Method for operating a wind turbine

Abstract

A method for operating a wind turbine wherein a parameter for a wind hitting the wind turbine is determined from present values for the generator speed and/or the wind speed at each point in time (t). A temporal change variable is formed for the parameter at each point in time (t). For the temporal change variables, which occurred in a past time interval, the third and/or fourth statistical moment is calculated for a distribution of the temporal change values in the time interval. If at least one of the statistical moments exceeds a predetermined value, then a detection signal is set for an extreme gust, which triggers one or both of the steps: increasing a setpoint value for the blade pitch angle starting from an actual value thereof, and reducing a setpoint value for the generator speed starting from an actual value thereof.

Claims

1. A method for operating a wind turbine, the method comprising the steps of: a) detecting at least one of a generator speed (n.sub.Gen(t)) and a wind speed (v.sub.wind(t)); b) on a regular or irregular basis, determining a parameter (A(t)) indicative of a wind hitting the wind turbine from present values for the at least one of a generator speed (n.sub.Gen(t)) and a wind speed (v.sub.wind(t)) at each point in time (t); c) forming a temporal change variable (dA(t)) for the parameter (A(t)) at each point in time (t); d) calculating at least one of a third and fourth statistical moment for the temporal change variables (dA(t)), which occurred in a past time interval ([t−T, t]), wherein the at least one of the third and fourth statistical moment (M.sub.3, M.sub.4) is calculated for a distribution of the temporal change values in the time interval ([t−T, t]); and, e) setting a detection signal (C.sub.EGLM) for an extreme wind gust if only one of the statistical moments (M.sub.3, M.sub.4) exceeds a predetermined cut-in threshold value, wherein said setting the detection signal triggers at least one of the following further method steps for protecting the wind turbine against wind damage: i. increasing a setpoint value for a blade pitch angle (θ.sub.set) causing the blade pitch to increase; and ii. reducing a setpoint value for a generator speed (n.sub.Gen,set) causing the generator speed to decrease.

2. The method of claim 1, wherein the setpoint value for the generator speed (n.sub.Gen,set) is reduced by an offset value (Δn) starting from an actual value for the generator speed (n.sub.Gen,act) or is set to a predetermined value (n.sub.fix).

3. The method of claim 1, wherein the setpoint value for the blade pitch angle (θ.sub.set) is increased by an offset value (Δθ) starting from an actual value for the blade pitch angle (θ.sub.act) or is set to a predetermined value (θ.sub.fix).

4. The method of claim 1, wherein the temporal change variable (dA(t)) is determined in the form of a differential quotient (ΔA) or a temporal derivative (dA/dt).

5. The method of claim 1, wherein the parameter (A(t)) is determined as the product of the generator speed (n.sub.Gen) and wind speed (v.sub.wind).

6. The method of claim 1, wherein the value of the wind speed (v.sub.wind) is a measured or an estimated value.

7. The method of claim 5, wherein the value for the generator speed (n.sub.Gen) is a measured value or a setpoint value for the generator speed specified by a controller.

8. The method of claim 1, wherein the third statistical moment is calculated as M.sub.3=E((X−μ).sup.3), wherein E(.Math.) specifies a formation of an expected value, X specifies values of the distribution and μ=E(X) specifies an expected value of the distribution.

9. The method of claim 1, wherein the fourth statistical moment is calculated as M.sub.4=E((X−μ).sup.4), wherein E(.Math.) specifies a formation of an expected value, X specifies values of the distribution and μ=E(X) an expected value of the distribution.

10. The method of claim 3, wherein the offset value (Δθ) for the blade pitch angle (θ) depends on the value of the fourth statistical moment (M.sub.4), wherein the offset value (Δθ) also increases as the value of the fourth statistical moment (M.sub.4) increases.

11. The method of claim 3, wherein the offset value (Δθ) for the blade pitch angle (θ) depends on the value of an out of plane bending moment of at least one blade, wherein the offset value (Δθ) also increases as the value of the bending increases.

12. The method of claim 1, wherein, if at least one of the statistical moments (M.sub.3, M.sub.4) falls below a predetermined shutdown threshold value, control of the wind turbine returns to previous operation over a predetermined time duration.

13. The method of claim 12, wherein control of the wind turbine only returns to previous operation if the third statistical moment (M.sub.3) is not positive.

14. The method of claim 1, wherein the detection signal (C.sub.EGLM) for an extreme gust can only be generated when the generator speed is greater than a minimum speed.

15. The method of claim 1, wherein the detection signal (C.sub.EGLM) for an extreme gust can only be generated when the wind speed is greater than a minimum wind speed.

16. The method of claim 1, wherein the detection signal (C.sub.EGLM) for an extreme gust can only be generated when the wind turbine is being operated and feeds power into the grid.

17. The method of claim 1, wherein the detection signal (C.sub.EGLM) for an extreme gust can only be generated when the third statistical moment (M.sub.3) is greater than a predetermined minimum value and/or increases.

18. The method of claim 12, wherein the time duration has a value of 10 seconds to 50 seconds.

19. The method of claim 12, wherein the time duration has a value of 20 seconds to 40 seconds.

20. The method of claim 2, wherein the setpoint value for the blade pitch angle (θ.sub.set) is increased by an offset value (Δθ) starting from an actual value for the blade pitch angle (θ.sub.act) or is set to a predetermined value (θ.sub.fix).

