The invention relates to the control of wind turbine installations, and in particular to the control of wind turbine installations during extreme operating conditions, such as when a high yaw error is detected.

Wind turbines must operate safely and reliably to extract energy from the incoming wind. To do this a number of control systems installed in the wind turbine or in the wind park are used to monitor the prevailing wind conditions and adjust the operating

parameters for the wind turbine accordingly. In most cases, and for horizontal wind turbines having wind turbine blades with variable pitch control, adjustment in this context means altering the pitch angle of the wind turbine blades to alter the amount of energy that the blades extract from the wind and the related load experienced by the blades as a result.

Most control schemes are divided into partial and full load operation. In partial load operation, the wind turbine controller may adjust the pitch of the blades to an optimal pitch angle, typically around zero degrees. This ensures that while ever the wind turbine is not producing the maximum amount of electricity it is able to produce, the rated value, it is controlled to capture as much energy from the wind as it can, and speed up the electrical generator connected to the wind turbine blades. Once the wind turbine has reached its rated value, it operates in full load operation, and control of the wind turbine blade pitch is used to turn the blades out of the wind to avoid overspeed of the generator and/or unsafe operation of the turbine.

It is known to detect extreme operating conditions, such as gusts of wind that because of a sudden increase in speed, or a sudden change in direction of the wind, could cause unsafe loads on the wind turbine components. Sudden direction change of the incident wind is a particular problem as, in normal operation, the wind turbine is turned or yawed into the wind for correct operation. Facing into the wind, ensures that the wind turbine extracts the maximum energy from the wind and also that the wind turbine is controlled in a safe manner. Typically, therefore a wind turbine might operate with a target relative yaw angle of zero degrees, indicating that the wind turbine is fully facing into the wind. However, if the wind direction suddenly changes, or the wind turbine is no longer able to turn into the wind, due to malfunction or loss of power, then the wind turbine might experience a yaw error. This is simply the difference between the direction the wind is blowing and the direction the wind turbine is facing.

High yaw errors can result in structural loads at the wind turbine and unsafe operating conditions. To avoid these situations, yaw errors must be detected quickly so that mitigating actions can be initiated to avoid high loads.

One of the difficulties with operating a wind turbine during high yaw error is suitable control of the wind turbine blades as they rotate. In normal operation, the loads on the blades are carefully balanced as the blades rotate around the hub using an appropriate pitch control algorithm.

The pitch control algorithm not only controls the blades according to the wind conditions, but also according to the position of the blades around the hub. For example, in wind shear situations, the blade that is pointing upwards into space, may experience more wind related load than the blade pointing downwards towards the ground, and the pitch control applied to the blades individually applies appropriate compensation to balance the loads on the hub and avoid undue loads acting in the tilt direction of the main bearing.

In extreme yaw events, however, the action of the incident wind on the blades is less predictable. It is possible, for example, that despite the correct pitch angle being applied to an individual blade, the angle of attack between that blade and the incident wind is such as to cause the blade to enter stall condition. This has been found to occur for both negative angle of attack (positive angles of attack are usual for lift generating conditions), as well as for small positive angles of attack. This can result in sudden drop off of the load on that respective blade, but no corresponding drop-off in load on the other blades. A sudden mismatch in load occurs leading to a high tilt and yaw moment, and as a result, the components of the wind turbine can be overloaded.

We have therefore appreciated that it would be desirable to provide a method and apparatus for control of the wind turbine to reduce structural loads on the wind turbine and to withstand high yaw error conditions.

SUMMARY OF THE INVENTION

In a first aspect, the invention provides a method for reducing the loading on a wind turbine during high yaw error events, the method comprising the steps of: detecting a wind turbine yaw error; and when a yaw error is detected, operating a load-reducing controller for the wind turbine to mitigate one or more structural loads associated with the yaw error, wherein the load reducing control engages a safe mode in which it operates to prevent the collective pitch control angle from falling below a minimum safe pitch angle.

As a result of the method, stall conditions on individual blades in high yaw events can be avoided or minimised, meaning that the load experienced by the rotor and main bearing are less likely to suffer from high fluctuations.

