Determining rotor azimuth angle of a wind turbine

Abstract

The invention relates to determining rotor azimuth angle of a wind turbine. A rotor speed of the wind turbine is obtained, a gain is applied thereto to obtain a gain-adjusted rotor speed signal, and a rotor speed integrator is applied to the gain-adjusted rotor speed signal to determine rotor azimuth angle. The invention involves receiving an acceleration sensor signal, from an acceleration sensor located in a rotor hub of the wind turbine, indicative of gravitational acceleration of the rotor hub relative to a rotation axis of the rotor hub, and determining a reference rotor azimuth angle based on the received acceleration sensor signal. The invention involves determining an azimuth angle error between the reference rotor azimuth angle and the rotor azimuth angle determined by the rotor speed integrator, and determining the gain to be applied to the obtained rotor speed signal based on the determined azimuth angle error.

Description

[0001] DETERMINING ROTOR AZIMUTH ANGLE OF A WIND TURBINE

[0002] TECHNICAL FIELD

[0003] The invention relates to determining rotor azimuth angle of a wind turbine. In particular, a reference rotor azimuth angle is determined based on an acceleration sensor signal from an acceleration sensor located in a rotor hub of the wind turbine, and an error between said reference rotor azimuth angle and a rotor azimuth angle obtained from a rotor speed integrator is used to determine a gain to be applied to a rotor speed signal being input into the rotor speed integrator.

[0004] BACKGROUND

[0005] A wind turbine typically incudes a tower, a nacelle atop the tower for housing one or more electrical components (e.g. converter), a rotor that rotates relative to the nacelle and tower, and a number of rotor blades (typically, three rotor blades) attached to the rotor. The rotational position of a wind turbine rotor - also referred to as the rotor azimuth angle - can be used for various different functions / purposes. These include being used for positioning the rotor and rotor blades during service or maintenance operations of the wind turbine, for calibrating other sensing systems of the wind turbine, and as an input into one or more control routines / process of the wind turbine, e.g. a load reducing control routine. It will be understood, therefore, that it is important to be able to obtain accurate determinations of the rotor azimuth angle of a wind turbine.

[0006] A known approach for obtaining wind turbine rotor azimuth angle is to apply an integrator function / module to an obtained rotor speed signal of the wind turbine, i.e. integrating the obtained speed of the wind turbine rotor with respect to time. The rotor speed signal may be obtained in any suitable manner, e.g. based on a signal obtained from a rotary encoder that measures a speed of a generator of the wind turbine. Such an approach is prone to capture any errors in the rotor speed measurement, accumulating these errors over time.

[0007] A known method for mitigating these accumulated errors is to include a dedicated reset sensor in the wind turbine for sensing when the wind turbine rotor starts a new revolution. Upon this being sensed, a signal is generated to reset the integrator so as to minimise errors being accumulated. Such a method can suffer the drawback that the dedicated sensor may not reliably be able to detect a target magnet protruding from a rotor lock disc of the rotor, and so not able to accurately measure when the rotor starts a new revolution. To address this issue, the sensor may be moved closer to the rotor lock disc; however, this can result in the magnet and sensor colliding under uneven loads of the rotor, which can damage or even destroy the magnet and / or sensor.

[0008] There remains a need to provide improved methods for obtaining rotor azimuth speed in an accurate and reliable manner. It is against this background to which the present invention is set.

[0009] SUMMARY OF THE INVENTION

[0010] According to an aspect of the invention there is provided a method of determining rotor azimuth angle of a wind turbine. The method comprises obtaining a rotor speed signal indicative of rotor speed of the wind turbine, applying a gain to the obtained rotor speed signal to obtain a gain-adjusted rotor speed signal or compensated rotor speed signal, and applying a rotor speed integrator to the gain-adjusted rotor speed signal to determine rotor azimuth angle of the wind turbine.

[0011] The method comprises receiving an acceleration sensor signal, from an acceleration sensor located in a rotor hub of the wind turbine, indicative of acceleration, due to gravity, in a radial direction of the rotor hub relative to a rotation axis of the rotor hub. The method comprises determining a reference rotor azimuth angle based on the received acceleration sensor signal. The method comprises determining an azimuth angle error between the reference rotor azimuth angle and the rotor azimuth angle determined by the rotor speed integrator. The method comprises determining the gain-adjusted to be applied to the obtained rotor speed signal based on the determined azimuth angle error.

