wind turbine 10 includes a tower 102, a nacelle 103 disposed at or on top of tower 102, and a rotor 104 operably coupled to a generator housed within nacelle 103. In addition to the generator, nacelle 103 houses other components required to convert wind energy into electrical energy, as well as various components required to operate, control, and optimize the performance of wind turbine 10. The rotor 104 of wind turbine 10 includes a central hub 105 and three rotor blades 106 projecting outward from the central hub 105. Rotor blades 106 are pitch-adjustable. Rotor blades 106 can be adjusted according to a collective pitch setting, in which each blade is set to the same pitch value. Additionally or alternatively, rotor blades 106 can be adjusted according to individual pitch settings, in which each blade 106 can be provided with a separate pitch set point. In other examples, it should be understood that the rotor may include fewer or additional rotor blades projecting outward from the central hub.

[0039] In this example, wind turbine 10 includes blade load sensors placed at or near each blade root 109 in a manner such that each sensor detects the load in a corresponding rotor blade 106. Blade load signals from such sensors can be used to determine how to adjust the pitch of each of the individual blades 106. Depending on the placement and type of sensor, loads can be detected in the flapping direction (in-plane / out-of-plane) or the edgewise direction (in-plane). For example, such sensors can be strain gauge sensors or optical Bragg sensors.

[0040] A method for controlling a wind turbine 10 according to an example of the present invention will now be described.

[0041] The method may be implemented by one or more controllers or other processing modules associated with wind turbine 10. In particular, one or more wind turbine controllers are configured to detect relatively rapid changes in blade loads and determine corresponding control actions in response. One or more wind turbine controllers may include or be connected to a thrust limiter controller that receives a maximum thrust level for wind turbine 10 and controls the rotor blade pitch angle to ensure that the loads experienced by wind turbine rotor 104 remain below the maximum thrust level (thereby reducing extreme loads and fatigue effects on one or more wind turbine components). Because gusts of wind may cause thrust loads to increase relatively quickly, determining the maximum thrust level based on blade loads that include high-frequency content, i.e., load variations that occur on short timescales (e.g., less than a few seconds, such as less than two seconds), with minimal delay ensures better or optimal performance of such a thrust limiter controller.

[0042] In an example, the controller(s) or processing module(s) may be located within wind turbine 10, such as within nacelle 103, within tower 102, or distributed at multiple locations within turbine 10 and communicatively connected to one another. Alternatively, the controller (or portions thereof) may be located external to wind turbine 10.

[0043] The controller(s) may be in the form of any suitable computing device, such as one or more functional units or modules implemented on one or more computer processors. Such functional units may be provided by suitable software running on any suitable computing substrate using conventional or custom processors and memory. One or more functional units may use a common computing substrate (e.g., they may run on the same server) or separate substrates, or one or both substrates may themselves be distributed among multiple computing devices. Computer memory may store instructions for executing the method performed by the controller, and the processor may execute the stored instructions to perform the method.

[0044] The described method uses measured flap loads experienced by rotor blades 106 to determine a maximum thrust level for wind turbine rotor 104. The maximum thrust level is a limit value for the total (aerodynamic) force acting on rotor 104 in a direction acting along the axis of rotation of rotor 104, which is used to control wind turbine 10. Measured flap loads, i.e., loads in the blade's flap direction, are obtained from blade load sensors, and each blade flap load measurement indicates a bending moment at the root 109 of rotor blade 106. The maximum thrust level is then determined by combining (e.g., averaging) the measured flap load signals obtained for each rotor blade 106 and adjusting the maximum thrust level based on changes in the combined signal. In this way, a thrust limit controller can control wind turbine 10 to mitigate excessive thrust without explicitly determining or estimating the actual thrust experienced by wind turbine 10.

[0045] Figure 2 The steps of a method 20 for controlling a wind turbine 10 according to an example of the present invention are schematically illustrated.

