Individual pitch control for mitigating whirling motion of wind turbine rotor blades

Abstract

The invention relates to adjusting individual pitch of wind turbine rotor blades. The invention involves obtaining an edge load signal indicative of edge loading on the individual rotor blades in a wind turbine rotor rotational frame, removing forward whirling frequency content corresponding to forward whirling motion from the edge signal, and applying a rotor azimuth-based transformation operation to obtain a backward whirling edge signal in a wind turbine rotor azimuth fixed frame. The backward whirling edge signal is split into first and second signals corresponding to disturbances at an edge frequency and at the edge frequency minus 2P, respectively. First and second control actions for mitigating backward whirling motion are applied to the first and second signals, respectively. The resulting signals are transformed back into rotor rotational frame to obtain individual pitch offset values which are then used to adjust pitch of the rotor blades.

Description

[0001] INDIVIDUAL PITCH CONTROL FOR MITIGATING WHIRLING MOTION OF WIND

[0002] TURBINE ROTOR BLADES

[0003] FIELD OF THE INVENTION

[0004] The invention relates to individual pitch control of rotor blades of a wind turbine, the individual pitch control being for mitigating whirling motion of the rotor blades. In particular, the invention relates to providing individual pitch control at different frequencies in order to mitigate whirling motion.

[0005] BACKGROUND

[0006] Wind turbines as known in the art include a wind turbine tower supporting a nacelle and a rotor with a number of (typically, three) pitch-adjustable rotor blades mounted thereto. A wind turbine is prone to vibrations such as nacelle movement or rotor blade movement. It is known that certain types of vibrations may be damped by active pitching of the rotor blades or adjusting generator torque.

[0007] Edgewise vibrations of the rotor blades are generally undesirable as the blades are typically weakly damped in that direction, and so there is a greater risk of damage to the blades as a result of vibrations in the blade edgewise direction compared to the blade flapwise direction, for instance. In particular, rotor blades may be excited to oscillate in whirling modes as a result of edgewise vibrations, which can increase wear on various wind turbine components.

[0008] Whirling motion in an edgewise direction of wind turbine rotor blades can be influenced by rotor disturbances at different frequencies. This is at least partly caused by dynamics of the wind turbine nacelle, specifically an asymmetry between dynamics in the tilt and yaw directions. It can therefore be challenging to design and implement wind turbine control routines that can mitigate or reduce whirling modes in an effective and complete manner.

[0009] It is against this background to which the present invention is set. SUMMARY OF THE INVENTION

[0010] The inventor of the present invention has understood that backward whirling motion in an edgewise direction of rotor blades of a wind turbine can be influenced by disturbances of the wind turbine rotor at both an edge frequency and at the edge frequency minus 2P. The edge frequency is the frequency of the rotor blade oscillations in the edgewise direction. 1P is a rotational frequency of the wind turbine rotor, i.e. a frequency with which a complete rotation of the rotor is completed, and so 2P is twice the rotor rotational frequency. The inventor of the present invention has understood that effective regulation / control of backward whirling content / motion can require implementing individual pitch control of the rotor blades at both the edge frequency and at the edge frequency minus 2P. In addition, the inventor has realised that each of these pitching frequencies are subject to different dynamics of the wind turbine system and to different pitching constraints, and so each of these pitching frequencies need to have different control actions in order to target backward whirling motion in an optimal manner. Furthermore, the inventor has realised that in order to achieve such optimal control, frequency content relating to backward whirling motion in a sensor signal indicative of edgewise oscillations needs to be isolated from other frequency content in such a sensor signal, and also the backward whirling frequency content needs to be split in an effective manner into the part resulting from edge frequency oscillations and the part resulting from oscillations at the edge frequency minus 2P.

[0011] According to an aspect of the invention there is provided a controller for a wind turbine having three rotor blades. The controller is for adjusting individual pitch of the three rotor blades. The controller is configured to obtain an edge load signal comprising three signal components each indicative of edge loading on a respective one of the rotor blades in a rotor rotational frame of the wind turbine. The edge load signal is based on edge load sensor measurements from one or more of the rotor blades. The controller is configured to remove, from the obtained edge load signal, forward whirling (FW) frequency content corresponding to FW motion in an edgewise direction of the rotor blades, to obtain a backward whirling (BW) edge signal. The controller is configured to apply a rotor azimuthbased transformation operation to the BW edge signal to obtain a transformed BW edge signal in a rotor azimuth fixed frame of the wind turbine. The controller is configured to generate first and second BW edge signals in the rotor azimuth fixed frame from the transformed BW edge signal. The first BW edge signal corresponds to disturbances at an edge frequency in the rotor rotational frame. The second BW edge signal corresponds to disturbances at the edge frequency minus 2P in the rotor rotational frame. The controller is configured to apply a first BW control action to the first BW edge signal to obtain a first BW edge control signal for mitigating BW motion in the edgewise direction of the rotor blades influenced by disturbances at the edge frequency in the rotor rotational frame. The controller is configured to apply a second BW control action to the second BW edge signal to obtain a second BW edge control signal for mitigating BW motion in the edgewise direction of the rotor blades influenced by disturbances at the edge frequency minus 2P in the rotor rotational frame. The controller is configured to apply an inverse m-blade coordinate transformation to the first and second BW edge control signals to obtain individual BW pitch reference offset values for the respective rotor blades in the rotor rotational frame. The controller is configured to transmit a control signal to adjust pitch of the rotor blades based on the obtained individual BW pitch reference offset values.

[0012] To remove the FW frequency content from the obtained edge load signal, the controller may be configured to apply an edge frequency-based (EF-based) transformation operation to the obtained edge load signal to obtain an EF-based transformed edge signal in an EF- based fixed frame. An angular frequency of a rotation angle used in the EF-based transformation operation may be the edge frequency. To remove the FW frequency content from the obtained edge load signal, the controller may be configured to apply a low pass filter to the EF-based transformed edge signal to remove the FW frequency content. As an alternative to the low pass filter, the controller may be configured to apply a notch filter to the EF-based transformed edge signal to remove the FW frequency content. The notch filter may isolate / select frequency content at twice the edge frequency.