In an embodiment of the invention, the minimum safe pitch angle is set as a function of the magnitude of the yaw error. In this manner the load reducing action can be applied in accordance with a desired dependency on the yaw error. In embodiments, the function is defined to provide an increasing or a constant minimum safe pitch angle with increasing yaw error. In this manner a simple, yet effective, load reducing behaviour can be applied. It may be beneficial to limit the operation of the load reducing controller to yaw errors which are detected in excess of a first threshold. The first threshold may be set in accordance with a load tolerance level above which a load reducing action is desired, or simply to avoid too much activation of the load reducing controller at low yaw errors.

The method may comprise determining the wind speed at the wind turbine, and determining the minimum safe pitch value and/or the function of the magnitude of the yaw error based on the wind speed, and/or on the rate of change of wind speed. This allows the controller to apply a safer operating condition when the wind speed is determined to be changing rapidly. Moreover, a higher variance of the wind direction is often observed at low wind speeds than at higher wind speeds and it may be beneficial to limit or reduce the operation of the load reducing controller for wind speeds below a given limit or threshold, or simply only to engage the load reducing controller for wind speed above a wind speed threshold, such as above between 5 m/s to 15 m/s, such as between 7 m/s to 12 m/s, such as 10 m/s. This may be beneficial since at low wind there is a much higher variance of the wind direction.

The method may comprise: detecting a wind turbine yaw error in excess of a first threshold, and setting the minimum safe pitch angle as a constant minimum safe pitch angle when the yaw error is detected in excess of a first threshold.

The method may further comprise: detecting whether the yaw error exceeds a second threshold level, higher than the first threshold, and if the second threshold is exceeded, increasing the minimum non-zero value. As a result, the controller can distinguish between moderate and severe yaw error events and apply more or less stringent control appropriately. The minimum safe pitch value is typically a non-zero positive pitch angle.

The method may comprise ceasing the operation of the safe mode when the yaw error is detected as falling below the first threshold. This may occur instantaneously, so that optimal production of power can be commenced immediately, or alternately, the safe mode operation may cease after a predetermined period of time has elapsed from detecting that the yaw error falls below the first threshold. This mode of operation may be safer, allowing for the fact that another yaw error event may quickly follow a first.

In one embodiment, detecting the yaw error may be based on one or more of detecting a wind direction change at the wind turbine, or detecting the wind direction in relation to the wind turbine. Alternatively or in addition, detecting the yaw error may comprise determining the difference between the measured and the estimated wind speed at the wind turbine. The detection may also be based on a CUSUM algorithm, as the CUSUM algorithm is useful for accurately detecting changes.

The wind turbine blade load-reducing control means may comprise operating a pitch control algorithm to make continuous adjustments to individual pitch angles for the blades. The wind direction change may be detected based on measurements taken at or for a second turbine in the vicinity of the wind turbine generator.

A corresponding controller for a wind turbine is also provided to carry out the method discussed above. Thus a wind turbine controller is provided which may implement the functionality, singly or in combination, of any of the method elements of the various embodiments of the present invention.

BRIEF DESCRIPTION OF THE DRAWINGS

Preferred embodiments of the invention will now be defined by way of example only and with reference to the drawings in which:

Figure 1 illustrates schematically the operation of a control system according to an example embodiment of the invention;

Figure 2 illustrates a pitch control algorithm according to an example embodiment of the invention;

Figure 3 illustrates a pitch control algorithm according to an example embodiment of the invention;

Figure 4 schematically illustrates examples of functions which relate a measured or determined yaw error to a minimum safe pitch angle; and

Figure 5 illustrates a situation with a pitch limit that is set as a linearly increasing function with measured or detected yaw error.

DETAILED DESCRIPTION OF EXAMPLE EMBODIMENTS

Examples of the invention provide safe mode operation for wind turbines operating during high yaw errors.

Figure 1 is a schematic illustration of the operation of a control system according to an example embodiment of the invention. The control system comprises a first detection block 2 for detecting a yaw error, in other words a situation in which the nacelle of the wind turbine, and the wind turbine rotor hub are not pointing into the wind. Some tolerance between an angle of zero degrees, in which the nacelle is pointing directly into the wind, and a non-zero (positive and negative) value is defined for normal operation. Once this value is exceeded however, a yaw error can be considered to have occurred.