[0046] At step 202 of method 20, a plurality of blade flap load signals are received at a wind turbine controller. Each blade flap load signal indicates a measured flap load on a respective one of rotor blades 106 and may be obtained from a respective blade load sensor. In an example, a blade flap load signal may be received for some or each of rotor blades 106.

[0047] In an example, each of the plurality of blade flap load signals may be passed through a high-pass filter to attenuate load variations below a cutoff frequency, thereby limiting subsequent analysis to higher frequency components and producing a substantially zero mean blade flap load. In other words, to enable each blade flap load signal to include only load deviations due to high frequency content, the processing module may apply filtering to attenuate low frequency content from each of the blade flap load signals, thereby producing a substantially zero mean blade flap load for each signal.

[0048] At step 204, the blade flap load signals are combined in a suitable manner to obtain a representative blade flap load signal for rotor blade 106. For example, multiple blade flap load signals, which may have passed through a high pass filter, may be added together and divided by the number of blade load signals to determine an average blade flap load signal.

[0049] To give an example, Figure 1 Each of the three rotor blades 106 of the wind turbine 10 shown in FIG. receives a blade flap load signal, and the signals may therefore be approximately 120 degrees out of phase. By summing the three blade flap load signals and determining an average blade flap load signal, the cyclic load variations at each individual blade are mitigated by the phase differences of the other signals. Thus, high-frequency content is preserved without cyclic variations that could falsely indicate short-timescale load variations, such as those due to gusty winds.

[0050] In an example, wind turbine control can be further improved by additionally filtering the remaining higher frequency cyclic load variations in the average blade flapping load signal as successive rotor blades 106 pass through the tower 102 during each rotation. To this end, a notch filter can be applied to the determined average blade flapping load signal. For example, the notch filter can be configured to attenuate 3P frequencies (i.e., three times the rotor frequency). In an example, the notch filter can also be configured to additionally or alternatively attenuate 6P and / or 9P frequencies (i.e., six or nine times the rotor frequency).

[0051] At step 206 , method 20 involves determining a maximum thrust level for wind turbine 10 based on the average blade flap load signal.

[0052] To this end, the average blade flapping load signal is analyzed to determine unacceptable variations (e.g., due to sudden changes in wind speed or direction) that require a reduction in the maximum thrust level to protect wind turbine 10. Therefore, as discussed in more detail below, variations in the average blade flapping load signal are analyzed relative to a standard or normal blade flapping load for prevailing conditions, such as current / recent wind speed and direction.

[0053] Therefore, to determine whether a reduction in the maximum thrust level is required, the average blade flap load signal may be compared with a reference load value that indicates a maximum allowable variation of the average blade flap load signal relative to a normal value of the average blade flap load signal. For example, the normal value of the average blade flap load signal may be a mean, median, or mode value of the average blade flap loads.

[0054] For example, a normalized value and / or a reference load value for the average blade flapping load signal can be determined or updated at a prescribed frequency based on successive intervals of the average blade flapping load signal. In this manner, the normalized value and / or the reference load value can be updated to account for current and / or recent wind conditions. Thus, the reference load value can be updated at regular intervals to accommodate long-term wind speed variations and / or to accommodate, for example, relatively stable or turbulent conditions during which wind speed variations may be relatively small or relatively large, respectively.

[0055] For example, a normal value and / or a reference load value of the average blade flapping load signal can be determined or updated at each time step using a sliding window that samples consecutive intervals of less than or equal to 1 minute. In this way, these values ​​can be considered to correspond to the current wind conditions. However, this example is not intended to limit the scope of the present invention, and in other examples, the interval can be, for example, less than or equal to 30 seconds, or up to 10 minutes, or even an hour.

[0056] In an example, the standard deviation of the average blade flapping load signal can be determined as a measure of acceptable variation from a reference load value. The use of the standard deviation allows the time taken to detect sudden load increases to be minimized, particularly in low turbulence conditions where the maximum thrust level can be set at a relatively high level. Minimizing reaction under such conditions is crucial to mitigating the thrust loads experienced by rotor 104. However, it should be understood that other measures of acceptable variation from a normal value may also be suitable.

