Method for controlling a wind energy farm taking wake effects into account

Abstract

A method for controlling a wind energy farm is disclosed. A wake state of the wind energy farm is determined, including determining wake chains defining wake relationships among the wind turbines of the wind farm under the current wind conditions. For at least one of the wind turbines of the wind energy farm, a lifetime usage is estimated, based on an accumulated load measure for the wind turbine. In the case that the estimated lifetime usage is below a predefined lifetime usage limit, the wind turbine is operated in an overrated state, while monitoring wake effects at each of the wind turbines. In the case that a downstream wind turbine detects wake effects above a predefined wake threshold level, at least one wind turbine having an upstream wake relationship with the downstream wind turbine is requested to decrease the generated wake, e.g. by decreasing power production.

Claims

1. A method for controlling a wind energy farm, the wind energy farm comprising a plurality of wind turbines, the method comprising: determining a wake state of the wind energy farm, including determining wake chains defining wake relationships among the plurality of wind turbines of the wind energy farm under current wind conditions, wherein a first wind turbine of the plurality of wind turbines comprises at least one downstream wind turbine in a first wake chain, and wherein a second wind turbine is a leaf node in a second wake chain, for at least the first wind turbine of the wind energy farm, estimating a first lifetime usage, based on an accumulated load measure for the first wind turbine, and in a case that the first estimated lifetime usage is below a first predefined lifetime usage limit, operating the first wind turbine in a first overrated state, estimating a second lifetime usage, based on an accumulated load measure for the second wind turbine, and in a case that the second estimated lifetime usage is below a second predefined lifetime usage limit, operating the second wind turbine in a second overrated state, at a first time and upon operating the first wind turbine operating in the first overrated state, monitoring wake effects due to at least the first wind turbine operating in the first overrated state at each of the wind turbines of the wind energy farm, in a case that a downstream wind turbine detects wake effects above a predefined wake threshold level, generating a control signal for the first wind turbine operating in the first overrated state, wherein the control signal requests a decrease in generated wake from the first wind turbine, and wherein the first wind turbine operating in the first overrated state comprises an upstream wake relationship with the downstream wind turbine, at a second time, controlling the first wind turbine operating in the first overrated state in accordance with the generated control signal, and operating the second wind turbine in accordance with the second overrated state.

2. The method of claim 1, wherein the monitoring wake effects at each of the wind turbines comprises monitoring loads on one or more parts of the plurality of wind turbines.

3. The method of claim 1, wherein the determining a wake state of the wind energy farm comprises detecting wake effects at the plurality of wind turbines of the wind energy farm.

4. The method of claim 1, wherein the estimating a lifetime usage for a given wind turbine comprises: measuring or estimating a bending moment of the wind turbine, calculating a fatigue load on the wind turbine, based on the measured or estimated bending moment, and comparing the calculated fatigue load to an expected fatigue load on the wind turbine, based on an age of the wind turbine.

5. The method of claim 4, wherein the bending moment is a tower bottom bending moment.

6. The method of claim 4, wherein the calculating a fatigue load on the wind turbine is performed using a rainflow count.

7. The method of claim 1, further comprising, in the case that a wind turbine is operated in an overrated state, re-estimating the first estimated lifetime usage, and in the case that the estimated lifetime usage reaches the predefined lifetime usage limit, discontinuing operating the wind turbine in the overrated state.

8. The method of claim 1, wherein the generating the control signal further comprises: the downstream wind turbine forwarding a wake detected signal to a central wind energy farm controller, the wake detected signal indicating that the downstream wind turbine has detected wake effects above a predefined wake threshold level, in response to receiving the wake detected signal, the central wind energy farm controller identifying at least one wind turbine of the wind energy farm having an upstream wake relationship with the downstream wind turbine, based on the determined wake state of the wind energy farm, and the central wind energy farm controller generating a control signal for at least one of the identified wind turbine(s) and forwarding the generated control signals to the identified wind turbine(s).

9. The method of claim 8, wherein at least one of the generated control signals request a decrease in power production of an upstream wind turbine.