[0013] The signal in the EF-based fixed frame obtained from the low pass filter may be the BW edge signal. The rotor azimuth-based transformation operation may correspond to a rotation operation about a rotation angle having an angular frequency equal to 1 P minus the edge frequency.

[0014] Prior to removing the FW frequency content from the obtained edge load signal, the controller may be configured to apply an m-blade coordinate transformation to the obtained edge load signal to obtain mutually orthogonal signals in the rotor azimuth fixed frame. The controller may be configured to apply a high pass filter to remove gravity-based content in the mutually orthogonal signals. The EF-based transformation operation may be applied to the signal in the rotor azimuth fixed frame obtained from the high pass filter. The EF-based transformation operation may correspond to a phase shift about a rotation angle having an angular frequency equal to the edge frequency minus 1P.

[0015] To obtain the edge load signal, the controller may be configured to apply a further high pass filter to the edge load sensor measurements to remove OP frequency content.

[0016] To apply the first BW control action to the first backward whirling edge signal, the controller may be configured to phase shift the first backward whirling edge signal about a first phase offset angle for mitigating backward whirling motion influenced by disturbances at the edge frequency. Alternatively, or in addition, to apply the second BW control action to the second backward whirling edge signal, the controller may be configured to phase shift the second backward whirling edge signal about a second phase offset angle for mitigating backward whirling motion influenced by disturbances at the edge frequency minus 2P.

[0017] To apply the first BW control action the controller may be configured to apply a first gain to the first backward whirling edge signal. Alternatively, or in addition, to apply the second BW control action the controller may be configured to apply a second gain to the second backward whirling edge signal.

[0018] The controller may be configured to apply a first amplitude-limiting factor to the first backward whirling edge control signal prior to the inverse m-blade coordinate transformation being applied. Alternatively, or in addition, the controller may be configured to apply a second amplitude-limiting factor to the second backward whirling edge control signal prior to the inverse m-blade coordinate transformation being applied.

[0019] To apply the inverse m-blade coordinate transformation to obtain the individual BW pitch reference offset values, the controller may be configured to apply a first inverse m-blade coordinate transformation to the first BW edge signal to obtain first individual BW pitch reference offset values for the respective rotor blades in the rotor rotational frame. The controller may be configured to apply a second inverse m-blade coordinate transformation to the second BW edge signal to obtain second individual BW pitch reference offset values for the respective rotor blades in the rotor rotational frame. The controller may be configured to combine the first individual BW pitch reference offset values and the second BW individual pitch reference offset values to obtain the individual BW pitch reference offset values.

[0020] The controller may be configured to remove, from the obtained edge load signal, BW frequency content corresponding to BW motion in the edgewise direction of the rotor blades, to obtain a FW edge signal. The controller may be configured to apply a FW rotor azimuth-based transformation operation to the FW edge signal to obtain a transformed FW edge signal in the rotor azimuth fixed frame. The controller may be configured to generate a first FW edge signal in the rotor azimuth fixed frame from the transformed FW edge signal. The first FW edge signal may correspond to disturbances at the edge frequency in the rotor rotational frame. The controller may be configured to apply a first FW control action to the first FW edge signal to obtain a first FW edge control signal for mitigating FW motion in the edgewise direction of the rotor blades influenced by disturbances at the edge frequency in the rotor rotational frame. The controller may be configured to apply a first FW inverse m-blade coordinate transformation to the first FW edge control signal to obtain first FW individual pitch reference offset values for the respective rotor blades in the rotor rotational frame. The controller may be configured to combine the first FW individual pitch reference offset values with the BW individual pitch reference offset values to obtain individual pitch reference offset values for the respective rotor blades in the rotor rotational frame. The control signal to adjust pitch of the rotor blades may be based on the obtained individual pitch reference offset values.

[0021] The controller may be configured to generate a second FW edge signal in the rotor azimuth fixed frame from the transformed FW edge signal. The second FW edge signal may correspond to disturbances at the edge frequency plus 2P in the rotor rotational frame. The controller may be configured to apply a second FW control action to the second FW edge signal to obtain a second FW edge control signal for mitigating FW motion in the edgewise direction of the rotor blades influenced by disturbances at the edge frequency plus 2P in the rotor rotational frame. The controller may be configured to apply a second FW inverse m-blade coordinate transformation to the second FW edge control signal to obtain second FW individual pitch reference offset values for the respective rotor blades in the rotor rotational frame. The controller may be configured to combine the second FW individual pitch reference offset values with the first FW individual pitch reference offset values and the BW individual pitch reference offset values to obtain the individual pitch reference offset values for the respective rotor blades in the rotor rotational frame. To remove the BW frequency content from the obtained edge load signal, the controller may be configured to apply a second edge frequency-based (EF-based) transformation operation to the obtained edge load signal to obtain a second EF-based transformed edge signal in a second EF-based fixed frame. An angular frequency of a rotation angle used in the second EF-based transformation operation may be negative edge frequency. The controller may be configured to apply a second low pass filter to the second EF-based transformed edge signal to remove the BW frequency content.

[0022] According to another aspect of the present invention there is provided a wind turbine comprising a controller as defined above.

[0023] According to another aspect of the invention there is provided a wind turbine comprising a controller as described above.

[0024] According to another aspect of the invention there is provided a method for controlling a wind turbine having three rotor blades. The method comprises obtaining an edge load signal comprising three signal components each indicative of edge loading on a respective one of the rotor blades in a rotor rotational frame of the wind turbine. The edge load signal is based on edge load sensor measurements from one or more of the rotor blades. The method comprises removing, from the obtained edge load signal, forward whirling frequency content corresponding to forward whirling motion in an edgewise direction of the rotor blades, to obtain a backward whirling edge signal. The method comprises applying a transformation operation to the backward whirling edge signal to obtain a transformed backward whirling edge signal in a rotor azimuthal fixed frame of the wind turbine. The method comprises generating first and second backward whirling edge signals in the rotor azimuth fixed frame from the transformed backward whirling edge signal. The first backward whirling edge signal corresponds to disturbances at an edge frequency in the rotor rotational frame. The second backward whirling edge signal corresponds to disturbances at the edge frequency minus 2P in the rotor rotational frame. The method comprises applying a first control action to the first backward whirling edge signal to obtain a first backward whirling edge control signal for mitigating backward whirling motion in the edgewise direction of the rotor blades influenced by disturbances at the edge frequency in the rotor rotational frame. The method comprises applying a second control action to the second backward whirling edge signal to obtain a second backward whirling edge control signal for mitigating backward whirling motion in the edgewise direction of the rotor blades influenced by disturbances at the edge frequency minus 2P in the rotor rotational frame. The method comprises applying an inverse m-blade coordinate transformation to the first and second backward whirling edge control signals to obtain individual pitch reference offset values for the respective rotor blades in the rotor rotational frame. The method comprises adjusting pitch of the rotor blades based on the obtained individual pitch reference offset values.