[0057] Since the plurality of blade flap load signals are each high pass filtered to have a zero mean load value in this example, the normal value of the average blade flap load signal in this example is a zero load value.

[0058] Thus, unacceptable variation can be considered as the average blade flap load signal being greater than a reference load value equal to X1 multiplied by the standard deviation about the mean, where X1 is a specified scalar coefficient. For example, X1 can be a positive integer.

[0059] If the average blade flap load signal remains less than or equal to a reference load value that is X1 multiplied by the standard deviation, the maximum thrust level may be set to the first thrust level T1 .

[0060] However, if the average blade flap load signal is greater than the reference load value, the maximum thrust level may be reduced to protect the wind turbine 10 .

[0061] To this end, the controller can include one or more rules, algorithms, or functions for reducing the maximum thrust level based on the average blade flap load signal being greater than a reference load value. Such rules, algorithms, or functions can allow for rapid adjustment of wind turbine 10 when the average blade flap load signal exceeds the reference load value. For example, the controller can be configured to determine the maximum thrust level based on a function for transitioning from a first thrust level T1 to a reduced second thrust level T2 for an average blade flap load value between a first reference load value and a second reference load value (such as X1 and X2 multiplied by a standard deviation from the mean). In this regard, X2 can be a specified scalar coefficient greater than or equal to X1 and can be, for example, a positive integer. To give an example, X1 can be equal to 4 and X2 can be equal to 6.

[0062] Thus, as the average blade flap load signal increases from X1 times the standard deviation to X2 times the standard deviation, the maximum thrust level decreases according to the transfer function. For example, as the average blade flap load signal increases from X1 times the standard deviation to X2 times the standard deviation, the transfer function may scale the maximum thrust level from a first thrust level T1 to a second thrust level T2. Thereafter, for an average load blade flap load greater than or equal to a second reference load value (X2 multiplied by the standard deviation), the maximum thrust level may be maintained at, for example, the second thrust level T2. Thus, the second thrust level T2 may define a minimum value for the maximum thrust level, which protects the wind turbine from short-timescale variations in wind speed / direction without causing unnecessary or excessive periods of reduced power production by the wind turbine 10.

[0063] In this manner, the average blade flapping load signal is used to determine short-timescale load deviations (which may occur due to, for example, gusts) by comparison with reference load deviations that are considered normal or acceptable for the prevailing wind conditions. Thus, the maximum thrust level can be adjusted accordingly to protect wind turbine 10 from damage.

[0064] In step 208, the determined maximum thrust level is output to control wind turbine 10. For example, the maximum thrust level may be output to one or more wind turbine controllers to ensure that the loads experienced by wind turbine rotor 104 remain below the maximum thrust level. For example, the maximum thrust level may be output to a thrust limiter controller that controls the rotor blade pitch angle to ensure that the loads experienced by wind turbine rotor 104 remain below the maximum thrust level.

[0065] For example, the thrust limit controller may transmit a collective pitch reference or operating point (to the pitch actuator system of wind turbine 10 ) based on the pitch angle of rotor blades 106 to be controlled.

[0066] In an embodiment, the thrust limiter may set a minimum pitch angle, which ensures that the rotor thrust is below a maximum thrust level. The thrust limiter controller may compare a desired pitch angle with the minimum pitch angle and output the desired pitch angle if the desired pitch angle is greater than the minimum pitch angle, and output the minimum pitch angle if the desired pitch angle is less than the minimum pitch angle.

[0067] In this way, sudden blade load changes, eg, due to wind gusts, may be distinguished from noise and typical cyclic variations, allowing the maximum thrust level to be quickly adapted to protect wind turbine 10 from excessive transient loads without unnecessarily reducing power production.