The yaw error detection block 2 receives one or more inputs a and b, which may be used individually or in combination in a yaw error detection calculation. In this example, one of the inputs is a measurement of the wind direction, and the other is a measurement of the wind speed. The wind direction measurement and the wind speed measurement may be measured using appropriate sensor equipment, including an anemometer, LIDAR, SODAR, or wind vane positioned on the wind turbine, or alternately on another wind turbine or weather mast in a position such that the conditions at the respective wind turbine can be deduced. LIDAR, and SODAR in particular, may operate to determine the wind conditions ahead of the wind turbine, so that pre-emptive action can be taken by the wind turbine controllers. In general, the yaw error may be detected in any way possible.

The yaw error detection block 2 processes the data received at the one or more inputs a and b and provides one or more outputs c and d to a number of connected controllers 4, 6, and 8. The controllers are responsible for controlling different aspects of the operation of the wind turbine, and so will each provide one or more outputs as a result. In Figure 1 , one of the controllers 4 is a pitch controller and provides a pitch angle control signal to one or more pitch actuators coupled to one or more respective wind turbine blades. Other controllers may also be provided in the control system including generator controller 6, and tilt and yaw controller 8.

The generator controller 6 provides control signals to the electrical windings of the generator. As is known in the art, the generator rotor is coupled to the main shaft of the wind turbine rotor, and generator control signals may be used to change the coupling between the electrical machine comprised by the generator and the mechanical machine comprised by the rotor in order to mitigate loads.

The tilt and yaw controller 8 adjusts the individual pitch angle applied to the blades to mitigate tilt and yaw moments on the main bearing shaft of the nacelle. Preferably, the tilt and yaw controller 8 is activated once a yaw error is detected.

The functional blocks illustrated in Figure 1 are intended to be schematic

representations only. In practice, the individual functions may be carried out by one or more hardware or software modules implementing the necessary control algorithms to effect the intended result. The functionality of a single block may for example also be combined with that of another block to form a combination. The pitch controller for example may be a distributed system having hardware components positioned in the nacelle and the rotor hub of the wind turbine respectively. The tilt and yaw controller may also be implemented as part of the pitch controller in the hub. The controllers illustrated in Figure 1 are examples of load reducing controllers installed as part of the wind turbine control system.

In this example, the yaw error detection block 2 is configured to process one or more of the inputs a and b using a CUSUM (cumulative sum) algorithm. The CUSUM algorithm is a statistical tool particularly well suited for monitoring changes in variables. It takes samples of a quantity (in this case wind speed and/or wind direction measurements) and combines these with sampling weights in a cumulative summation. When the cumulative summation exceeds a pre-determined set value, then a change is said to have been detected. The use of the CUSUM algorithm in yaw error detection block 2 is particularly advantageous as it provides a robust detection technique limiting the number of false positives. In other embodiments other processing techniques or algorithms for detecting a change in the wind speed and/or direction could be used.

In this example, the yaw error detection block 2 operates on the basis of at least a first threshold, and an optional second threshold, for the detection of an extreme yaw error. In this way, the severity of the yaw error may be quickly quantified and an appropriate signal passed on to the downstream controllers. Depending on the severity of the signal, the controllers may then take different actions appropriate to the circumstances. For example, one threshold may correspond to detection of a moderate yaw error, while another may correspond to an extreme yaw error. Depending on which of the first or second thresholds are passed, the detection block 2 may therefore output one or more detection flags indicating to the controller blocks, the severity of the yaw error event. In the example shown in Figure 1 for example, the output d, shown as a dashed line, may correspond to an output or 'flag' indicating a moderate yaw event, while the output c, shown as a solid line, may correspond to an output or flag indicating an extreme yaw event signal.

Although in the present example, each of the controllers is shown as receiving both outputs c and d, the controllers 4, 6 and 8 some controllers may receive only one of the outputs c and d, as appropriate. For example, it may be desirable to engage certain of the controllers 4, 6 and 8 only when an extreme yaw error event is detected, and take no action for the controller if only a medium yaw error event is detected. The use of outputs that indicate different severities therefore simplifies the control signals that need to be passed to the connected controllers.

Based on the inputs c and d (the output from the yaw error detection block 2) received at the controller, the controllers will take different action according to their programming.

Figure 1 is a schematic diagram and it will be appreciated that a single output or signal could be used to carry information relating to both moderate and extreme events and that there need not be separate dedicated control paths for each output in order to implement this.