10. The method of claim 1, further comprising identifying at least one wind turbine which does not have an upstream wake relationship with any of the other wind turbines of the wind energy farm, and wherein the estimating a lifetime usage for at least one of the wind turbines of the wind energy farm includes estimating a lifetime usage for the at least one identified wind turbine.

11. A wind energy farm, comprising: a plurality of wind turbines; and a control system configured to control the plurality of wind turbines by performing an operation, comprising: determining a wake state of the wind energy farm, including determining wake chains defining wake relationships among the plurality of wind turbines under current wind conditions, for at least a first wind turbine of the plurality of wind turbines, estimating a first lifetime usage, based on an accumulated load measure for the first wind turbine, and in a case that the first estimated lifetime usage is below a first predefined lifetime usage limit, operating the first wind turbine in a first overrated state, estimating a second lifetime usage, based on an accumulated load measure for a second wind turbine of the plurality of wind turbines, and in a case that the second estimated lifetime usage is below a second predefined lifetime usage limit, operating the second wind turbine in a second overrated state, at a first time and upon operating the first wind turbine operating in the first overrated state, monitoring wake effects at each of the plurality of wind turbines due to at least the first wind turbine operating in the first overrated state, upon a downstream wind turbine detecting wake effects above a predefined wake threshold level, generating a control signal for the first wind turbine operating in the first overrated state, the control signal requesting a decrease in generated wake from the first wind turbine, wherein the first wind turbine operating in the first overrated state comprises an upstream wake relationship with the downstream wind turbine, at a second time, controlling the first wind turbine operating in the first overrated state in accordance with the generated control signal, and operating the second wind turbine in accordance with the second overrated state.

12. The wind energy farm of claim 11, wherein the monitoring wake effects at each of the plurality of wind turbines comprises monitoring loads on one or more parts of the plurality of wind turbines.

13. The wind energy farm of claim 11, wherein the determining a wake state of the wind energy farm comprises detecting wake effects at the plurality of wind turbines.

14. The wind energy farm of claim 11, wherein the estimating a lifetime usage for a given wind turbine comprises: determining a bending moment of the wind turbine, calculating a fatigue load on the wind turbine, based on the determined bending moment, and comparing the calculated fatigue load to an expected fatigue load on the wind turbine, based on an age of the wind turbine.

15. The wind energy farm of claim 14, wherein the bending moment is a tower bottom bending moment.

16. The wind energy farm of claim 14, wherein the calculating a fatigue load on the wind turbine is performed using a rainflow count.

17. The wind energy farm of claim 11, further comprising: upon operating in an overrated state, re-estimating the first estimatd lifetime usage; and upon the re-estimated lifetime usage reaching the first predefined lifetime usage limit, discontinuing operating the first wind turbine in the overrated state.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) The invention will now be described in further detail with reference to the accompanying drawings in which

(2) FIGS. 1-3 illustrate wake chains in a wind energy farm being controlled in accordance with a method according to an embodiment of the invention,

(3) FIG. 4 is a block diagram illustrating a method according to an embodiment of the invention,

(4) FIG. 5 illustrates determination of lifetime usage for use in a method according to an embodiment of the invention, and

(5) FIG. 6 is a flow chart illustrating a method according to an embodiment of the invention.

DETAILED DESCRIPTION OF THE DRAWINGS

(6) FIGS. 1-3 illustrate four wind turbines 1a, 1b, 1c, 1d arranged in a wind energy farm being controlled in accordance with a method according to an embodiment of the invention. FIGS. 1-3 illustrate wake chains of the wind turbines 1a, 1b, 1c, 1d under various operating conditions.

(7) The direction of the incoming wind is illustrated by arrow 2, and the wake generated by the respective wind turbines 1a, 1b, 1c, 1d is illustrated by cones 3 connected to the respective wind turbines 1a, 1b, 1c, 1d.