[0025] According to another aspect of the invention there is provided a non-transitory, computer- readable storage medium storing instructions thereon that when executed by one or more processors cause the one or more processors to perform the method described above.

[0026] BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Examples of the invention will now be described with reference to the accompanying drawings, in which:

[0028] Figure 1 schematically illustrates a wind turbine in accordance with an aspect of the invention;

[0029] Figure 2 schematically illustrates a controller of the wind turbine of Figure 1 in accordance with an aspect of the invention;

[0030] Figure 3 schematically illustrates control blocks implemented by the controller of Figure 2 in accordance with examples of the invention;

[0031] Figure 4 summarises the steps of a method performed by the controller of Figure 2 in accordance with the invention; and,

[0032] Figures 5(a)-(e) illustrates simulation results obtained by implementing the method according to Figure 4, compared against simulation results obtained by implementing previous known methods.

[0033] DETAILED DESCRIPTION

[0034] Figure 1 illustrates, in a schematic view, an example of a wind turbine 10. The wind turbine 10 includes a tower 102, a nacelle 103 disposed at the apex of, or atop, the tower 102, and a rotor 104 operatively coupled to a generator housed inside the nacelle 103. In addition to the generator, the nacelle 103 houses other components required for converting wind energy into electrical energy and various components needed to operate, control, and optimise the performance of the wind turbine 10. The rotor 104 of the wind turbine 10 includes a central hub 105 and three rotor blades 106 that project outwardly from the central hub 105. Moreover, the wind turbine 10 comprises a control system including one or more controllers (not shown in Figure 1). The controller(s) may be placed inside the nacelle 103, in the tower 102 or distributed at a number of locations inside (or externally to) the turbine 10 and communicatively connected to one another. The rotor blades 106 are pitch-adjustable. The rotor blades 106 can be adjusted in accordance with a collective pitch setting, where each of the blades are set to the same pitch value. In addition, the rotor blades 106 are adjustable in accordance with individual pitch settings, where each blade 6 may be provided with an individual pitch setpoint.

[0035] In the described example, the wind turbine 10 includes blade load sensors 107 placed at each blade root in a manner such that the sensor 107 detects loading in the blade 106. Blade load signals from such sensors may be used to determine how to adjust the pitch of each of the individual blades 106. The blade load sensors 107 are configured to detect loading in the edge (edgewise) direction, i.e. in the rotor plane. In some examples, the blade load sensors 107 may also be configured to detect loading in the flap (flapwise) direction, i.e. in / out of the rotor plane. The blade load sensors 107 may be strain gauge sensors or optical Bragg-sensors, for instance. As the sensors are placed on the rotating blades 106, such load signals for each of the adjustable rotor blades 106 are measured in a rotating reference frame of the rotor 104 (rotor rotational frame).

[0036] Figure 2 schematically illustrates an example of a control system or overall controller 20 of the wind turbine 10 that includes a (feedback) speed controller or control block 202 implemented to determine collective pitch actuation signals for controlling pitch of the rotor blades 106. In the illustrated implementation, the speed controller 202 minimises a speed error (speedrot - speeds) between the actual rotor speed (speedrot) and a reference rotor speed (speedref) in order to output a requested power P (in the form of a power setpoint) and a collective pitch reference, 0COi. The collective pitch reference as determined by the speed controller 202, in view of the rotor speed, may also take further sensor values into account. This is referred to in Figure 2 as a measurement set, ms, being input into the speed controller 202. The feedback speed controller 202 may be implemented by a PI (proportional-integral), PID (proportional-integral-derivative), or similar control scheme. In one example, the speed controller 202 may alternatively be a model predictive controller which, based on minimising a cost function, is arranged to determine the collective pitch reference and / or the power reference.

[0037] Figure 2 further illustrates a control block / module or controller 204, of the overall controller 20. In the controller 204, pitch modification signals or pitch reference offset values (A0i, A02, A0S) are determined based on one or more input signals 205. In the described example, the input signals 205 include measured blade load signals from the blade load sensors 107, specifically blade loading in the edgewise direction of the rotor blades. The pitch reference offset values (A01, A02, A0s) are superimposed onto the collective pitch reference 9COito provide resulting or overall pitch modification signals (0A, 0B, 0C) that can be applied to pitch actuators of the rotor blades 106 individually.

[0038] The controller 204 is in particular a cyclic or individual blade pitch controller for targeting / controlling edgewise whirling content caused by edgewise vibrations of the rotor blades 106. When the rotor of a wind turbine is turning, oscillations of the blades relative to their edgewise axes can cause movement of the blade in the same plane as the plane of rotation of the rotor. It will be appreciated that the rotor shaft is effectively mounted at one of its ends and is unsupported at the hub end where the blades are attached. As edgewise oscillation of the blades excites the rotor with a force that is transverse to its longitudinal axis, then in resonant conditions this may result in the rotational axis of the rotor shaft describing an erratic pattern of motion. It is this phenomenon that may be referred to as ‘edgewise whirling’ or, simply, ‘whirling’. Oscillation in whirling modes can increase wear on various wind turbine components, and is therefore undesirable.

[0039] The phase differences between the edgewise oscillations of the blades determines whether whirling occurs in the same direction as the rotor rotation, which may be referred to as ‘forward whirl’ or a ‘forward whirling mode’, or whether whirling occurs in a direction opposite to that of the rotor rotation, which may be referred to as ‘backward whirl’ or a ‘backward whirling mode’.