[0012] The gain may be determined to minimise the azimuth angle error.

[0013] The steps of the method may be performed repeatedly as part of a feedback loop process. The steps of the method may be performed substantially continuously.

[0014] The method comprises, prior to determining the azimuth angle error, applying a normalisation function to the reference rotor azimuth angle and to the rotor azimuth angle determined by the rotor speed integrator. The azimuth angle error may be determined based on the respective normalised rotor azimuth angles. In some examples, determining the gain based on the determined azimuth angle error may comprise: determining a phase delay based on the determined azimuth angle error; and, determining the gain based on the phase delay. The phase delay may be positive or negative.

[0015] The determined azimuth angle error may be a cyclic varying signal. The method may comprise applying a transform to the cyclic varying signal to obtain the phase delay. The phase delay may be a monotonic signal over at least one period of the cyclic varying signal.

[0016] In some examples, applying the transform may comprise: for values of the determined azimuth angle error greater than one, calculating a difference between one and the azimuth angle error; and, for values of the determined azimuth angle error less than one, calculating an absolute value of the determined azimuth angle error.

[0017] A further transform may be applied to the obtained phase delay such that the determined phase delay - to be used to determine the gain to be applied to the obtained rotor speed signal - is in an interval between -0.5 and 0.5.

[0018] The method may comprise applying a filter to the determined phase delay prior to determining the gain to the obtained rotor speed signal. The filter may be configured to reduce disturbances in the received acceleration sensor signal caused by tower oscillations of the wind turbine.

[0019] The received acceleration sensor signal may be indicative of acceleration, due to gravity, of the rotor hub in first and second mutually orthogonal radial directions.

[0020] In some examples, determining the reference rotor azimuth angle may comprise applying a trigonometric function to components of the received acceleration sensor signal in the first and second mutually orthogonal radial directions.

[0021] The reference rotor azimuth angle may be determined in dependence on an orientation and position of the acceleration sensor in the rotor hub.

[0022] The method may comprise applying an offset to the received acceleration sensor signal to compensate for a centrifugal force or centripetal acceleration contribution included therein. The offset may be proportional to the square of the rotor speed or the angular rotor speed in the obtained rotor speed signal. The reference rotor azimuth angle may be determined based on the centrifugal-compensated or centripetal-acceleration-compensated acceleration sensor signal.

[0023] The offset may be applied to the components of the transformed acceleration sensor signal, i.e. the acceleration sensor signal which has been transformed from the rotating reference frame into a stationary reference frame.

[0024] The rotor speed in the obtained rotor speed signal may be determined based on a received generator speed signal indicative of generator speed of the wind turbine. Optionally, the generator speed signal may be received from a rotary encoder of the wind turbine.

[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 processor to perform a method as defined above.

[0026] According to another aspect of the invention there is provided a controller for determining rotor azimuth angle of a wind turbine. The controller is configured to: obtain a rotor speed signal indicative of rotor speed of the wind turbine; apply a gain to the obtained rotor speed signal to obtain a gain-adjusted rotor speed signal; and apply a rotor speed integrator to the gain-adjusted rotor speed signal to determine rotor azimuth angle of the wind turbine.

[0027] The controller is configured to receive an acceleration sensor signal, from an acceleration sensor located in a rotor hub of the wind turbine, indicative of acceleration, due to gravity, in a radial direction of the rotor hub relative to a rotation axis of the rotor hub. The controller is configured to determine a reference rotor azimuth angle based on the received acceleration sensor signal. The controller is configured to determine an azimuth angle error between the reference rotor azimuth angle and the rotor azimuth angle determined by the rotor speed integrator. The controller is configured to determine the gain to be applied to the obtained rotor speed signal based on the determined azimuth angle error.