[0068] It is therefore contemplated that the present invention will provide enhanced responsiveness of the wind turbine to relatively rapid changes in wind speed, protecting the wind turbine from significant thrust load variations, thereby extending the life of the wind turbine without unnecessary curtailment.

[0069] Figure 3 A shows a graph 30 of a wind speed curve 301 that depicts wind speed measurements at the wind turbine 10 during an example period. In the example shown, a brief increase in wind speed is evident after a period of approximately 22 seconds. For comparison, Figure 3 B shows a graph 31 of a first blade flapping load curve 302, a second blade flapping load curve 303, and a third blade flapping load curve 304 received from respective blade load sensors during an example period, and Figure 3 C shows a graph 32 of an average blade flap load signal curve 305 determined based on the first blade flap load curve 302 , the second blade flap load curve 303 and the third blade flap load curve 304 according to step 204 of the method 20 .

[0070] As in Figure 3 B and Figure 3 As is apparent from Figure C, load fluctuations at the individual blade load sensors are mitigated in the determined average blade flapping load curve 305, such that the normal value of the average blade flapping load curve 305 is zero, and deviations caused by short-time-scale increases in wind speed (after 22 seconds) can be distinguished from the relatively high-frequency deviations of the remaining effects. Thus, according to step 206 of method 20, corresponding load thresholds can be set to reflect acceptable and unacceptable load deviations, and short-term deviations can be effectively mitigated, thereby maintaining maximum thrust levels without unnecessarily pitching the blades and / or reducing power production.

[0071] Figure 4 A processing or control module arrangement 40 that may be implemented by a wind turbine controller according to an example of the present invention is schematically illustrated. Specifically, a first blade flapping load signal 401 is obtained from a first blade load sensor, a second blade flapping load signal 402 is obtained from a second blade load sensor, and a third blade flapping load signal 403 is obtained from a third blade load sensor. In the illustrated embodiment, each of the first blade flapping load signal 401, the second blade flapping load signal 402, and the third blade flapping load signal 403 is filtered by applying a corresponding high-pass filter 41 to attenuate low-frequency content, thereby generating a zero-mean load for each blade flapping load signal 401, 402, 403. In other words, the processing module may apply filtering to remove low-frequency content from each of the blade flapping load signals 401, 402, 403 so that the blade flapping load signals 401, 402, 403 only include load deviations due to high-frequency content.

[0072] High frequency content may be defined in any suitable manner. However, as a purely illustrative example, high frequency content may be content greater than 0.5 Hz. Similarly, low frequency content may be defined in any suitable manner. As a purely illustrative example, low frequency content may be content less than 0.1 Hz.

[0073] In some examples, the high pass filter applied to each blade flap load signal 401 , 402 , 403 may be tuned to pass content near the first flapping eigenfrequency of the corresponding rotor blade 106 , which will vary depending on the specific size of rotor 104 and rotor blade length.

[0074] The high pass filtered signals are then added together at summing junction 42 and the added signal is divided by the number of blade flap load signals 401, 402, 403 received at division junction 43 (three in this example) to determine an average blade load signal 404. In the example shown, a notch filter 44 is applied to the average blade load signal 404 to attenuate cyclic effects (also known as 3P frequencies), such as those produced when successive rotor blades 106 pass the tower 102 during each rotation.

[0075] The determined average blade loading signal 404 is then used to determine a maximum thrust level 405. In particular, the average blade loading signal 404 is passed to the processing module 45 for determining the maximum thrust level 405 based on the average blade loading signal 404. The average blade loading signal is split into two signal paths 404A and 404B. The first signal path 404A serves as one of the inputs into the thrust conversion module 50. The average blade loading signal is abbreviated as 'abls' and is shown as a value along the x-axis of a graph that maps the maximum thrust level (abbreviated here as 'mtl') to a function of the average blade loading signal.