(8) In the situation illustrated in FIG. 1, all of the wind turbines 1a, 1b, 1c, 1d are properly yawed into the incoming wind 2, i.e. the rotor of each wind turbine 1a, 1b, 1c, 1d is directly facing the incoming wind 2. Accordingly, the cones 3 illustrating the wake generated by the wind turbines 1a, 1b, 1c, 1d are pointing substantially in the same direction as the incoming wind 2.

(9) When one wind turbine 1a, 1b, 1c, 1d is arranged in the wake created by one of the other wind turbines 1a, 1b, 1c, 1d, a wake relationship exists among the two wind turbines 1a, 1b, 1c, 1d. In the situation illustrated in FIG. 1, wind turbine 1d is arranged in the wake of wind turbines 1a, 1b and 1c. Thus, wind turbine 1d may be regarded as a downstream wind turbine with respect to each of wind turbines 1a, 1b and 1c, and each of wind turbines 1a, 1b and 1c may be regarded as upstream wind turbines with respect to wind turbine 1d. No wind turbine is arranged in the wake of wind turbine 1d, and therefore wind turbine 1d is not an upstream wind turbine with respect to one or more other wind turbines when the wind direction is as shown in FIG. 1.

(10) Similarly, wind turbine 1c is arranged in the wake of wind turbine 1a. Accordingly, wind turbine 1c may be regarded as a downstream wind turbine with respect to wind turbine 1a, and wind turbine 1a may be regarded as an upstream wind turbine with respect to wind turbine 1c. And, as described above, wind turbine 1c may also be regarded as an upstream wind turbine with respect to wind turbine 1d.

(11) Wind turbines 1a and 1b are not arranged in the wake of any other wind turbine. Accordingly, wind turbines 1a and 1b are not downstream wind turbines with respect to one or more other wind turbines. However, wind turbine 1d is arranged in the wake of wind turbine 1b, and wind turbines 1c and 1d are both arranged in the wake of wind turbine 1a. Accordingly, as described above, wind turbine 1b may be regarded as an upstream wind turbine with respect to wind turbine 1d, and wind turbine 1a may be regarded as an upstream wind turbine with respect to each of wind turbines 1c and 1d.

(12) The wake relationships described above are illustrated by arrows 4, indicating which of the wind turbines 1a, 1b, 1c, 1d are affected by wake effects generated by which other wind turbines 1a, 1b, 1c, 1d. The arrows 4 may be referred to as wake chains.

(13) In the situation illustrated in FIG. 2, the wind direction 2 is substantially the same as in the situation illustrated in FIG. 1. However, in the situation illustrated in FIG. 2, a yaw error is present in wind turbine 1c, i.e. the rotor of wind turbine 1c is not pointing directly into the incoming wind 2. As a consequence, the cone 3 illustrating the wake created by wind turbine 1c is not pointing in the same direction as the incoming wind 2. Thereby wind turbine 1d is no longer arranged in the wake of wind turbine 1c. The wake relationships of the wind turbines 1a, 1b, 1c, 1d in FIG. 2 are thereby as follows. Wind turbine 1d is a downstream wind turbine with respect to wind turbines 1a and 1b, and is not an upstream wind turbine with respect to any other wind turbine. Wind turbine 1c is a downstream wind turbine with respect to wind turbine 1a, and is not an upstream wind turbine with respect to any other wind turbine. Wind turbine 1b is an upstream wind turbine with respect to wind turbine 1d, and is not a downstream wind turbine with respect to any other wind turbine. Finally, wind turbine 1a is an upstream wind turbine with respect to wind turbines 1c and 1d, and is not a downstream wind turbine with respect to any other wind turbine.

(14) In the situation illustrated in FIG. 3, the direction of the incoming wind 2 has changed as compared to the situation illustrated in FIG. 1. All of the wind turbines 1a, 1b, 1c, 1d are properly yawed into the incoming wind 2, i.e. the rotor of each wind turbine 1a, 1b, 1c, 1d is directly facing the incoming wind 2.