[0028] According to another aspect of the invention there is a wind turbine comprising a controller as defined above. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Examples of the invention will now be described with reference to the accompanying drawings, in which:

[0030] Figure 1 is a schematic diagram of a wind turbine in accordance with an example of the invention;

[0031] Figure 2 schematically illustrates a controller of the wind turbine of Figure 1 in accordance with an example of the invention;

[0032] Figure 3 shows the steps of a method performed by the controller of Figure 2 in accordance with an example of the invention;

[0033] Figure 4 schematically illustrates a rotor plane of the wind turbine of Figure 1 , as well as example positions of an accelerometer in a rotor hub of the wind turbine of Figure 1 ;

[0034] Figure 5 shows plots of rotor position of the wind turbine of Figure 1 against time as determined using each of: an acceleration signal from the accelerometer of Figure 4; an integrator of the controller of Figure 2 that has drifted from a correct signal; a previous, known method;

[0035] Figure 6 shows normalised versions of the plots of Figure 5 of rotor position of the wind turbine of Figure 1 against time as determined using each of the acceleration signal and the integrator;

[0036] Figure 7 shows a plot of the rotor position error between the two normalised plots of Figure 6 over time;

[0037] Figure 8 a plot of the rotor position error between the two normalised plots of Figure 6 over a longer time period than Figure 7;

[0038] Figure 9 shows a phase delay of the integrator of the controller of Figure 2 determined using the rotor position error of Figure 7; Figure 10 shows a transformed version of the phase delay of Figure 9 to be within a defined range;

[0039] Figure 11 shows the signals of Figure 6 when the method of Figure 3 is applied repeatedly as part of a feedback loop; and,

[0040] Figure 12 is a plot of the phase delay over time determined using the signals of Figure 11 .

[0041] DETAILED DESCRIPTION

[0042] 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.

[0043] The rotor 104 and rotor blades 106 rotate about a rotor axis and define a rotor plane. In particular, wind impinging on the rotor blades 106 drives rotation of the rotor 104 and rotor blades 106, thereby capturing wind energy that can be converted to electrical energy, e.g. for supply to the grid.

[0044] The wind turbine 10 includes an acceleration sensor in the form of an accelerometer 107 located in the rotor hub 105. In the described example, the accelerometer 107 is a three- axis accelerometer configured to detect acceleration of the rotor hub in three mutually orthogonal directions. The orientation and position of the accelerometer 107 within ( / relative to) the rotor hub 105 will be known. The accelerometer 107 may be part of a sensing device I (inertial) measurement unit that includes further sensors. For instance, such a measurement unit may additionally include a (three-axis) gyroscope configured to measure angular motion / velocity of the rotor hub 105.

[0045] The present invention provides a method and controller for determining a rotational position of the wind turbine rotor 104 - hereinafter referred to as the rotor azimuth angle - in an accurate and reliable manner. The invention is advantageous in that it provides for accurate and reliable determinations of rotor azimuth angle without needing to use a dedicated reset sensor for resetting a rotor speed integrator used to determine rotor azimuth angle. The invention beneficially makes use of existing hardware of the wind turbine - namely, the accelerometer 107 in the rotor hub 105 -to determine how to mitigate accumulated errors in the rotor speed integrator. The removal of the need to use a dedicated reset sensor reduces the electrical and mechanical complexity of implementing a rotor azimuth angle determination, which in turn can reduce component costs, reduce maintenance costs / time, and free up space in the rotor hub. The approach of the present invention is more reliable than previous approaches that use a dedicated reset sensor as the risk of said sensor being damaged / destroyed is removed. Further advantages associated with the present invention will become apparent in the following.

[0046] Figure 2 schematically illustrates a controller 20 of the wind turbine 10. The controller 20 is for determining rotor azimuth angle of the wind turbine rotor 104. The determined rotor azimuth angle may be used for any suitable purpose. For instance, the determined rotor azimuth angle may be used for positioning the rotor and rotor blades during service or maintenance operations of the wind turbine, for calibrating other sensing systems of the wind turbine, and as an input into one or more control routines / process of the wind turbine, e.g. a load reducing control routine. The controller 20 may be placed inside the rotor hub 105, in 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.

[0047] The controller 20 is configured to receive a rotor speed signal 201 indicative of a measurement / determination of a (rotational) speed of the wind turbine rotor 104. The rotor speed signal 201 may be obtained from any suitable source. The rotor speed signal may be determined based on a measured speed of a generator of the wind turbine 10. The wind turbine 10 may be provided with a rotary encoder configured to measure generator speed. The rotor speed may then be determined as the measured generator speed divided by a gearbox speed ratio of the wind turbine 10 (generator speed compensated by gear ratio). This may be regarded as a particularly accurate method for obtaining the rotor speed signal 201.