[0040] Whirling motion - in particular, both forward and backward whirling motion - can be influenced by rotor disturbances at different frequencies. This is because of the specific dynamics of the wind turbine nacelle 103. For instance, this could result from an asymmetry between dynamics in the tilt and yaw directions, and / or because tilt-yaw signals / moments are not orthogonal (in phase). It is the case that backward whirling motion can be influenced by rotor disturbances at both an edge frequency and an edge frequency minus 2P (‘edge-2P’ frequency). The edge frequency is the frequency of the rotor blade oscillations in the edgewise direction. 2P is twice the rotor rotational frequency.

[0041] This is explained further via the following. Edge loading that corresponds to backward whirling motion may be expressed as: where refers to the edge loading on the three rotor blades 106, |Mfw| refers to the magnitude of the backward whirling content, o)eis the edge frequency, and t is time.

[0042] This may be projected from the rotor rotational frame into a fixed frame (that is based on a rotor azimuth of the wind turbine 10) via application of an m-blade (multi-blade) coordinate transformation in the form of a Coleman transform to obtain: where Md, Mqare mutually orthogonal signals in the fixed frame, and a>rotis the rotational frequency of the rotor 104, i.e. 1 P (where a>rott = p is the phase).

[0043] The d and q channels / directions in the fixed frame are related to the tilt and yaw directions of the wind turbine 10 (or the wind turbine rotor 104). This relation may be characterised by a transfer function matrix as follows: where Gd^tutis a transfer function from the d direction to the tilt direction, etc., and MtM, Myaware (mutually orthogonal) signals in the tilt and yaw directions.

[0044] It will be appreciated that while whirling motion acts to excite tilt and yaw motion in equal measure, the response of the nacelle 103 along the tilt axis may be different from the response along the yaw axis. In particular, there may be no response (or an insignificant response) in the tilt direction, such that the backward whirling edge loading in the fixed frame may be (or at least assumed to be):

[0045] Fixed frame oscillations of this form then produce content at the edge frequency and at the edge-2P frequency. Considered in reverse, disturbances at the edge-2P frequency in the rotor rotational frame can excite the backward whirling frequency in the fixed frame, which then as a result of the asymmetry between tilt and yaw outlined above, can have an influence on the rotor at the edge frequency, exciting the backward whirling motion.

[0046] It can be shown in a corresponding manner that forward whirling motion can be influenced by rotor disturbances at both the edge frequency and the edge frequency plus 2P (‘edge+2P’ frequency), where the latter term again arises because of the asymmetry of response in the tilt and yaw directions.

[0047] The present invention relates to mitigating edgewise whirling content caused by edgewise vibrations of the rotor blades 106. The present invention is advantageous in that individual blade pitch control is determined and implemented to target disturbances at each of the different frequencies that cause whirling motion. In this way, whirling motion can be fully addressed, thereby leading to a greater reduction in wind turbine component wear caused by such whirling motion. This contrasts with some previous approaches, in which only part of the underlying cause of whirling motion is targeted.

[0048] The present invention is furthermore advantageous in that frequency content at each of the different frequencies to be targeted is isolated effectively - both from each other, and from other frequency content in the received sensor signal - meaning that appropriate control actions to target each frequency act on the appropriate part of the frequency content, resulting in more successful mitigation of the whirling motion. This is achieved by performing appropriate signal transformations to split I pull apart different components of a signal in a manner that facilitates isolation of certain frequency content. This is discussed in greater detail below. The present approach contrasts with some previous approaches that rely on the accuracy of certain filters, e.g. notch filters, to isolate target frequency content in a frequency signal, which may be close to other frequency content in the signal that it is not desired / necessary to act on. The present approach also contrasts with some previous approaches that rely on accurate knowledge of the edge frequency, which can be difficult to estimate as it is a function of the operating point of the wind turbine.

[0049] Figure 3 schematically illustrates a control strategy to be implemented by the controller 204 in accordance with examples of the invention, the control strategy comprising implementation of a number of control blocks / modules as described herein. As mentioned above, the controller 204 is configured to receive measurements of blade edge loading, from the blade load sensor 107 of each of the three rotor blades 106, in the form of a blade edge load sensor signal Mabccomprising the respective blade edge load measurements Ma, Mb, Mc. These measurements are obtained in the rotor rotational frame (or abc frame).

[0050] In the described example, a high-pass filter 31 is applied to the received sensor signal 205 to remove a OP component therein. In the described example, the resulting signal is set at block 32 as a blade edge load signal M^bcto be used in the subsequent analysis, where the blade edge load signal M^bcis a three-dimensional vector with components Ma,Mb,Mb. In different examples, the control block 32 may estimate one or more of the blade edge load components M , Mb, Mb. For instance, if a blade load measurement is not received from one (or more) of the blade load sensors 107, e.g. because of a sensor fault, then then corresponding blade edge load component to be used in the blade edge load signal M^bcmay be estimated based on the blade edge load measurements received from the other blade load sensors 107 and / or from one or more other sensors of the wind turbine 10. Such estimation may be performed in any suitable manner known to the skilled person.

[0051] At control block 33, the controller 204 is configured to apply an m-blade coordinate transformation in the form of a Coleman transformation to the edge load signal M^bc. As is apparent from the above discussion, the Coleman transformation takes the three rotating signals into a fixed reference frame along a first reference direction d and a second reference direction q. In particular, application of the Coleman transformation provides the fixed frame signal / vector Mq, which comprises the mutually orthogonal components The transformation makes it possible to distinguish backward whirling from forward whirling, both being different manifestations of the edge 1stmode.

[0052] When the phase i = a)rott in the Coleman transformation is set as the rotor azimuth, backward whirling appears at the edge frequency minus 1 P (‘edge-1 P’), while forward whirling appears at the edge frequency plus 1 P (‘edge+1 P’). In such a case, the fixed frame may therefore be referred to as a rotor azimuth fixed frame. The edge frequency may be estimated via a look-up table or similar in a memory module. As mentioned above, however, the edge frequency may be specific to a specific wind turbine structure but change for different operating points of the wind turbine.