[0076] The processing module 45 analyzes the average blade load signal 404 to determine whether there is a sudden, transient increase in blade load that requires a reduction in the maximum thrust level. To this end, the processing module 45 determines a reference load value based on the second signal path 404B as a further input to the thrust conversion module 50, the further input defining the thrust limits of acceptable load variations for normal operation under the current conditions, and the processing module 45 determines the maximum thrust level by comparing the average blade load signals abls 404, 404A with such reference load values ​​R1, R2, 406, 407.

[0077] like Figure 4 As shown, for this purpose, the processing module 45 may therefore include a standard deviation module 46 , a low-pass filter 47 , first and second gain modules 48 , 49 and a thrust conversion module 50 .

[0078] In particular, the average blade load signal 404B is first passed to the standard deviation module 46, which determines the standard deviation of the average blade load signal 404, 404B to measure the typical variation of the average blade load signal 404 under the current conditions. The standard deviation may be determined at regular intervals or at each time step, for example, to adapt to the current wind conditions, particularly their relative turbulence.

[0079] Processing module 45 determines the standard deviation of the average blade flap load signal as a measure of acceptable variation, allowing control module assembly 40 to minimize detection time in low turbulence conditions. In particular, control module assembly 45 is relatively sensitive to load variations in low turbulence conditions compared to relatively (high) turbulence conditions. This is appropriate because, due to other components of the thrust limiter, the maximum thrust level is typically set to a high static thrust in low turbulence conditions, making wind turbine 10 susceptible to sudden wind speed increases. Conversely, in extreme turbulence conditions, the maximum thrust level is typically set to a low level by other components of the thrust limiter, allowing for a reasonably higher tolerance.

[0080] A low pass filter 47 removes high frequency content from the output of the standard deviation module 46, and the filtered output is passed to a first gain module 48 and a second gain module 49. The first gain module 48 generates a first input signal 406 to the thrust conversion module 50, and the second gain module 48 generates a second input signal 407 to the thrust conversion module 50 for defining respective reference loads for comparison with the average blade flapping load signals abls 404, 404A.

[0081] Specifically, the first input signal 406 is equal to X1 times the low-pass filtered standard deviation of the average blade load signal and forms a first reference load value R1. Similarly, the second input signal 407 is equal to X2 times the low-pass filtered standard deviation of the average blade load signal and forms a second reference load value R2. Here, X1 is a first gain determined by the first gain module 48, and X2 is a second gain determined by the second gain module 49, and X2 is greater than or equal to X1 to provide a larger reference load value.

[0082] In the above example, it will be appreciated that due to the high-pass filter 41, the normal value is a zero load value, making it unnecessary to explicitly determine the normal value as a mean value. However, in other examples, a normal value (e.g., as a mean, mode, or median) of the average blade load signal may be determined for consecutive intervals using a sliding window instead. Furthermore, in alternative embodiments, the first reference load value and the second reference load value may be determined by adding the first input signal 406 and the second input signal 407, respectively, to the determined normal value.

[0083] The thrust conversion module 50 receives the first and second input signals 406, 407, and the average blade flapping load signals abls 404, 404A and determines a corresponding maximum thrust level. Considering this in more detail, the thrust conversion module 50 uses the first input signal 406 to determine a lower load reference value (in this case, equal to X1 multiplied by the standard deviation) and uses the second input signal 407 to determine an upper load reference value (in this case, equal to X2 multiplied by the standard deviation). Thus, the thrust conversion module 50 determines a first thrust level T1 for an average blade flapping load that is less than or equal to the lower reference load value. For average blade flapping load values ​​between the lower reference load value and the upper reference load value, the thrust conversion module 50 applies a function that linearly reduces the maximum thrust level from the first thrust level T1 to a second thrust level T2. If the average blade flapping load value is greater than or equal to the upper reference load value, the thrust conversion module 50 sets the maximum thrust level to the second thrust level T2 to protect the wind turbine 10 accordingly.

