SYSTEM AND METHOD FOR CONTROLLING OPTIMAL YAW

Description

Field of the Invention

[0001] The present invention relates to a system and a method for optimal yaw control of a wind turbine, comprising a tower carrying a rotatable nacelle rotated by a yaw motor, which nacelle comprises at least one generator connected by a shaft to a rotor, which rotor comprising one or more wings, which nacelle further comprises means for detecting wind direction and wind velocity, which system performs measurement and storing data related to power production, wind velocity and wind direction.Background of the Invention

[0002] US 2014 / 0186176A discloses a method of determining a degree of yaw error of a wind turbine. The method includes obtaining wind pressure measurement values from in front of the rotor plane, which wind pressure measurement values exhibit a periodic nature related to a rotation of a spinner of the wind turbine, performing a signal processing step to process the wind pressure measurement values to determine a phase offset of the wind pressure measurement values relative to an angular reference, and deriving a yaw error angle from the phase offset. A method of establishing a relationship between a phase offset of wind pressure measurement values and a degree of yaw error of a wind turbine, a yaw error determination unit, and a wind turbine are also provided.

[0003] US 2008 / 01111379 A1 discloses a wind energy installation including a pylon, a pilot-able machine housing arranged on the pylon, a rotor configured to rotate on an end face of the machine housing and to drive a generator for production of electrical power, measurement devices for measurement of wind speed and direction and of an electromechanical quantity, a controller for the pivoting device, and a calibration module for the controller that comprises an efficiency measure element configured to calculate an efficiency measure for the production of the electromechanical quantity via the wind speed and an evaluation device. The evaluation device includes a classifier with outputs for positive and negative wind directions and calculation elements for determining mean values of the outputs and a comparator element which determines the difference between the mean values and outputs a calibration signal corresponding to the difference to the controller for the pivoting device.

[0004] US 2011 / 0101691 A1 discloses a system and method for optimal yaw control in which information relating to the power production and wind direction is logged and stored and subsequently analyzed, during which the yaw error is compared to the actual power production, and in case the maximum power production occurs at a yaw error greater than or less than 0, the yaw measurement is corrected with an offset value, namely the yaw error which produces the maximum power. The yaw control is based on measurements over a limited time period of 10 minutes.

[0005] US 2015 / 086357 A1 and US 2010 / 066087 A1 also discloses systems and methods for perform yaw corrections based on data relating measured yaw error to measured power output of the wind turbine. However, both disclosures require additional parameters for determining the optimal yaw control.Object of the Invention

[0006] The object of this invention is to optimize power productions of wind turbines. A further object is to let a wind turbine optimize the yaw position of a nacelle to the wind direction. A further object of the invention is to optimize the power production mostly by use of existing hardware in a wind turbine i.e. a system with reduced complexity.

[0007] Furthermore, it is an objective to obtaining a stable system with reduced risk of oscillation by use of a build-in delay in correction / regulation of the wind turbine.Description of the Invention

[0008] The object can be fulfilled by a system, as disclosed in the preamble of claim 1 and further modified in that power production measured in a positive direction to an actual yaw position is accumulated in a first storage related to the measured wind direction and that power production measured in a negative direction to the actual yaw position is accumulated in a second storage related to the measured wind direction, which system is configured to compare the power production accumulated for positive and negative yaw errors to estimate a difference in power production, which system is configured to further perform a yaw correction based on the comparing of power production for positive and negative yaw errors.

[0009] By this system, it can be achieved that existing components operating in a wind turbine can be used by this system in order to optimize the power production of the wind turbine by letting the wind turbine perform self-calibrating yaw control. By this system the wind turbine will during normal operation monitor the power production in relation to measured wind speed and wind direction. By this system stochastic wind direction change in relation to the yaw position of the wind turbine, wind from different positive or negative directions related to the yaw position can be measured and related data referring to positive or negative wind directions related to the actual yaw position are accumulated in two different storages. By comparing the data related to power production plus or minus to the yaw direction can be compared and the difference in actual measured production is used for new calibration of the yaw control. In this way all necessary components for performing the self-calibrating yaw control already exist in the wind turbine. All new features that are necessary in order to use this system are a question about a new software package placed in existing controllers. By optimizing the yaw position of a wind turbine it is, possible to increase the power production of the wind turbine between two to five percent. The increase in production is of course dependent on the wind speed, so with extremely low wind speed optimizing gives only very little results. Also in very high wind situations other regulation systems such as pitch control is reducing the power production, therefore, the system as disclosed here is operational in wind speed between maybe 5-15 metres per second. Statistically, this wind velocity seems to be existing most of the time. Therefore, this system can be very important in optimizing production of wind turbines.