[0053] At control block 34, a further filter, e.g. a high-pass filter, is applied to the obtained signal Mfg in the rotor azimuth fixed frame in order to remove a component of the signal resulting from gravity. The resulting signal Mdqincludes only frequency content relating to backward and forward whirling motion. In order to determine and apply appropriate control actions for targeting backward and / or forward whirling motion, the frequency content relating to backward whirling motion needs to be isolated from the frequency content relating to forward whirling motion in the signal Mdq. However, as mentioned above, the backward and forward whirling appear in the rotor azimuth fixed frame at edge-1 P and edge+1 P frequencies, respectively. This means that the frequency content is separated only by 2P, which is a relatively small gap to be able to effectively isolate one of the components from the other, e.g. with a bandpass filter.

[0054] In the described example, the controller 204 is therefore beneficially configured to perform another transformation operation to increase the separation between the backward and forward whirling frequency content / components so that one of the components may then be more readily isolated from the other. In particular, the controller 204 implements different branches of control blocks to deal with I target the backward and forward whirling motion separately.

[0055] Considering the branch for targeting backward whirling motion, at control block 35a the controller 204 is configured to apply a transformation operation in the form of a rotation operation to the signal Mdq. In the described example, the rotation operation rotates the signal Mdqabout an angle based on the edge frequency a>e. In particular, a phase shift of (&)e- a)rot)t is applied to shift the signal Mdqfrom the rotor azimuth fixed frame (dq frame) to a backward whirling edge frequency-based (EF-based) fixed frame (dBqBframe).

[0056] In the dBqBframe, the backward whirling content in the signal Mdqis shifted from edge-1 P to OP and the forward whirling content is shifted from edge+1 P to twice the edge frequency. The edge frequency may for instance be around 4P, although this will vary in dependence on the specific wind turbine and operating point as mentioned above. Nonetheless, it is apparent that this shifting operation 35a increases the separation of the backward and forward whirling content.

[0057] It is noted that in different examples, instead of shifting the signal Mdqdirectly from the dq frame to the dBqBframe, the signal Mdqmay first be transformed back to the abc frame via application of an inverse Coleman transformation, and then transformed from the abc frame to the dBqBframe via application of an m-blade transformation having a phase based on the edge frequency.

[0058] It is further noted that although the phase shift at control block 35a is based on the edge frequency, in different examples the phase shift may be based on a different value. In particular, the rotation angle for this phase shift I rotation operation may be set such that sufficient separation of the backward and forward whirling components is obtained. This separation is maximised when the angular frequency is set to the edge frequency (in radians per second), as in the example described herein. However, the rotation angle may be set to any suitable value, e.g. 4P, that provides sufficient frequency separation of the backward and forward whirling components.

[0059] Returning to Figure 3, the forward whirling content may then readily be removed from the resulting signal in the dBqBframe to obtain a backward whirling edge signal, i.e. a signal including only backward hiding frequency content. In the described example, a low pass filter is applied at control block 36a to remove the forward whirling content, and then at control block 37a an inverse backward whirling EF-based transformation operation is applied to return the backward whirling edge signal from the dBqBframe to the dq frame to obtain the signal / vector M ^, including the mutually orthogonal components Mdw, Mqw.

[0060] The signal Mdincludes backward whirling frequency content arising from both edge frequency disturbances and edge-2P disturbances, as explained above. In the present invention, different control actions are beneficially applied via different pitching frequencies for targeting the backward whirling motion caused by the respective different disturbance frequencies.

[0061] To achieve this, the following two vectors in the dq frame are defined: The first of these (the ‘edge vector’) generates a cyclic pitch action at the edge frequency and the second (the ‘edge-2P vector’) generates a cyclic pitch action at edge-2P. It is seen that summing these two vectors would give a non-zero d component at the backward whirling frequency and a zero q component at the backward whirling frequency, thereby capturing the asymmetry of the tilt and yaw dynamics as explained above.

[0062] By expressing this in the form of these two vectors, rather than a single vector with only a non-zero d component, the control actions applied to each resulting pitching frequency (i.e. edge frequency and edge-2P) can be controlled separately.

[0063] Figure 3 illustrates a first sub-branch of control blocks that act on the first (edge) vector for backward whirling motion. In particular, at control block 381a a gain kBW1is applied to the first vector for mitigating backward whirling motion influenced by disturbances at the edge frequency. At control block 391a, a phase shift 8BW1is applied, where 8BW1is the target phase offset for the cyclic pitch signal at the edge frequency that mitigates backward whirling motion. In the described example, at control block 401a the controller 204 is configured to apply an amplitude-limiting factor to the resulting signal, where the amplitude-limiting factor may be defined based on amplitude saturation for cyclic pitch saturation at the edge frequency for backward whirling control. Finally, an inverse m-blade transformation in the form of an inverse Coleman transformation is applied at control block 411a to obtain a set of individual pitch modification signals 'nthe rotor rotational frame for mitigating backward whirling motion cause by edge frequency disturbances.

[0064] Figure 3 also illustrates a second sub-branch of control blocks, which act on the second (edge-2P) vector for backward whirling motion. In a corresponding manner to the first subbranch outlined above, in the second sub-branch control blocks 382a, 392a, 402a, 412a the following actions are respectively performed: a gain kBW2for mitigating backward whirling motion influenced by disturbances at the edge-2P frequency is applied; a phase shift 8BW2is applied (where 8BW2is the target phase offset for the cyclic pitch signal at the edge-2P frequency that mitigates backward whirling motion); an amplitude-limiting factor is applied to the resulting signal (where the amplitude-limiting factor may be defined based on amplitude saturation for cyclic pitch saturation at the edge-2P frequency for backward whirling control); and, a Coleman transformation is applied to obtain a set of individual pitch modification signals Gabc2in the rotor rotational frame for mitigating backward whirling motion caused by edge-2P frequency disturbances. The two sets of individual pitch modification signals for mitigating backward whirling motion abe1* Gabc2may then be combined, e.g. added together, and in an example in which backward whirling motion only is to be mitigated, the controller 204 may output the combined signal as the pitch modification signals A6>123 to be combined with the collective pitch reference from the speed controller 202.