[0084] For example, the first thrust level T1 may be greater than or equal to a maximum expected thrust level of the wind turbine 10. In an example, the first thrust level T1 and the second thrust level T2 may be determined based on the current operation of the wind turbine 10 and may be set by other parts of the wind turbine control, such as by reference to one or more lookup tables.

[0085] The thrust levels T1 and T2 may be preset levels that are set during the tuning process. These thrust levels may be the thrusts that are maximally permissible given the level of variation in the average blade flapping load signal (maximum permissible thrust). The thrust level T1 may be set during the tuning process to an appropriate maximum thrust level for stable wind conditions (i.e., wind conditions that result in a low level of variation in the average blade load), while the thrust level T2 may be set during the tuning process to an appropriate maximum thrust level for varying wind conditions (i.e., wind conditions that result in a high level of variation in the average blade load). Reference levels R1 and R2 may be set during the tuning process to set how fast the transition from a high maximum thrust level to a low maximum thrust level should be. In this way, the maximum thrust level of the wind turbine may be adapted in a flexible manner to specific wind turbine design variants and specific wind turbine site conditions.

[0086] Before the maximum thrust level signal 405 is output to one or more wind turbine controllers that determine corresponding control actions, the determined maximum thrust level is maintained by a hold module 51, which applies a hold function for stability purposes. Hold module 51 can provide greater stability under certain conditions. For example, the hold time can be a predetermined period on the order of approximately 10 seconds, such as between 2 and 20 seconds, and ensures that wind turbine 10 is not controlled to pitch inward immediately after pitching outward. Otherwise, this could lead to instability or a limit cycling phenomenon, where thrust limiting is continuously turned on and off. A short hold period results in a quick return to the high thrust level T1 when fluctuations in the average blade load signal decrease, while a long hold period results in a longer maintenance of the low thrust level.

[0087] The hold function may be implemented as an asymmetric hold function. In the asymmetric hold function, the determined maximum thrust level is only held if the average blade load signal decreases, i.e. the asymmetric hold function may be implemented such that the determined maximum thrust level is held for a predetermined period for a decreasing average blade flapping load signal and the determined maximum thrust level is adjusted for an increasing average blade flapping load signal. Figure 4, illustrated by arrows 52 and 53. In the case of a given average blade load value measured as indicated by arrow 53, the corresponding maximum thrust level is determined. The following average blade load value can be larger or smaller, causing the resulting maximum thrust level to change as indicated by arrow 53. If the average blade load signal is towards a smaller value (higher maximum thrust level), the holding module maintains the value, while if the average blade load signal is towards a higher value (lower maximum thrust level), the value is set to the maximum thrust level. If, after expiration of a predetermined period, the value remains below the initial value, the maximum thrust level is increased, typically by using a preset slope to avoid signal jumps. If a higher value has been detected during the predetermined period, the predetermined period is reset each time a higher value is detected.

[0088] Output signal 405 may be used as input to a thrust limiter controller, for example, where rotor blade pitch angle is controlled to ensure that loads experienced by wind turbine rotor 104 remain below a maximum thrust level (thereby reducing fatigue of one or more wind turbine components).

[0089] Because the maximum thrust level signal 405 is intended to protect the wind turbine 10 from relatively rapid load changes, the maximum thrust level signal 405 can be updated relatively frequently or substantially continuously to ensure that such changes are reflected in the determination of the maximum thrust level. For example, the blade flapping loads can be sampled at a defined sampling rate of the controller 40. The method for obtaining the maximum thrust level can be performed for each set of sampled data (blade loads), i.e., at each time step. Alternatively, the maximum thrust level signal 405 can be updated at a defined time step, i.e., at defined time intervals.

[0090] Thus, the present invention provides enhanced wind turbine protection and reduced tower loads without unnecessarily reducing power production.

[0091] Figure 4 The diagram illustrates a specific advantageous implementation of the present invention that includes a number of specific elements. Although each described element represents an advantageous embodiment of that element, not all elements need to be implemented in the manner shown in order to achieve an advantageous working implementation of the present invention.