[0048] An alternative method for determining the rotor speed may be based on sensor signals obtained only from the inertial measurement unit of the wind turbine 10, in particular from the accelerometer 107 and the gyroscope. Specifically, a first estimated rotational speed of the rotor hub 105 may be determined based on the angular velocity signal from the gyroscope, and a second estimated rotational speed of the rotor hub 105 may be determined based on the acceleration signal from the accelerometer 107. A correction value may be determined based on a difference between the first and second estimated rotor speeds, and the rotor speed is determined by applying the correction to the first estimated rotor speed. This method may be regarded as providing a reliable and accurate rotor speed measurement that is robust against external noise.

[0049] In some examples, the controller 20 may include a rotor speed selector module. In such examples, the controller 20 may be configured to receive more than one rotor speed signal, in particular from different sources I obtained via different methods (such as the ones outlined above). The selector module may then be configured to select which of the received rotor speed signals is to be used by other modules of the controller 20 to determine rotor azimuth angle (as described below). Beneficially, if the rotor speed signal from one source is not received, e.g. because of a fault, then the rotor speed selector module can select another source to provide the rotor speed signal in order to ensure continued operation of the controller 20. In some examples, the rotor speed selector module may combine rotor speed measurements / estimations from a plurality of sources to obtain a rotor speed signal to be used by the controller 20 to determine rotor azimuth angle, which may increase the accuracy of the determination.

[0050] The rotor speed signal 201 is for input into an integrator module 202 of the controller 20. The integrator 202 is configured to integrate rotor speed with respect to time to obtain a rotor azimuth angle signal 203.

[0051] As outlined above, the integrator 202 can suffer from drift away from a correct / accurate determination of rotor azimuth angle as a result of accumulated errors from the rotor speed signal 201. To address this, the present invention uses acceleration measurements from the accelerometer 107 to compensate for the integrated speed measurement errors in the integrator 202. In particular, the controller 20 includes a reference estimation unit / module 204 for determining a reference rotor azimuth angle signal 205 based on the received acceleration signal 206 from the accelerometer 107, and a gain unit / module 207 for determining a gain signal 208 to be applied to the obtained rotor speed signal 201 , e.g. via a summation function unit / module 209, prior to it being input into the integrator 202. In particular, the gain is determined based on an error between the reference azimuth angle signal 205 and the determined azimuth angle signal 203 output by the integrator 202, as part of a feedback loop process. This will be described in greater detail below. Figure 3 shows the steps of a method 30 performed by the controller 20 to determine rotor azimuth angle of the wind turbine 10. At step 301 , the method 30 involves obtaining the rotor speed signal 201 indicative of rotor speed of the wind turbine 10. As described above, the rotor speed signal may be obtained from any suitable source, and is for input into the rotor speed integrator 202.

[0052] At step 302, the method 30 involves receiving the acceleration sensor signal 206 from the acceleration sensor 107 located in the wind turbine rotor hub 105. The acceleration signal 206 is indicative of acceleration, due to gravity, in a radial direction of the rotor hub 105 relative to a rotation axis of the rotor hub 105. In the described example, it is desired to obtain the acceleration in two mutually orthogonal directions in the rotor plane, i.e. in the radial direction relative to the rotor axis, as this will then be used to determine an indication of the azimuth angle of the rotor. The estimation unit 204 may perform a coordinate transform of the acceleration signal 206 from a rotating reference frame into a stationary reference frame in order to obtain the mutually orthogonal directions in the rotor plane. Thus, the one or more components of the acceleration sensor signal 206 may be transformed from a rotating reference frame rotating with the rotor 104 into a stationary reference frame.

[0053] It is particularly desired to obtain the acceleration caused by gravity in the two mutually orthogonal directions. The accelerometer 107 in the rotor hub 105 is excited primarily by gravity; however, the acceleration signal 206 will also include contributions as a result of centrifugal force as well as rotor rotational speed accelerations and vibrations of the rotor hub 105 and nacelle 103. The method 30 may involve steps to compensate for ( / remove) the centrifugal force contribution in the acceleration signal 206. This may be performed by the estimation unit 204 of the controller 20.