The operation of the pitch controller 4 will now be described in more detail. Figure 2 illustrates a pitch control operation according to an example embodiment of the invention wherein the minimum safe pitch angle is set as a constant minimum safe pitch angle for yaw errors in excess of a given (first) threshold. At time an extreme yaw error event is detected by the yaw error detection block 2, and a yaw error event flag is sent to the pitch controller 4. As a result, at time the pitch controller enters a safe mode operation in which it subsequently applies a safe lower pitch lower limit to the instantaneous pitch angle commands output to the pitch actuators of respective wind turbine blades. This is typically a non-zero, positive value. However, as the pitch angle can be defined relative to any appropriate reference points, it may in fact be any suitable value (including a negative value) for the wind turbine to give safe operation. This pitch limit is shown in detail in Figure 2 by the dashed horizontal line, and is labelled as θι,. It shall be referred to as a minimum pitch limit, or a pitch floor.

Figure 2 illustrates on the vertical axis the instantaneous pitch angle command for a single wind turbine blade. Time is represented on the x axis. In normal operation, that is for times from t=0 to the pitch controller 4 calculates an appropriate pitch angle for the wind turbine blade based on a number of inputs, including instantaneous wind speed, load, generator load, power demand and so on. The wind turbine pitch controller may also apply cyclical pitch control to compensate for tilt and yaw loads on the turbine.

At time ti, the safe mode operation is commenced and the pitch controller 4 applies a lower limit to the pitch angle command that is output from controller 4 to the respective pitch actuators of each of the individual blades. In other words, irrespective of the calculated pitch angle, the pitch controller is prevented from supplying a pitch angle command at a value that is below the set lower limit. This may be achieved simply by applying a mask or envelope over the underlying algorithm calculating the desired pitch angle, so that where a lower pitch angle would otherwise be set, the pitch controller rounds up the value to the value of the lower limit. This may be achieved through simple programming logic.

As can be seen in Figure 2, without the lower limit being applied, the pitch controller would have output pitch angle control signals at values below the lower limit around time t=t ₂ . The non-limited pitch angle commands are shown as the curved dashed lines.

However, as a result of the safe mode operation, the pitch angle command signal is not allowed to fall below the value of the minimum pitch limit θι, _, but can otherwise still vary as dictated by the usual pitch control algorithm.

It will be appreciated that the actual pitch reference signal can be defined in a number of ways. In general, the pitch angle may be defined as the geometrical angle between a chord of the blade profile and the rotor plane at a given radius. For example, the pitch angle may therefore be the angle of the blade tip with reference to the rotor plane. Other locations on the blade surface will have potentially different angles of attack due to the twist in the blade from the tip to the root. Selecting the location on the blade span where the pitch angle is defined is merely a matter of convention.

In the graph of Figure 2, the lower limit for the pitch angle is illustrated as pitch angle Q _L . In this example, the angle is set to a value of 7 degrees, where a zero pitch angle is defined as where the chord is coincident with the rotor plane at a radius of 82%. As indicated by the vertical doubled headed arrow on the right hand side of the graph, the pitch limit Q _L \s tuneable, in that it may be varied according to the circumstances. For example, in situations where the wind speed and direction are determined to be varying rapidly, a higher value for the minimum pitch limit may be appropriate. The pitch controller 4 may make this determination based on wind speed and wind direction signals received at its inputs, performing the required differentiation to determine rate of change where desired.

Additionally, or alternatively, the pitch controller may tune the minimum pitch limit Q _L depending on the type of flag output by the yaw error detection block 2 and received by the pitch controller. For example, where the yaw error detection block 2 outputs an output d indicating a moderate yaw error, the pitch controller may limit the pitch angle to remain above the safe pitch value, 7 degrees in this example. Where the yaw error detection block outputs an output c indicating a more extreme yaw error, the pitch controller 4 may limit the pitch angle to remain above values of a higher value of 9 degrees. This is illustrated in Figure 3 to which reference should now be made.

In Figure 3, the yaw error detection block 2 detects an extreme yaw error at time t ₃ and issues an output c to the pitch controller (the output at time is assumed to indicated a moderate yaw error - output d). This causes the pitch controller to apply a minimum pitch limit to the output pitch control angle that is higher than that applied at time t ₂ . As a result, the pitch control angle algorithm continues to calculate the instantaneous pitch angle commands for output from the pitch controller block 4, but disregards values that fall below the new pitch floor, such as those occurring at times t ₄ and t ₅ . The pitch controller therefore locks values that should ordinarily be lower at the minimum pitch limit while safe mode is in operation. In the example shown, the pitch floor is a non-zero positive value that is constant in time until a change in the yaw error event or environmental conditions is detected, at which point it may be adjusted.