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) The invention will now be described with reference to the drawings wherein:

(2) FIG. 1 shows a time series of input variables, their temporal derivative, their mean, their standard deviation as well as their skewness and kurtosis,

(3) FIG. 2 shows the distribution of the derived input variable at point in time t=50 s,

(4) FIG. 3 shows the distribution at point in time t=53.76 s,

(5) FIG. 4 shows the distribution of the derived input variable at t=68 s; and,

(6) FIG. 5 shows the distribution of the derived input variable at point in time t=85 s.

DESCRIPTION OF THE PREFERRED EMBODIMENTS OF THE INVENTION

(7) FIG. 1 shows the temporal progression in a time scale shown in seconds from point in time t=30 s to t=110 s. The upper representation shows the temporal progression of a parameter. The parameter is for example a value for the generator speed and/or for the wind speed. An estimated wind speed can preferably be used as the parameter. The shown diagram shows a dual line progression, wherein the line 10 is based on measured values for the wind speed and the line 12 on the estimated values for the wind speed. It can be clearly seen that the values of the wind estimator progress straighter but apparently also after the point in time indicated by the separation line 14; approximately at 54 s, the wind estimator with its values 12 specifies systematically lower values than the measured wind. The diagram below shows the temporal change in the input signal, wherein it is geared here towards the temporal change in the estimated signal 12. From both curves for the input signal and for its derivative, it is only qualitatively suggested at a high level that a gust occurred at point in time 14. The following diagram is geared towards the mean of the derived input variable. It can be clearly seen that the mean reaches a maximum 16 with a clear temporal delay with respect to the separation line 14.

(8) The standardized skewness of the distribution is plotted in the diagram located further below. Here as well, the skewness clearly reaches its maximum after point in time 14. It is hard to see with the naked eye but the data statistically shows that the skewness here increases faster and thus exceeds a threshold value earlier than the mean with its value 16 and the standard deviation with its value 18. This is also clear in the kurtosis in the diagram located below it. Also here, the maximum value 22 is clearly after point in time 14, but the slope to the maximum value 22 is steeper. For the comparison, please note that the ordinate has a considerably different scale. If the mean is on a scale of 5×10.sup.−3, then the kurtosis moves on a scale of 10, thus a factor that is 2,000 times greater. If one takes into consideration that this concerns the evaluation of statistical data, to which a certain fluctuation adheres, it also becomes clear that, with the skewness and the kurtosis, that is, the third and fourth statistical moments, there are variables that are better suited for a threshold comparison.

(9) FIG. 2 shows the distribution of the wind speed changes at point in time t=50 s. The dashed line shows that a normal distribution of the speed changes is mainly present here. The data was recorded for a time interval of 30 s. In terms of the length of the time interval, it should be taken into account that the longer the time interval the more measured values are available for the statistical evaluation, whereby the values are more accurate. On the other hand, a longer time interval can potentially always indicate a situation in which the wind conditions have changed, whereby the statistical accuracy decreases.

(10) For the present example, a time interval of 30 s proved especially beneficial for the evaluation. However, time intervals of 10 s to 50 s can also be used.

(11) FIG. 3 shows the distribution of the wind speed changes at point in time t=53.76 s. This mainly corresponds with point in time 14 from FIG. 1. It can be clearly seen that larger values, for example values greater than 0.01, occur increasingly in the speed change. It should thus be expected that here the third statistical moment is activated as the measure for the asymmetry of the distribution and provides a clear value.

(12) FIG. 4 shows the point in time t=68 s, which is drawn in FIG. 1 as point in time 24. It can be seen here that even greater positive wind speed changes occur. A large third statistical moment should thus be present. At the same time, it should also be expected that the distribution in FIG. 4 can no longer be interpreted well as a normal distribution; in this respect, the fourth statistical moment should have a positive value as the kurtosis measure.

(13) FIG. 5 shows the distribution of the wind speed change at point in time 26 in FIG. 1. An overhang of negative wind speed changes can be clearly seen; this can be due to the fact that positive speed changes from an earlier point in time fall out of the time frame of the statistics. For example, a time frame of 30 s can be provided so that all speed changes that lie further in the past are no longer taken into consideration. What is to be expected at this point in time is a negative third moment. Even the fourth moment as kurtosis and measure for the deviation from the normal distribution has a peak again here, which is indicated in FIG. 1 with 28.

(14) In the above discussion, the aspect of the standardization of the moments for skewness and kurtosis was not covered. The standardization also depends on the standard deviation so that for turbulent wind with a large standard deviation the higher statistical moments are smaller, which can require an adjustment of threshold values. Furthermore, when calculating the distribution for the past time interval, a time duration of 30 s was assumed and the distribution was calculated continuously. For this, a series of numerical standard methods exist, which allow a continuous, numerical calculation of the distribution.

(15) It is understood that the foregoing description is that of the preferred embodiments of the invention and that various changes and modifications may be made thereto without departing from the spirit and scope of the invention as defined in the appended claims.