[0054] The centrifugal force contribution induces an offset in the acceleration signal 206 that is proportional to the rotor hub speed squared in the two measurement axes of the accelerometer 107 that are perpendicular to the main shaft of the rotor 104, i.e. the two mutually orthogonal axes in the rotor plane. The method 30 may therefore include a step of applying an offset to the received acceleration sensor signal 206 to compensate for a centrifugal force contribution included therein. In particular, the offset may be proportional to the square of the rotor speed in the obtained rotor speed signal 201 , in which case the rotor speed signal 201 is received by the estimation unit 204 as shown in Figure 2. With reference to Figure 4, one approach for determining the offset to be applied to the received acceleration signal 206 to compensate for centrifugal force contributions in the signal 206 is described. The offset is determined based on the (known) position and orientation of the accelerometer 107 in the rotor hub 105 relative to the axis of rotation. Figure 4 schematically illustrates the rotor plane 401 of the wind turbine 10. In particular, Figure 4 shows the rotor plane 401 when (a specific) one of the rotor blades 106 is pointing directly downwards, which corresponds to a rotor azimuth angle, 6, of zero radians, i.e. 0 = 0.

[0055] Figure 4 schematically illustrates two possible positions A,Anof the accelerometer 107. In position A the accelerometer 107 is located directly below (vertically downwards from) the rotor axis 402 (i.e. centre of the rotor plane 401), i.e. on a centre line 403 of the rotor plane 401 , at a distance rafrom the centre 402 of the rotor plane 401. On the other hand, in position Anthe accelerometer 107 is offset from the centre line 403 by an angle 0off, and located a distance rnfrom the centre 402, as shown in Figure 4. Note that 0off= 0 at position A. As illustrated in Figure 4, the positioning of the accelerometer 107 will impact the manner in which the centrifugal force Fcis captured on the accelerometer axis. The accelerometer 107 measures acceleration ax, ayin the mutually orthogonal directions x and y in the rotor plane 401 (as well as parallel to the rotor axis in direction z). Here, the measured acceleration ax, aymay be the sensor signal 206 transformed from the rotating reference frame to the stationary reference frame.

[0056] The offset angle 0offmay be determined as 0off= tan-1(dx / dy), where dx, dyare the known distances of the accelerometer 107 from the rotor plane centre 402 in the x and y directions, respectively. The radius rnis determined as rn= dx2+ dy2

[0057] Also, the centrifugal force Fcis then determined as c — ^rot Ti where vrotis the angular rotor speed, and the centrifugal force components along the accelerometer axes x and y, e.g. in the stationary frame, would be Fcsin 0offand Fccos 0off. For convenience, the mass is considered to be one in the above formula. These components can then be removed from the total measurement of the accelerometer 107 along the respective axis, thereby providing the centrifugal force offset to the received accelerometer signal 206. In practice, the accelerometer 107 is affected by and may measure the centripetal acceleration (and additional accelerations, e.g. gravity). The centripetal acceleration measured by the accelerometer is given wA2-rn, where omega is the angular velocity. Thus, the accelerometer 107 may not measure Fc, but instead the centripetal acceleration which is related to Fc. The components of the measured centripetal acceleration along the accelerometer axes correspond to those of the centrifugal force Fc.

[0058] Thus, instead of using the measurements from the accelerometer 107 directly, by mapping the sensor signal 206 into the stationary rotor frame and then applying the centrifugal force compensation, or centripetal acceleration compensation, the position and orientation of the accelerometer 107 can be chosen arbitrarily. Thus, mounting locations of accelerometers in the hub can be chosen based on considerations of locations which provide the simplest mounting.

[0059] The other disturbances in the received acceleration signal 206, e.g. caused by vibrations of the rotor hub 105 and nacelle 103, may be neglected as they have no continuous impact on the acceleration signal 206 and in any case are negligible in magnitude relative to the contribution from gravity. The resulting acceleration signal is therefore two components that are 90 degrees out of phase, providing sinusoidal signals with an amplitude close to 1G and a frequency proportional to the rotor hub speed.