[0010] According to the invention data related to power production in positive and negative wind directions is accumulated during at least a defined time period.

[0011] In order to have the relatively best conditions for yaw calibration it is necessary to perform measurement of positive and negative power production over a defined time period. You can probably have a result after a few hours of operation but the result will be much better for a longer period of time. In fact, a system could operate where the self-calibrating yaw control continues in operation. Probably in a system where the oldest data in the accumulation registers is deleted. In that way the system will continuously be operating but only having influence on data from the defined time period.

[0012] According to the invention the defined time period is longer than 24 hours. It is possible in one embodiment to achieve a good result which has been achieved in an assimilated system. In practical use of the system the result will probably show that period as long as one week will be much more effective for having the correct yaw correction. According to the invention the system is adapted to use at least a yaw correction algorithm, which system is further adapted to transmit a signal Perr to a PI controller which performs integration of the signal Perr, which system is further adapted to communicate a yaw offset set point γδ to a yaw controller which system is further adapted to communicate yaw correction signals to a wind turbine.

[0013] Hereby only a few extra electronic or software modules are necessary in order to let this system operate in an existing wind turbine.

[0014] In a further preferred embodiment for the invention can the wind turbine be adapted to communicate a yaw signal γ representing the actual yaw position of the wind turbine back to the yaw controller, which yaw signal γ is communicated further back to the yaw corr. algorithm. Hereby the feed forward regulation that is performed for the yaw control also uses a feedback system were actual yaw position is used for a new input for the yaw correction algorithm. Because of the signal forward direction there is a PI regulator there is a time delay in the forward regulations which is probably necessary for avoiding oscillations in the yaw system.

[0015] In a further preferred embodiment for the invention the yaw correction algorithm can be adapted to receive input P from the actual power production of the wind turbine. Hereby it can be achieved that not only actual measured wind direction and wind velocity is used for the yaw control, but also data representing actual power production is used as an input to the yaw correction algorithm. Hereby, any deviation in power production related to plus / minus wind direction will be part of the accumulated plus / minus values in the system.

[0016] In a further preferred embodiment for the invention the system can be adapted to perform sufficient filtering and estimate average power to the left and right of the yaw offset γδ. Hereby a further filtration can be performed which can reduce the content of relatively high frequency yaw control commands. Also potential noise that is generated from other electrical systems that could be picked up in signal lines will be reduced in that filtration.

[0017] The pending patent application further discloses a method, according to claim 6, for self-calibrating yaw control as previous disclosed and in the following steps of operation: a. let the wind direction change stochastically and measure power production of a wind turbine for positive and negative yaw errors, b. perform accumulation of data representing estimated power production for positive and negative yaw errors, c. perform the accumulation of data in at least a defined time period, d. use the accumulated data to estimate a power difference Perr between measured positive and negative yaw generated power, e. transmit the estimated power difference Perr (42) to a PI controller which integrates the signal to generate a yaw offset set point γδ, f. use the yaw offset set point γδ in the yaw controller to control the yaw direction of the wind turbine.

[0018] Hereby, it can be achieved that this method leads to a highly effective way for yaw correction of wind turbines. This can be achieved mostly by existing equipment in the wind turbine. Only a few new software modules are necessary in order to increase the power production. The reason for this power correction is that for many wind turbines actual wind directions are measured by measuring systems that in some situations have measuring failures. Also the position of the typical direction measurement for wind placed at the top of the nacelle will have some influence on the wind passing along the nacelle. But also turbulence generated by the rotating wings can in some situations have influence on the actual measurement of wind direction. Also for wind turbines operated at land, disturbance of the wind can lead to a measurement failure. An effect of measuring failures which has probably never developed is simply the change of wind direction in relation to the height over the earth surface. With wind turbines operating with extremely long wings, there can be sufficient change in wind direction from the wind height measuring the wind direction at the nacelle and the wind that is hitting the wings of the wind mill when they are at the top position. Therefore, there are a lot of good arguments for using the self-calibrating yaw control.

[0019] By an alternative embodiment for this invention can the Power production used as input for the algorithm be full or partly be supplemented by Cp as input. Cp is defined as the effect of the rotor, calculated from the generated power P, the wind speed V, air density Rho, the rotor area A, calculated by the following formula: Cp = P / (½ x A x Rho x V 3< ).