[0092] Many modifications may be made to the examples described without departing from the scope of the appended claims.

Claims

1. A method of controlling a wind turbine comprising a rotor and a plurality of rotor blades, the method comprising: receiving a plurality of blade flap load signals indicative of measured flap loads on respective rotor blades, each blade flap load signal being received from a blade flap load sensor of a respective rotor blade; determining an average blade flap load signal based at least in part on the plurality of blade flap load signals; and A maximum thrust level of a thrust limit controller is determined based on the average blade flap load signal and a reference load value indicating a maximum allowable variation of the average blade flap load signal from a normal value. 2 . The method according to claim 1 , further comprising determining the reference load value based on corresponding intervals of the average blade flapping load signal, optionally wherein the reference load value is determined at a prescribed frequency.

3. The method according to claim 1 or claim 2, wherein: The reference load value is based on a standard deviation of the average blade flap load signal.

4. A method according to any preceding claim, wherein: The maximum thrust level is determined as: a first thrust level if the average blade flapping load signal is less than or equal to the reference load value; and If the average blade flapping load signal is greater than the reference load value, then a second thrust level is selected, wherein the second thrust level is less than the first thrust level.

5. The method according to claim 4, wherein Determining the maximum thrust level further includes: comparing the average blade flap load signal to a further reference load value, the further reference load value indicating a large variation of the average blade flap load signal relative to the normal value; and If the average blade flap load signal is greater than the further reference load value, the maximum thrust level is determined as the second thrust level.

6. The method according to claim 5, wherein: The method further includes determining the maximum thrust level based on a function for transitioning from the first thrust level to the second thrust level if the average blade flap load signal is greater than the reference load value but less than or equal to the further reference load value.

7. The method according to claim 6, wherein: The function comprises a proportional decrease from the first thrust level towards the second thrust level according to a difference between the average blade flap load signal and the reference load value.

8. A method according to any one of claims 5 to 7, when dependent on claim 3, wherein The additional reference load value is also based on the standard deviation of the average blade flap load signal, the reference load value and the additional reference load value being equal to respective X1 times and X2 times the standard deviation of the average blade flap load signal relative to the mean of the average blade flap load signal, wherein X2 is greater than X1.

9. The method according to claim 8, wherein Determining at least one of the reference load value and / or the further reference load value comprises applying a low pass filter to a standard deviation of the average blade flap load signal.

10. A method according to any preceding claim, wherein each of the plurality of blade flap load signals is high pass filtered prior to determining the average blade flap load signal based thereon; and Wherein, before determining the maximum thrust level based on the average blade flapping load signal, a notch filter is applied to the average blade flapping load signal.

11. A method according to any preceding claim, wherein: Apply the hold function to the determined maximum thrust level.

12. The method according to claim 11, wherein The hold function is an asymmetric hold function implemented to hold the determined maximum thrust level for a predetermined period for a decreasing average blade flap load signal and to adjust the determined maximum thrust level for an increasing average blade flap load signal.

13. The method of any preceding claim, further comprising outputting the determined maximum thrust level to the thrust limit controller to control the pitch angles of the plurality of rotor blades based on the determined maximum thrust level.

14. A non-transitory computer-readable storage medium having stored thereon instructions which, when executed by one or more processors, cause the one or more processors to perform the method according to any preceding claim.

15. A control system for a wind turbine comprising a rotor and a plurality of rotor blades, the control system comprising one or more controllers configured to: receiving a plurality of blade flap load signals indicative of measured flap loads on respective rotor blades, each blade flap load signal being received from a blade flap load sensor of a respective rotor blade; determining an average blade flap load signal based at least in part on the plurality of blade flap load signals; and A maximum thrust level of a thrust limit controller is determined based on the average blade flap load signal and a reference load value indicating a maximum allowable variation of the average blade flap load signal from a normal value.

16. A wind turbine comprising a control system according to claim 15.

Images