[0060] Returning to Figure 3, at step 303 the method 30 involves, at the estimation unit 204 of the controller 20, determining a reference rotor azimuth angle based on the received acceleration sensor signal 206. This may be achieved by applying an inverse tangent function (arctangent function) to the mutually orthogonal components ax, ayin the rotor plane 401 of the acceleration signal 206 (or the offset components ax, aythat have been compensated for centrifugal force). This operation may take into consideration the signs of the axes / components so that a result in the range -TT / 2 to TT / 2 is obtained, and this may further be transformed to be in the range 0 to 2n. A further offset may be applied to ensure that the reference rotor azimuth angle is obtained in a correct quadrant of the rotor plane 401.

[0061] Figure 5 shows a number of comparative plots of rotor position (i.e. rotor azimuth angle) against time. In particular, a first plot 501 (solid line) shows rotor position as determined by a previous method using a rotor speed integrator with a dedicated azimuth sensor to generate a reset pulse as outlined above. A second plot 502 shows rotor position as determined by a rotor speed integrator 202 in the absence of a generated reset pulse. Indeed, it is seen that the determined rotor position 502 drifts away from, or is offset from, the determined rotor position 501 when a reset pulse is used. A third plot 503 (dashed line) shows rotor position as determined in step 303 of the method as described above, i.e. the reference rotor azimuth angle signal 205 determined based on the rotor acceleration signal 206. It is seen that the rotor position determined using the rotor acceleration signal 206 is almost indistinguishable from the rotor position determined when a reset pulse is used to reset the rotor speed integrator.

[0062] Returning to Figure 3, at step 304 the method 30 involves, at the gain unit 207 of the controller 20, determining an azimuth angle error between the reference rotor azimuth angle signal 205 from the reference estimation unit 204 and the rotor azimuth angle signal 203 output by the rotor speed integrator 202. The azimuth angle error may be regarded as the difference in rotor position at a given time between the reference rotor azimuth angle signal 205 from the reference estimation unit 204 and the rotor azimuth angle signal 203 output by the rotor speed integrator 202. It is apparent from Figure 5 described above that, in the absence of the integrator being reset, or some other form of compensation being applied, the output signal 203 from the integrator 202 drifts from an accurate / correct determination of rotor azimuth angle.

[0063] In order to determine the azimuth angle error more readily, in some examples a normalisation function may be applied to each of the reference rotor azimuth angle signal 205 and the output signal 203 from the integrator 202 prior to determining the azimuth angle error. That is, the error may be determined based on the normalised versions of the respective signals 203, 205. The normalisation function(s) may involve transforming the signals 203, 205 to vary between zero and one. Figure 6 shows a normalised plot 602 of the rotor speed integrator output signal 502 of Figure 5, i.e. in the absence of a reset pulse or other compensation, and a normalised plot 603 of the reference rotor azimuth angle plot 503 of Figure 5, i.e. based on the acceleration signal 206. Figure 7 shows a plot of the azimuth angle error against time calculated as a difference between the normalised versions of the reference rotor azimuth angle and the output of the rotor speed integrator in the absence of a reset pulse or other compensation, i.e. the difference between the plots 602, 603 in Figure 6. It will be appreciated that the error / drift from the integrator 202 may be positive or negative. In the described example, the integrator signal may be at 70% of a full rotor cycle / revolution when the reference rotor azimuth angle is at 100%, i.e. the rotor cycle has completed. As the signals are cyclic in nature, in this example the integrator signal may be regarded as being 30% delayed or 70% ahead of the reference. These values correspond to the maximum and minimum values of the error plotted in Figure 7. As the integrator signal drifts away from the correct rotor azimuth angle as a result of error accumulation, the maximum and minimum values of the difference between integrator and reference signals will change proportionally to the drift in each rotor cycle. This is illustrated in Figure 8, which is a plot of the azimuth angle error against time shown over a longer time period than in Figure 7.

[0064] Returning to Figure 3, at step 305 the method 30 involves, at the gain unit 207 of the controller 20, determining the gain signal 208 to be applied to the obtained rotor speed signal 201 based on the determined azimuth angle error. The gain may for instance be determined in a manner that minimises the determined azimuth angle error, i.e. to attempt to reduce the error to zero. The gain signal 208 may be in the form of a monotonic signal or even a constant value signal. In order to obtain the gain signal 208 from the cyclic varying azimuth error signal (e.g. as shown in Figures 7 and 8), a transform may be applied to transform the cyclic varying error signal to a monotonic or constant value signal. As mentioned above, the integrator drift can be regarded as a delay relative to the reference signal or as being ahead of the reference signal. The delay may be being positive or negative depending on whether the integrator signal is behind or ahead of the reference signal. In the described example, the integrator drift is regarded as being a delay relative to the reference rotor azimuth angle. The transform in this case may be to calculate a difference between one and the azimuth angle error for values of the determined azimuth angle error greater than one, and to calculate an absolute value of the determined azimuth angle error for values of the determined azimuth angle error less than one.