[0020] In some situations where the algorithm is calculating a correction signal based on P or Cp input signals, the algorithm is only calculating correction under some defined conditions. a - Calculate correction for defined intervals of direction of the nacelle or direction of the wind relative to the nacelle. Hereby, it is possible to stop using the algorithm from directions where e.g. landscape characteristics such as hills at one side of the wind turbine or if there is shadow effect from one defined direction, then the calculation of correction values could be cancelled for some defined conditions. b. - only perform correction calculation for defined intervals of the turbulence intensity of the wind. Hereby, the shadow effects from other wind turbines or turbulence from the atmosphere or turbulence from the wind introduced by the terrain or shadows from buildings etc. which all can reduce the effectiveness of the algorithm. c. - only calculate correction for one or more wind speed intervals. That could be 5 - 9 m / s, hereby the algorithm can be used in the wind speed where it is most effective and not used where the algorithm is less efficient or not working. d. - further it is possible to perform a combination of the conditions a, b and c.

[0021] Hereby, it is possible to adjust the algorithm to the actual use in a wind turbine. The algorithm for yaw correction can of course increase the power production for a wind turbine, but in some conditions this algorithm will probably not operate optimal. Therefore, it is highly effective if the algorithm is in use where the conditions are optimal and where the yaw correction can be formed and the power production can be increased.Description of the Drawing

[0022] Fig. 1discloses theoretical power generation as function of yaw misalignment. Fig. 2discloses a block diagram overview of the Yaw Alignment Correction algorithm. Fig. 3discloses the method to use this information in order to estimate a power difference. Detailed Description of the Invention

[0023] Since the wind is a varying energy resource, it is necessary to track the wind direction with a turbine rotor in order to optimize the power intake. This is done by letting a yaw motor rotate the turbine nacelle according to a yaw sensor (wind vane) by a yawing algorithm. The wind vane and the airflow around the rotor and the nacelle is often offset from the true wind direction, leading to a general drop in power, as illustrated in Figure 1. The potential in AEP of correcting the yaw offset is up to several percentages, thus being a strong argument for inventing an automatic correction algorithm. The invention described in this document includes a yaw controller which tracks a specific power-optimal yaw set point.

[0024] Figure 1: The blue line is theoretical power 24 as function of yaw misalignment. The yaw measurement is offset by 5°, leading to a power optimal yaw marked by γδ 28. As indicated by arrows, the actual wind direction will have some fluctuations from the point where the yaw adjustment is fixed for a period. Wind direction will change plus and minus to the defined direction. For this patent application we have realized that yaw misalignment often happens simply because the measurement of the wind direction is performed in a rather primitive way, for example by a wind vane placed on the top of a nacelle. Turbulence or shadow effects can change the wind direction and the rotating rotor can have some influence on the actual wind direction. Also the actual shape of the nacelle can have some influence on the actual wind direction. By continuous measurement of actual produced power of a wind turbine, which often takes place, it is possible to perform a correlation to the measured wind direction. In the pending application, this performed an accumulation of power production plus / minus in relation to the yaw position. It is probably necessary to measure the power production for a rather long time in order to have a reliably better adjustment of the yaw position.

[0025] Figure 2: Block diagram overview of the Yaw Alignment Correction algorithm 40. An algorithm calculates a power error P err 42 used by a PI controller 44. This, in turn, gives a yaw offset setpoint γδ 46 which is tracked by a yaw controller 48. γ: yaw error measurement, P: power measurement. Figure 2, which describe how the parts interface to the turbine 4. During normal operation the wind direction 20 changes stochastically, leading to a variation in yaw error γ and power P 24. Hereby, a more or less continuous yaw correction algorithm is operating. This yaw correction algorithm is probably also a module where the accumulation of power production takes place. At least there is a power input to that module. After a defined delay, communication of the Perr 42 to the PI controller 44 will probably start. This PI controller will perform integration of the signal that is received, so that any quick response in the perr signal 42 will be very much delayed. The output from the PI controller is the yaw correction signal 46, which is transmitted into the yaw controller 48. This yaw controller 48 generates a commanding signal 50 to the wind turbine and forces the wind turbine to change its yaw direction. The actual yaw signal 54 is transmitted back to the yaw controller 48 and back to the yaw correction algorithm 40. That way, a continuous adjustment of the yaw signal can be performed, even when the yaw controller 40 has performed with a relatively long time delay, which could be more than 24 hours. In order to get reliable data it is possible to change the yaw direction of the nacelle into a more correct position and hereby maybe increase the yield of power up to 5 percent.

[0026] Figure 3 describes the method to use this information in order to estimate a power difference, P err , which indicates an imbalance between positive- and negative yaw error. Feeding P err to a PI controller it is then possible to track zero imbalances by minimizing P err . The output of the PI controller is the yaw offset setpoint γδ, which is used as the set point for the yaw controller. Thereby the yaw controller becomes self-calibrating in order to maximize the power output : The average power to the left 26 and right 32 of the yaw offset γδ 46 is estimated with sufficient filtering. The difference, P err 42 indicates whether it is possible to increase the power output by moving the yaw offset set point 46. At figure 3 it is clearly indicated that there is a difference in power production between 26 and 32. Therefore this figure indicates that a yaw correction in the correct direction of the P1 26 probably could give a better yield.