[0065] In the control strategy illustrated in Figure 3, forward whirling content is also to be mitigated. Referring back to the output signal Mdqfrom the control block 34, a separate branch of control blocks is implemented for targeting forward whirling motion. At control block 35b the controller 204 is configured to apply a rotation operation to the signal Mdqabout an angle based on the negative edge frequency -o)e. In particular, a phase shift of (&)e+ a)rot)t is applied to shift the signal Mdqfrom the rotor azimuth fixed frame (dq frame) to a forward whirling edge frequency-based (EF-based) fixed frame (dFqFframe).

[0066] In a corresponding manner to control block 36a, at control block 36b a low pass filter is applied to this time remove backward whirling content from the obtained signal, before applying an inverse forward whirling EF-based transformation operation to return the forward whirling edge signal from the dFqFframe to the dq frame. This obtains a signal / vector Mdthat includes the mutually orthogonal components Mdw, Mqw.

[0067] In a corresponding manner to the backward whirling motion signal, the signal Mdincludes forward whirling frequency content arising from both edge frequency disturbances and edge+2P disturbances. In order to be able to apply separate control actions for targeting the forward whirling motion caused by the respective different disturbance frequencies, two vectors in the dq frame are defined are defined in a corresponding manner to that described for backward whirling frequency above. A first of these vectors (the ‘edge vector’) generates a cyclic pitch action at the edge frequency and the second (the ‘edge+2P vector’) generates a cyclic pitch action at edge+2P.

[0068] Figure 3 illustrates a first sub-branch of control blocks that act on the first (edge) vector for forward whirling motion. In a corresponding manner to the sub-branches for backward whirling motion outlined above, in the first sub-branch control blocks 381b, 391 b, 401b, 411b for forward whirling motion the following actions are respectively performed: a gain k jyi for mitigating forward whirling motion influenced by disturbances at the edge frequency is applied; a phase shift 8FW1is applied (where 8FW1is the target phase offset for the cyclic pitch signal at the edge frequency that mitigates forward whirling motion); an amplitude-limiting factor is applied to the resulting signal (where the amplitude-limiting factor may be defined based on amplitude saturation for cyclic pitch saturation at the edge frequency for forward whirling control); and, a Coleman transformation is applied to obtain a set of individual pitch modification signals A0™1in the rotor rotational frame for mitigating forward whirling motion caused by edge frequency disturbances.

[0069] In a further corresponding manner, in the second sub-branch control blocks 382b, 392b, 402b, 412b for forward whirling motion the following actions are respectively performed: a gain kFW2for mitigating forward whirling motion influenced by disturbances at the edge+2P frequency is applied; a phase shift 8FW2is applied (where 8FW2is the target phase offset for the cyclic pitch signal at the edge+2P frequency that mitigates forward whirling motion); an amplitude-limiting factor is applied to the resulting signal (where the amplitude-limiting factor may be defined based on amplitude saturation for cyclic pitch saturation at the edge+2P frequency for forward whirling control); and, a Coleman transformation is applied to obtain a set of individual pitch modification signals A0™2in the rotor rotational frame for mitigating forward whirling motion caused by edge+2P frequency disturbances.

[0070] The two sets of individual pitch modification signals for mitigating forward whirling motion ^abc2rnay then be combined, and in an example in which both backward and forward whirling motion is to be mitigated, the controller 204 may then determine an overall combined signal - using each of the combined backward and forward whirling pitch offsets - as the pitch modification signals A6>123.

[0071] Figure 4 summarises the steps of a method 100 performed by the controller 204 in accordance with an example of the invention. At step 1001 , the method 100 involves obtaining an edge load signal that includes three signal components each indicative of edge loading on a respective one of the wind turbine rotor blades 106 in a rotor rotational frame of the wind turbine 10. The edge load signal may be edge load sensor measurements from the three rotor blades 106.

[0072] At step 1002, the method 100 involves removing forward whirling (FW) frequency content from the edge load signal obtained in step 1001 , to obtain a backward whirling (BW) edge signal. Optionally, this may involve applying a transformation operation to obtain a transformed edge signal in a fixed frame (e.g. the dBqBframe) of the wind turbine 10. An angular frequency of a rotation angle used in the transformation operation may be (an estimate of) the edge frequency or any other suitable frequency / value that allows for sufficient separation of the BW and FW frequency content in the fixed frame signal. A low pass filter may then be applied to the transformed edge signal to remove the FW frequency content.

[0073] At step 1003, the method 100 involves applying a transformation operation to the BW edge signal to obtain a transformed BW edge signal in a fixed frame (e.g. the dq frame) of the wind turbine 10. In some cases, the transformation operation may correspond to a phase shift about a rotation angle having an angular frequency equal to the edge frequency minus 1P.

[0074] At step 1004, the method 100 involves generating first and second BW edge signals / vectors in the fixed frame from the transformed BW edge signal. The first BW edge signal corresponds to disturbances at the edge frequency in the rotor rotational frame. The second BW edge signal corresponds to disturbances at the edge-2P frequency in the rotor rotational frame. This may involve defining vectors as described in the above example, i.e. an ‘edge vector’ and an ‘edge-2P vector’ that respectively generate a cyclic pitch action at the edge frequency and at edge-2P. Optionally, summing these two signals / vectors results in one component of the vector being zero.

[0075] At step 1005, the method 100 involves applying a first BW control action to the first BW edge signal to obtain a first BW edge control signal for mitigating BW motion in the edgewise direction of the rotor blades influenced by disturbances at the edge frequency in the rotor rotational frame. The first BW control action may include application of one or both of a first controller gain and a first phase offset.

[0076] Also at step 1005, the method 100 involves applying a second BW control action to the second BW edge signal to obtain a second BW edge control signal for mitigating BW motion in the edgewise direction of the rotor blades influenced by disturbances at the edge- 2P frequency in the rotor rotational frame. The second BW control action may include application of one or both of a second controller gain and a second phase offset.

[0077] At step 1006, the method 100 involves applying an inverse m-blade coordinate transformation, e.g. Coleman transformation, to the first and second BW edge control signals to obtain individual BW pitch reference offset values for the respective rotor blades in the rotor rotational frame. This may involve applying a first and second inverse m-blade coordinate transformations to the respective first and second BW edge signals to obtain first and second individual BW pitch reference offset values for the respective rotor blades in the rotor rotational frame. These may then be combined, e.g. summed, to obtain the individual BW pitch reference offset values.