[0065] Figure 9 shows a plot of the phase delay 901 obtained from the azimuth angle error 701 of Figure 7 using the described transform. The phase delay 901 in Figure 9 indicates a phase delay of 70%, which is equivalent to a phase delay of -30%. For definiteness, a further transform may be applied to the phase delay to obtain a value between -0.5 and 0.5 (-50% and 50%). A phase shift delay 1001 obtained according to this further transform is shown in Figure 10. Note that the phase delay is a metric that relates (current) rotor azimuth angle error of the integrator and a rotor revolution, independently of how long it takes to complete the rotor revolution. The gain to be applied to the rotor speed signal 201 is then determined based on the phase delay obtained from the determined azimuth angle error. For instance, the gain may be determined as one plus the determined phase delay.

[0066] Returning again to Figure 3, at step 306 the method 30 involves applying the gain signal 208 determined by the gain unit 207 to the obtained rotor speed signal 201 to obtain a gain-adjusted rotor speed signal 210 that is to be input into the integrator 202. In particular, the determined gain is applied to the obtained rotor speed by the module 209, which may for instance comprise a multiplier function to multiply the two signals 201 , 208 together. At step 307, the method 30 involves applying the rotor speed integrator 202 to the gain- adjusted (or compensated) rotor speed signal 210 to determine the rotor azimuth angle signal 203.

[0067] The rotor azimuth angle signal 203 is also fed back to the gain unit 207 so that the described method steps may be performed repeatedly as part of a feedback loop process. The steps of the method 30 may be performed substantially continuously, e.g. at each sampling point of the controller 20. For example, the sampling points may be repeated with a time step of a few microseconds, e.g. 5ms, 10ms, 20ms such as time steps within a range of 1-100ms. Thus, the rotor azimuth angle signal 203 may be updated at each sampling point so that the signal 203 is updated several times during a 360° rotation of the hub 105, such as at least 5 times, at least 10 times or at least 50 times during a 360° rotation. Thus, even at low rotor speeds the rotor azimuth angle can be determined with high accuracy since possible azimuth angle errors do not accumulate over long periods. For example, if the rotor is braked, the azimuth is still correctly determined until the rotor is brought to a full stop

[0068] Figure 11 shows a plot of the reference azimuth angle 1101 and the azimuth angle 1102 determined by the integrator 202 over time when such a feedback loop is performed according to the described method 30. It is seen that the integrator output 1102 tends towards the reference azimuth angle 1101 over time. This means that phase delay I gain tends towards zero over time, which is illustrated in the plot 1201 of phase delay over time in Figure 12 for the example of Figure 11 .

[0069] It is noted that, while it is beneficial to use the reference rotor azimuth angle determined based on the rotor hub accelerometer 107 to determine a phase delay and gain for the purposes described above, it remains beneficial to determine rotor azimuth angle that is to be used for instance as input to one or more control routines of the wind turbine based on the obtained rotor speed signal 201. This is because vibrations of the wind turbine 10, such as side-side vibrations of the tower 102 can influence the accuracy of the reference rotor azimuth angle determined based on the acceleration signal 206. For instance, there may be a discontinuity in the signal, e.g. when a rotor blade 106 passes the tower 102. On the other hand, rotor speed measurement is not influenced by tower vibrations and so can use the integral of the rotor speed signal without there being any discontinuity.

[0070] Many modifications may be made to the described examples without departing from the scope of the appended claims.

[0071] 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.