If, at any time, the yaw error detection block 2 detects that yaw error event has passed, the safe mode operation may end, and the minimum pitch limit can be removed or set to zero.

Excessive loads on the pitch turbine blades, and therefore the rotor, main shaft and connected components, are largely a result of differences between the actual angle of attack between the wind turbine blade and the intended angle of attack according to the given pitch angle. In the above example, a zero degree pitch angle corresponds to pitching the wind turbine blade into the wind to extract the maximum amount of energy from the incident wind. In this configuration, the blades' pressure and suction surfaces are positioned to experience maximum lift from the wind, and therefore any associated loading from the force of the wind. In strong wind conditions, the wind turbine blades are feathered, or angled out of the wind, thereby reducing the loads. This corresponds to an increasingly positive pitch angle. However in actual operation, where a sudden gust of wind is experienced, the wind direction can change before the pitch angle can be adjusted to compensate. In this scenario, and depending on the direction of the wind, the actual angle of attack for the individual wind turbine blade may correspond to a negative pitching of the blade into the wind.

If a large negative angle of attack is experienced by one of the blades, the loading on that blade will drop as the blade enters a stall condition, and the loads on the rotor hub will then be out of balance. This may produce a disruptive and even cyclical loading on the rotor hub.

The lower limit on the pitch angle therefore provides a margin of error between the wind turbine blade and the expected wind direction (indicated by a pitch angle of zero in Figures 2 and 3). The margin of error reduces the likelihood that sudden changes in wind direction would push the actual angle of attack into a negative pitch, where stall and loading conditions are problematic. At other times, the margin of error between the pitch angle command and the zero pitch angle means that the loading on the blade is kept lower than the maximum. This builds some leeway and margin into the system.

Figure 4 schematically illustrates examples of functions which relate a measured or determined yaw error to a minimum safe pitch angle or pitch limit 0 _L The figure illustrates two simple functions. In general other and more complex function may be used. One function is a step function 40 which implements setting a constant pitch limit above a first threshold Th1. Another function is a linearly increasing function 41 which linearly increases the minimum pitch limit for yaw errors above an activation level 42. The activation level 42 may be set at any appropriate level. As an example it may be set at the first threshold Th1 , in which case a linearly increasing pitch limit only starts at the first threshold, linearly increasing with increasing yaw error. Other examples include piecewise linear functions, progressively increasing functions, functions that flattens out at high yaw errors, etc, Moreover, the activation of the minimum pitch limit may be further dependent on a given minimum wind speed.

Figure 5 illustrates, in the lower graph, a similar situation as Figure 2, however with a pitch limit 50 set as a linearly increasing function with measured or detected yaw error. The upper graph of Figure 5 shows an example of the yaw error 51 as function of time, together with an activation threshold 52. At the time ti the yaw error 51 becomes larger than the threshold 52.

As the pitch limit 50 is set to increase linearly with the yaw error, the shape of the pitch limits 50 reflects the shape of the yaw error 51. Below the activation threshold 52, the pitch limit is set as the aerodynamically optimal pitch angle 53, but above the threshold, a small yaw error 51 A results in a small pitch limit 50A, whereas a larger yaw error 51 B results in a larger pitch limit 50B. In this manner the actual pitch angle 54 is affected in dependence on the magnitude of the yaw error. In the illustrated example, the pitch angle 54 is limited at times around t ₂ and around t ₃ . As a result, of the operation described above, the wind turbine can be safely operated despite an extreme yaw events occurring.

In the example described above, the minimum safe pitch angles have been expressed as values of 7 degrees and also 9 degrees. In practice values in the range of zero degrees to 10 degrees have also been found acceptable. Pitch controllers may send pitch reference signals which are negative in sign (up to minus 5 degrees for example) in certain circumstances, so a minimum pitch angle of zero will still in fact serve to provide the advantages provide by the invention.

In alternative embodiments of the invention, the minimum safe pitch angle or angles may be tuned off-line, that is determined in advance through testing of the wind turbine in various stall and extreme yaw event conditions. The determined values may then be set in memory for activation during extreme events. The testing may be repeated periodically if desired and the safe pitch angle or angles stored in memory may be updated.

Embodiments of the invention have been described by way of example only.

Nothing in the above examples is intended to limit the scope of the invention as defined by the following claims.