[0078] At step 1007, the method 100 involves transmitting a control signal to adjust pitch of the rotor blades based on the obtained individual BW pitch reference offset values. This may involve combining the individual BW pitch reference offset values with a collective pitch reference from a speed controller of the wind turbine, and outputting the resulting signal dABCto a pitch actuator of the wind turbine 10, which then actuates the pitch of the rotor blades 106 in accordance with the determined individual pitch references.

[0079] In the above-described method, only backward whirling motion in an obtained edge signal is targeted. However, it will be understood that forward whirling motion may alternatively or additionally be targeted. In either case, it may be that forward whirling motion as a result of edge frequency disturbances only is targeted, or it may be that forward whirling motion as a result of edge frequency and edge+2P disturbances is targeted. In some cases, the relatively high frequency pitching at edge+2P may not be included for reasons of excessive pitch bearing wear and / or because forward whirling may in some cases not be as much of an issue as backward whirling.

[0080] Figure 5 shows indicative and comparative simulation results of the described approach. In particular, Figures 5(a)-(e) show plots of rotor blade bearing fatigue, blade root edge fatigue, minimum edgewise extreme loading, maximum edgewise extreme loading, and edgewise backward whirling content, respectively, at a characteristic free wind speed for three different wind turbine control setups. A first control setup corresponds to an edgewise controller that includes a combination of cyclic blade pitch at the edge frequency for countering backward whirling motion, and cyclic blade pitch at the edge frequency for countering forward whirling motion. A second control setup is the same as the first but where a 2P cyclic pitch controller of the wind turbine is disabled. A third control setup is the same as the second but where implementation of an additional pitching action at the edge-2P frequency as described herein is provided. It may be seen that the third control setup - i.e. in accordance with the approach described herein - results in the lowest of each of the rotor blade bearing fatigue 51c, blade root edge fatigue 52c, maximum edgewise extreme loading 54c and edgewise backward whirling content 55c, and also the highest minimum edgewise extreme loading 53c, out of the three control setups considered. Figures 5(b), 5(d) and 5(e) show that that the 2P cyclic pitching controller can be useful as disabling it can result in increased blade edge fatigue (from point 52a to point 52b), increased edgewise extreme maximum loading (from point 54a to point 54b), and increased edgewise backward whirling content (from point 55a to point 55b). However, it may also be seen in Figures 5(a) that the 2P cyclic pitching controller being enabled can significantly increase blade bearing fatigue (from point 51b to point 51a). The present approach - i.e. points 51c-55c - is beneficial as it reduces blade fatigue / loading while not increasing blade bearing fatigue.

[0081] Many modifications may be made to the described examples without departing from the scope of the appended claims.

[0082] The described controller 20 may be in the form of any suitable computing device, for instance 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 customer processors and memory. The one or more functional units may use a common computing substrate (for example, they may run on the same server) or separate substrates, or one or both may themselves be distributed between multiple computing devices. A computer memory may store instructions for performing the methods performed by the controller, and the processor(s) may execute the stored instructions to perform the method.

[0083] Although the above description generally refers to a Coleman transformation (or inverse Coleman transformation) for transforming between rotational and fixed reference frames, it will be understood that other suitable m-blade transformations may be used.

Claims

CLAIMS1. A controller for a wind turbine having three rotor blades, the controller being for adjusting individual pitch of the three rotor blades, the controller being configured to: obtain an edge load signal comprising three signal components each indicative of edge loading on a respective one of the rotor blades in a rotor rotational frame of the wind turbine, the edge load signal being based on edge load sensor measurements from one or more of the rotor blades; remove, from the obtained edge load signal, forward whirling (FW) frequency content corresponding to FW motion in an edgewise direction of the rotor blades, to obtain a backward whirling (BW) edge signal; apply a rotor azimuth-based transformation operation to the BW edge signal to obtain a transformed BW edge signal in a rotor azimuth fixed frame of the wind turbine; generate first and second BW edge signals in the rotor azimuth fixed frame from the transformed BW edge signal, wherein the first BW edge signal corresponds to disturbances at an edge frequency in the rotor rotational frame, and wherein the second BW edge signal corresponds to disturbances at the edge frequency minus 2P in the rotor rotational frame; apply a first BW control action to the first BW edge signal to obtain a first BW edge control signal for mitigating BW motion in the edgewise direction of the rotor blades influenced by disturbances at the edge frequency in the rotor rotational frame; apply a second BW control action to the second BW edge signal to obtain a second BW edge control signal for mitigating BW motion in the edgewise direction of the rotor blades influenced by disturbances at the edge frequency minus 2P in the rotor rotational frame; apply an inverse m-blade coordinate transformation to the first and second BW edge control signals to obtain individual BW pitch reference offset values for the respective rotor blades in the rotor rotational frame; and, transmit a control signal to adjust pitch of the rotor blades based on the obtained individual BW pitch reference offset values.

2. A controller according to Claim 1 , wherein to remove the FW frequency content from the obtained edge load signal, the controller is configured to: apply an edge frequency-based (EF-based) transformation operation to the obtained edge load signal to obtain an EF-based transformed edge signal in an EF-basedfixed frame, wherein an angular frequency of a rotation angle used in the EF-based transformation operation is the edge frequency; and, apply a low pass filter to the EF-based transformed edge signal to remove the FW frequency content.

3. A controller according to Claim 2, wherein the signal in the EF-based fixed frame obtained from the low pass filter is the BW edge signal, and wherein the rotor azimuthbased transformation operation corresponds to a rotation operation about a rotation angle having an angular frequency equal to 1P minus the edge frequency.

4. A controller according to any previous claim, wherein, prior to removing the FW frequency content from the obtained edge load signal, the controller is configured to: apply an m-blade coordinate transformation to the obtained edge load signal to obtain mutually orthogonal signals in the rotor azimuth fixed frame; apply a high pass filter to remove gravity-based content in the mutually orthogonal signals.