Claims

CLAIMS1. A method of determining rotor azimuth angle of a wind turbine, the method comprising: obtaining a rotor speed signal indicative of rotor speed of the wind turbine; applying a gain to the obtained rotor speed signal to obtain a gain-adjusted rotor speed signal; and applying a rotor speed integrator to the gain-adjusted rotor speed signal to determine rotor azimuth angle of the wind turbine, the method comprising: receiving an acceleration sensor signal, from an acceleration sensor located in a rotor hub of the wind turbine, indicative of acceleration, due to gravity, in a radial direction of the rotor hub relative to a rotation axis of the rotor hub; determining a reference rotor azimuth angle based on the received acceleration sensor signal; determining an azimuth angle error between the reference rotor azimuth angle and the rotor azimuth angle determined by the rotor speed integrator; and determining the gain to be applied to the obtained rotor speed signal based on the determined azimuth angle error.

2. A method according to Claim 1 , wherein the gain is determined to minimise the azimuth angle error.

3. A method according to Claim 1 or Claim 2, wherein the steps of the method are performed repeatedly as part of a feedback loop process; optionally, wherein the steps of the method are performed substantially continuously.

4. A method according to any previous claim, the method comprising, prior to determining the azimuth angle error, applying a normalisation function to the reference rotor azimuth angle and to the rotor azimuth angle determined by the rotor speed integrator, wherein the azimuth angle error is determined based on the respective normalised rotor azimuth angles.

5. A method according to any previous claim, wherein determining the gain based on the determined azimuth angle error comprises: determining a phase delay based on the determined azimuth angle error; and, determining the gain based on the phase delay.

6. A method according to Claim 5, wherein the determined azimuth angle error is a cyclic varying signal, and wherein the method comprises applying a transform to the cyclic varying signal to obtain the phase delay, wherein the obtained phase delay is a monotonic signal over at least one period of the cyclic varying signal.

7. A method according to Claim 6, wherein applying the transform comprises: for values of the determined azimuth angle error greater than one, calculating a difference between one and the azimuth angle error; and, for values of the determined azimuth angle error less than one, calculating an absolute value of the determined azimuth angle error.

8. A method according to any of Claims 5 to 7, the method comprises applying a filter to the determined phase delay prior to determining the gain, wherein the filter is configured to reduce disturbances in the received acceleration sensor signal caused by tower oscillations of the wind turbine.

9. A method according to any previous claim, wherein the received acceleration sensor signal is indicative of acceleration, due to gravity, of the rotor hub in first and second mutually orthogonal radial directions.

10. A method according to Claim 9, wherein determining the reference rotor azimuth angle comprises applying a trigonometric function to components of the received acceleration sensor signal in the first and second mutually orthogonal radial directions.

11. A method according to Claim 10, wherein the reference rotor azimuth angle is determined in dependence on an orientation and position of the acceleration sensor in the rotor hub.

12. A method according to any previous claim, the method comprising applying an offset to the received acceleration sensor signal to compensate for a centrifugal force contribution included therein, wherein the offset is proportional to the square of the rotor speed in the obtained rotor speed signal, and wherein the reference rotor azimuth angle is determined based on the centrifugal-compensated acceleration sensor signal.

13. A method according to any previous claim, wherein the rotor speed in the obtained rotor speed signal is determined based on a received generator speed signal indicative of generator speed of the wind turbine; optionally, wherein the generator speed signal is received from a rotary encoder of the wind turbine.

14. A method according to any previous claim, further comprising performing a coordinate transformation of the components of the acceleration sensor signal (206) from a rotating reference frame into a stationary reference frame.

15. A controller for determining rotor azimuth angle of a wind turbine, the controller being configured to: obtain a rotor speed signal indicative of rotor speed of the wind turbine; apply a gain to the obtained rotor speed signal to obtain a gain-adjusted rotor speed signal; and apply a rotor speed integrator to the gain-adjusted rotor speed signal to determine rotor azimuth angle of the wind turbine, the controller being configured to: receive an acceleration sensor signal, from an acceleration sensor located in a rotor hub of the wind turbine, indicative of acceleration, due to gravity, in a radial direction of the rotor hub relative to a rotation axis of the rotor hub; determine a reference rotor azimuth angle based on the received acceleration sensor signal; determine an azimuth angle error between the reference rotor azimuth angle and the rotor azimuth angle determined by the rotor speed integrator; and determine the gain to be applied to the obtained rotor speed signal based on the determined azimuth angle error.

16. A wind turbine comprising a controller according to Claim 15.

Images