(15) As a consequence of the changed direction of the incoming wind 2, the directions of the cones 3 illustrating the wake generated by the wind turbines 1a, 1b, 1c, 1d have also been changed. Thereby the wake relationships of the wind turbines 1a, 1b, 1c, 1d in FIG. 3 are as follows. Wind turbine 1d is a downstream wind turbine with respect to wind turbine 1b, and is not an upstream wind turbine with respect to any other wind turbine. Wind turbine 1c is a downstream wind turbine with respect to wind turbine 1b, and is not an upstream wind turbine with respect to any other wind turbine. Wind turbine 1b is an upstream wind turbine with respect to wind turbines 1c and 1d, and is not a downstream wind turbine with respect to any other wind turbine. Finally, wind turbine 1a is neither a downstream wind turbine nor an upstream wind turbine with respect to any other wind turbine.

(16) The wake chains illustrated in FIGS. 1-3 are, thus, defined by the current wind conditions, including the current wind direction 2. Furthermore, the current wind speed, and possibly the prevailing turbulence conditions, will have an influence on the lengths of the cones 3, and will thereby have an impact on the wake chains.

(17) The wake chains may be obtained by detecting wake effects at each of the wind turbines 1a, 1b, 1c, 1d. This could, e.g., include monitoring loads on one or more parts of the wind turbines 1a, 1b, 1c, 1d which are characteristic for specific wake effects, such as increased turbulence, occurring at the position of the wind turbine 1a, 1b, 1c, 1d.

(18) FIG. 4 is a block diagram illustrating control of a wind energy farm in accordance with a method according to an embodiment of the invention. The wind energy farm comprises a number of wind turbines 1, four of which are illustrated. The wind energy farm further comprises a central wind energy farm controller 5 arranged to handle the overall control of the wind energy farm, including coordination of the control of the individual wind turbines 1 of the wind energy farm. Accordingly, each of the wind turbines 1 of the wind energy farm is arranged to communicate with the central wind energy farm controller 5.

(19) In the method illustrated in FIG. 4, when one of the wind turbines 1 of the wind energy farm detects wake effects above a predefined threshold, it forwards a wake detected signal to the central wind energy farm controller 5. The detected wake effects could, e.g., be in the form of a detection of specific loads on one or more parts of the wind turbine 1, the loads being characteristic for undesired wake effects occurring at the wind turbine 1. Alternatively or additionally, the detected wake effects could include detecting a certain turbulence pattern at the wind turbine 1.

(20) In response to the receipt of a wake detected signal from a wind turbine 1, the central power plant controller 5 identifies at least one wind turbine 1 of the wind energy farm which has an upstream wake relationship with the wind turbine 1 which forwarded the wake detected signal. Accordingly, at least some of the wind turbines 1 which contribute to the wake occurring at the wind turbine 1 which forwarded the wake detected signal are identified. This identification could, e.g., be performed using previously determined wake chains of the kind illustrated in FIGS. 1-3. All wind turbines 1 which contribute to the wake may be identified, or only some of these wind turbines 1 may be identified, e.g. the ones providing the largest contributions.

(21) Then the central wind energy farm controller 5 generates a control signal for at least one of the identified upstream wind turbines 1 and forwards the generated control signals to the respective upstream wind turbines 1. The control signals request the upstream wind turbines 1 to decrease the generated wake. This could, e.g., include decreasing the power production of the wind turbines 1. Control signals may be generated for all of the identified upstream wind turbines 1, or only for some of the identified upstream wind turbines 1, e.g. the ones providing the largest contributions to the wake effects.

(22) Furthermore, the central wind energy farm controller 5 may request one or more of the wind turbines 1 of the wind energy farm to operate in an overrated state if it can be assumed that this will not decrease the expected lifetime of the wind turbines 1. To this end the lifetime usage for the wind turbines 1 is estimated, based on an accumulated load measure for the wind turbines 1. If the estimated lifetime usage for a given wind turbine 1 is below a predefined lifetime usage limit, it is considered safe to operate the wind turbine 1 in an overrated state, and the central wind energy farm controller 5 therefore forwards a control signal to the wind turbine 1, requesting it to operate in an overrated state. However, if a downstream wind turbine 1 subsequently detects that this results in wake effects above a predefined wake threshold level, then the wind energy farm controller 5 may request the upstream wind turbine 1 to stop operating in an overrated state.