5. A controller according to Claim 4 when dependent on Claim 2, wherein the EF-based transformation operation is applied to the signal in the rotor azimuth fixed frame obtained from the high pass filter, and wherein the EF-based transformation operation corresponds to a phase shift about a rotation angle having an angular frequency equal to the edge frequency minus 1 P.

6. A controller according to any previous claim, wherein to obtain the edge load signal, the controller is configured to apply a further high pass filter to the edge load sensor measurements to remove OP frequency content.

7. A controller according to any previous claim, wherein to apply the first BW control action to the first backward whirling edge signal, the controller is configured to phase shift the first backward whirling edge signal about a first phase offset angle for mitigating backward whirling motion influenced by disturbances at the edge frequency, and / or wherein to apply the second BW control action to the second backward whirling edge signal, the controller is configured to phase shift the second backward whirling edge signal about a second phase offset angle for mitigating backward whirling motion influenced by disturbances at the edge frequency minus 2P.

8. A controller according to any previous claim, wherein to apply the first BW control action the controller is configured to apply a first gain to the first backward whirling edge signal, and / or wherein to apply the second BW control action the controller is configured to apply a second gain to the second backward whirling edge signal.

9. A controller according to any previous claim, wherein the controller is configured to apply a first amplitude-limiting factor to the first backward whirling edge control signal prior to the inverse m-blade coordinate transformation being applied, and / or wherein the controller is configured to apply a second amplitude-limiting factor to the second backward whirling edge control signal prior to the inverse m-blade coordinate transformation being applied.

10. A controller according to any previous claim, wherein to apply the inverse m-blade coordinate transformation to obtain the individual BW pitch reference offset values, the controller is configured to: apply a first inverse m-blade coordinate transformation to the first BW edge signal to obtain first individual BW pitch reference offset values for the respective rotor blades in the rotor rotational frame; apply a second inverse m-blade coordinate transformation to the second BW edge signal to obtain second individual BW pitch reference offset values for the respective rotor blades in the rotor rotational frame; and, combine the first individual BW pitch reference offset values and the second BW individual pitch reference offset values to obtain the individual BW pitch reference offset values.

11. A controller according to any previous claim, the controller being configured to: remove, from the obtained edge load signal, BW frequency content corresponding to BW motion in the edgewise direction of the rotor blades, to obtain a FW edge signal; apply a FW rotor azimuth-based transformation operation to the FW edge signal to obtain a transformed FW edge signal in the rotor azimuth fixed frame; generate a first FW edge signal in the rotor azimuth fixed frame from the transformed FW edge signal, wherein the first FW edge signal corresponds to disturbances at the edge frequency in the rotor rotational frame; apply a first FW control action to the first FW edge signal to obtain a first FW edge control signal for mitigating FW motion in the edgewise direction of the rotor blades influenced by disturbances at the edge frequency in the rotor rotational frame;apply a first FW inverse m-blade coordinate transformation to the first FW edge control signal to obtain first FW individual pitch reference offset values for the respective rotor blades in the rotor rotational frame; and, combine the first FW individual pitch reference offset values with the BW individual pitch reference offset values to obtain individual pitch reference offset values for the respective rotor blades in the rotor rotational frame, wherein the control signal to adjust pitch of the rotor blades based on the obtained individual pitch reference offset values.

12. A controller according to Claim 11 , the controller being configured to: generate a second FW edge signal in the rotor azimuth fixed frame from the transformed FW edge signal, wherein the second FW edge signal corresponds to disturbances at the edge frequency plus 2P in the rotor rotational frame; apply a second FW control action to the second FW edge signal to obtain a second FW edge control signal for mitigating FW motion in the edgewise direction of the rotor blades influenced by disturbances at the edge frequency plus 2P in the rotor rotational frame; apply a second FW inverse m-blade coordinate transformation to the second FW edge control signal to obtain second FW individual pitch reference offset values for the respective rotor blades in the rotor rotational frame; and, combine the second FW individual pitch reference offset values with the first FW individual pitch reference offset values and the BW individual pitch reference offset values to obtain the individual pitch reference offset values for the respective rotor blades in the rotor rotational frame.

13. A controller according to Claim 11 or Claim 12, wherein to remove the BW frequency content from the obtained edge load signal, the controller is configured to: apply a second edge frequency-based (EF-based) transformation operation to the obtained edge load signal to obtain a second EF-based transformed edge signal in a second EF-based fixed frame, wherein an angular frequency of a rotation angle used in the second EF-based transformation operation is negative edge frequency; and, apply a second low pass filter to the second EF-based transformed edge signal to remove the BW frequency content.

14. A wind turbine comprising a controller according to any previous claim.

15. A method for controlling a wind turbine having three rotor blades, the method comprising: obtaining an edge load signal comprising three signal components each indicative of edge loading on a respective one of the rotor blades in a rotor rotational frame of the wind turbine, the edge load signal being based on edge load sensor measurements from one or more of the rotor blades; removing, from the obtained edge load signal, forward whirling frequency content corresponding to forward whirling motion in an edgewise direction of the rotor blades, to obtain a backward whirling edge signal; applying a transformation operation to the backward whirling edge signal to obtain a transformed backward whirling edge signal in a rotor azimuthal fixed frame of the wind turbine; generating first and second backward whirling edge signals in the rotor azimuth fixed frame from the transformed backward whirling edge signal, wherein the first backward whirling edge signal corresponds to disturbances at an edge frequency in the rotor rotational frame, and wherein the second backward whirling edge signal corresponds to disturbances at the edge frequency minus 2P in the rotor rotational frame; applying a first control action to the first backward whirling edge signal to obtain a first backward whirling edge control signal for mitigating backward whirling motion in the edgewise direction of the rotor blades influenced by disturbances at the edge frequency in the rotor rotational frame; applying a second control action to the second backward whirling edge signal to obtain a second backward whirling edge control signal for mitigating backward whirling motion in the edgewise direction of the rotor blades influenced by disturbances at the edge frequency minus 2P in the rotor rotational frame; applying an inverse m-blade coordinate transformation to the first and second backward whirling edge control signals to obtain individual pitch reference offset values for the respective rotor blades in the rotor rotational frame; and, adjusting pitch of the rotor blades based on the obtained individual pitch reference offset values.

Images