[0027] It is possible by this patent application to increase the power production of a wind turbine in any place where a wind turbine is operating. The optimization of the production seems to be effective for medium wind maybe starting from 5 metres per second and ending at approximately 15 metres per second. At very low wind velocities it is probably possible to increase the yield but it is very difficult to measure the positive effect. Above 15 metres per second, wind mills are starting maximal production and other regulation means, such as pitch control or maybe stall control will reduce the production, so the method of self-adjustment of the yaw has only minimal effect. But measured over a year, it is a fact that between 5 and 15 metres per second is where most of the power production is performed by nearly all wind turbines. Therefore, the effect of the pending application, which has shown that it is possible to increase the power production maybe up to 5 percent, could be very important, not only at single operating wind mills but probably also at wind farms at land or wind farms at sea. In wind farms every single wind turbine will optimize its yaw position according to the actual wind situation. In that way different shadowing effects are maybe compensated in a highly effective way. List of numbers wind turbine4power production, P24power production measured in a positive direction26power optimal yaw position28power production measured in a negative direction32yaw correction algorithm40signal Perr42PI controller44yaw offset set point, γδ46yaw controller48yaw correction signal50yaw signal, γ54

Claims

1. System (2) adapted for optimal yaw control of a wind turbine (4), comprising a tower carrying a rotatable nacelle rotated by a yaw motor, which nacelle comprises at least one generator connected by a shaft to a rotor, which rotor comprises one or more wings, which nacelle further comprises means for detecting wind direction and wind velocity, which system is configured to perform measurement and to store data related to power production (24) of the wind turbine (4), wind velocity and wind direction, characterized in that power production (26) of the wind turbine (4) measured in a positive direction to an actual yaw position (28) is accumulated in a first storage related to the measured wind direction, and that power production (32) of the wind turbine measured in a negative direction to the actual yaw position (28) is accumulated in a second storage (34) related to the measured wind direction , which system is configured to compare the power production (26, 32) accumulated for positive and negative yaw errors to estimate a difference in power production, which system is configured to further perform a yaw correction based on the comparing of power productions for the positive and negative yaw errors, whereby data related to the power production in positive and negative wind directions related to the actual yaw position (28) are accumulated during at least a defined time period (36), which defined time period (36) is longer than 24 hours, wherein the system is adapted to use at least a yaw correction algorithm (40), which system is further adapted to transmit a signal (42) which derives from the estimated difference in power production to a PI controller (44) which performs integration of the signal Perr (42), which system is further adapted to communicate a yaw offset set point γδ (46) to a yaw controller (48), which system is further adapted to communicate yaw correction signals (50) to the wind turbine (4).

2. System according to claim 1, characterized in that the wind turbine (4) is adapted to communicate a yaw signal γ (54) representing the actual yaw position of the wind turbine (4) back to the yaw controller (48), which yaw signal γ (54) is communicated further back to the yaw correction algorithm (40).

3. System according to any one of the preceding claims, characterized in that the yaw correction algorithm (40) is adapted to receive input P (56) from the actual power production of the wind turbine (4).

4. System according to claim 3, characterized in that the input P (56) used as input for the algorithm is fully or partly supplemented by Cp as input, wherein Cp is defined as effect of the rotor, calculated from the generated power P, wind speed V, air density Rho, rotor area A, calculated by the following formula: Cp = P / (½ x A x Rho x V3).

5. System according to one of the claims 1-4, characterized in that the system is adapted to perform filtering and estimate average power production to the left and right of the yaw offset γδ threshold and average power production for yaw errors smaller than the yaw offset.

6. Method for self-calibrating yaw control as disclosed in one of the claims 1-5, characterized in the following steps of operation: a. let the wind direction change stochastically and measure power production of a wind turbine for positive and negative yaw errors, b. perform accumulation of data representing measured power production for positive and negative yaw errors, c. perform the accumulation of data in at least a defined time period, which defined time period (36) is longer than 24 hours and preferably that the defined period (36) is as long as one week, d. use the accumulated data to estimate a power difference Perr (42) between measured positive and negative yaw generated power, e. transmit the estimated power difference Perr (42) to a PI controller which integrates the signal to generate a yaw offset set point γδ (46), f. use the yaw offset set point γδ (46) in the yaw controller (48) to control the yaw direction of the wind turbine.