(23) FIG. 5 shows three curves, illustrating determination of lifetime usage for use in a method according to an embodiment of the invention. The upper curve 6 shows measured generator torque of a wind turbine as a function of time. The middle curve 7 shows an estimated damage rate on the generator of a wind turbine, caused by the generator torque, as a function of time. The lower curve 8 shows accumulated damage on the generator as a function of time.

(24) It can be seen from the upper curve 6 that the generator torque is substantially constant for most of the time. However, an increase in generator torque appears during a time period from approximately t=54 to approximately t=85. This could, e.g., be due to the wind turbine being operated in an overrated state, or due to the wind turbine being arranged in the wake of one or more other wind turbines.

(25) It can be seen from the middle curve 7 that the increase in generator torque results in a corresponding increase in the estimated damage rate on the generator.

(26) It can be seen from the lower curve 8 that the accumulated damage on the generator increases steadily for most of the time. However, during the time interval where the generator torque is increased, the accumulated damage on the generator increases faster, reflecting the increased estimated damage rate illustrated in the middle curve 7.

(27) The accumulated damage on the generator illustrated in the lower curve 8 may be used for estimating a lifetime usage for the wind turbine. In order to ensure that the wind turbine is able to operate during an entire design lifetime of the wind turbine without introducing excessive fatigue loads, the accumulated damage should be maintained below a level corresponding to an expected accumulated damage at any given time during the lifetime of the wind turbine. If the wind turbine has, for a period of time, been operated with a damage rate which is below a design damage rate, the wind turbine can subsequently be operated at a higher damage rate for a limited time period, without causing the accumulated damage to exceed the design accumulated damage at that point in time. Thus, when this is the case, the wind turbine may, e.g., be allowed to operate in an overrated state. Furthermore, an increased damage rate may be allowed under certain wind conditions or if the wind turbine can subsequently be operated in a less aggressive manner.

(28) FIG. 6 is a flow chart illustrating a method according to an embodiment of the invention. The process is started at step 9. At step 10 a wake state for the wind energy farm is determined. This could, e.g., be done in the manner described above with reference to FIGS. 1-3.

(29) At step 11, a lifetime usage for one of the wind turbines of the wind energy farm is determined. This could, e.g., be done in the manner described above with reference to FIG. 5.

(30) At step 12 it is investigated whether or not the estimated lifetime usage is below a predefined lifetime usage limit. If this is not the case, operation of the wind turbine in an overrated state will most likely decrease the expected lifetime of the wind turbine. Therefore, in this case, the process is forwarded to step 13, and the wind turbine is operated in a normal manner, and the process is ended at step 14.

(31) In the case that step 12 reveals that the estimated lifetime usage is below the predefined lifetime usage limit, it is considered safe to operate the wind turbine in an overrated state, in the sense that this is not expected to cause a significant decrease in the expected lifetime of the wind turbine. Therefore, in this case, the process is forwarded to step 15, and the wind turbine is operated in an overrated state, thereby increasing the total power output of the wind energy farm.

(32) While the wind turbine is operated in the overrated state, wake effects are monitored at the wind turbines arranged downstream with respect to the wind turbine, at step 16. This could, e.g., include monitoring various loads on the downstream wind turbines.

(33) At step 17 it is investigated whether or not one or more of the downstream wind turbines has detected wake effects above a predefined wake threshold level. If this is not the case, it is considered safe to continue operating the wind turbine in the overrated state, and the process is therefore returned to step 15.

(34) In the case that step 17 reveals that one or more of the downstream wind turbines has detected wake effects above the predefined wake threshold level, it is no longer considered safe to operate the wind turbine in the overrated state. Therefore the process is, in this case, forwarded to step 18, where the power production of the wind turbine is decreased, e.g. to the nominal power production level, and the process is ended at step 14.

(35) Steps 11-18 may be repeated or performed simultaneously for one or more further wind turbines of the wind turbines of the wind energy farm, e.g. for each of the wind turbines of the wind energy farm.