Computer implemented method for controlling one or more wind turbines in a wind farm

By leveraging communication and environmental data evaluation between wind turbines in a wind farm, control commands are generated to adjust the control parameters of the wind turbine, solving the problem of excessive load caused by gusts or waves, extending component life and reducing power costs.

CN115244296BActive Publication Date: 2024-11-12SIEMENS GAMESA RENEWABLE ENERGY AS
View PDF 10 Cites 0 Cited by

Patent Information

Application Number
CN202180022325.8
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Priority Date
2020-03-19
Filing Date
2021-03-04
Publication Date
2024-11-12
Estimated Expiration
2041-03-04

AI Technical Summary

Technical Problem

In environmental events such as gusts or waves, wind turbines at the wind farm need to be controlled instantly to avoid excessive loads and component failures, but hard shutdowns lead to huge loads and reduced life.

Method used

By setting up communication between wind turbines upstream and downstream in the wind farm, the possibility of predetermined events is evaluated using environmental data and stress data, and a control command is generated to adjust the control parameters of the wind turbine to avoid excessive loads, including shutdowns, soft shutdowns, , adjust the yaw angle and pitch angle and other countermeasures.

Benefits of technology

It effectively reduces the excessive load of wind turbines, extends component life, reduces the cost of power equalization, and improves the overall performance of the wind farm.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115244296B_ABST
    Figure CN115244296B_ABST
Patent Text Reader

Abstract

The invention relates to a method for computer-implemented control of one or more wind turbines (1, 2) in a wind farm. The wind farm comprises a first upstream wind turbine (1) and a second downstream wind turbine (2), wherein the following steps are performed at each of one or more time points during operation of the wind farm: i) obtaining environmental data (ED) and stress data (SD) of the first wind turbine (1), the environmental data (ED) and the stress data (SD) being obtained by sensors (13) installed at the first wind turbine (1); ii) determining status information (STI), the status information indicating whether a predetermined event exists when the environmental data (ED) and the stress data (SD) are obtained, wherein the predetermined event is The invention relates to a method for controlling the first wind turbine (1) immediately; iii) broadcasting a message (M) into the wind farm, wherein the message (M) contains environmental data (ED) and a timestamp (TS) as event information (EVI); iv) evaluating the event information (EVI), wherein the event information (EVI) is whether the predetermined event at the first wind turbine (1) will affect the second wind turbine (2); v) if the evaluation indicates that the predetermined event will affect the second wind turbine (2), generating a control command (CO) for controlling the second wind turbine (2) to counteract the predetermined event.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to a computer-implemented method and system for controlling one or more wind turbines in a wind farm, wherein the wind farm comprises a first upstream wind turbine and a second downstream wind turbine. Background Art

[0002] As known to those skilled in the art, a wind turbine of a wind farm comprises an upper section having a rotor and a nacelle located on top of a tower, wherein the upper section can be rotated around a vertical yaw axis in order to change the yaw angle of the respective turbine. The yaw angle of the wind turbine is usually adjusted so that the rotor of the wind turbine faces the wind. To this end, a wind sensor (i.e. anemometer) is installed on the respective wind turbine to estimate the wind direction. Generally, in order to maximize the generation of electrical energy in a wind farm, yaw misalignment should be avoided. Therefore, the wind turbine is controlled so that the yaw misalignment angle is as small as possible.

[0003] There are environmental events such as gusts or waves that require immediate control of the wind turbine, such as a hard stop, in order to avoid exceeding critical loads and resulting component failure. However, a hard stop results in a huge load on the turbine and may reduce the overall lifespan.

[0004] EP 2788620 B1 discloses a method for warning wind turbine generators in a wind farm of extreme wind events. The wind farm comprises a group of one or more first type WTGs (wind turbine generators) located in an inner area of ​​the wind farm and a plurality of second type WTGs located in an outer area of ​​the wind farm at least partially surrounding the inner area. The first type WTGs are structurally weaker than the second type WTGs and / or rated for higher performance than the second type WTGs.

[0005] EP 1790851 A2 discloses a method and system for controlling a wind farm power plant, comprising a central processing and control unit operably coupled to wind turbines in the wind farm to receive and send data and control signals to each wind turbine, thereby reducing fatigue loads and complying with power limits. Summary of the invention

[0006] The object of the present invention is to provide a simple method in order to avoid significant loads on the turbines of a wind park in the event of environmental events such as wind gusts or waves.

[0007] This object is achieved by the independent claim. Preferred embodiments of the invention are defined in the dependent claims.

[0008] The present invention provides a computer-implemented method for controlling one or more wind turbines in a wind farm. The wind farm comprises an upstream first wind turbine and a downstream second wind turbine, i.e., a pair of first and second wind turbines. The method is applied to this pair, and may also be applied to several such pairs. Alternatively, the method naturally applies to one or more upstream first wind turbines and two or more downstream second wind turbines. Each of the first and second wind turbines comprises an upper section at the top of a tower, which is pivotable about a vertical yaw axis and has a nacelle and a rotor with rotor blades. The rotor is attached to the nacelle, and the rotor blades are rotatable by the wind about a horizontal rotor axis.

[0009] According to the method of the invention, the following steps i) to v) are performed at each of one or more time points during the operation of the wind farm.

[0010] In step i), environmental data and stress data of the first wind turbine are obtained. In the following, the terms "environmental data" and / or "stress data" refer to digital data. The term "obtaining environmental data and stress data" means that the data is received by a computing unit that implements the method of the present invention. The environmental data and stress data are current data (i.e., the latest data) acquired by one or more first sensors installed at the first wind turbine. The term "sensor" refers to any sensing device configured to acquire environmental data or operating data of a wind turbine, the environmental data such as wind information, temperature, etc., and the operating data such as rotation speed, yaw angle, pitch angle, strain, vibration of the tower, vibration of the nacelle, etc. The acquired environmental data and stress data make it possible to determine whether the actual control parameters of the first wind turbine result in the requested or predetermined performance and whether they are within a predetermined load range.

[0011] In step ii), state information is determined based on the acquired environmental and stress data, the state information indicating whether a predetermined event exists when acquiring the environmental and stress data. The predetermined event requires immediate control of the first wind turbine to avoid excessive loads on the turbine and resulting component failure.

[0012] In step iii), in case the status information indicates a predetermined event, a message is broadcasted into the wind farm. The broadcast message contains at least some of the environmental data and a time step as event information.

[0013] In step iv) it is evaluated upon receipt of the broadcast message whether the predetermined event at the first wind turbine will affect the event information of the second wind turbine.

[0014] In step v), in case the evaluation indicates that the predetermined event will affect the second wind turbine, a control command for controlling the second wind turbine to counteract the predetermined event is generated.

[0015] According to the invention, evaluating the event information consists of or includes determining a probability value of whether the event information is true, wherein generating a control command is initiated only when the probability value exceeds a predetermined probability threshold.

[0016] The method of the invention provides a simple and direct method for controlling a first and / or a second wind turbine in the event that the first wind turbine is subjected to a predetermined event requiring immediate control of the first wind turbine. To this end, in the event that the status information generated by the evaluation of the environmental data and the stress data indicates a predetermined event, the acquired environmental data and stress data of the first wind turbine are broadcast to the wind farm. If the evaluation indicates that the predetermined event will also affect the second wind turbine, countermeasures can be taken by generating a control command that adjusts one or more control parameters of the second wind turbine to avoid excessive load and damage.

[0017] The method directly or indirectly utilizes the communication between the first wind turbine and the second wind turbine. The method provides a warning system for the second wind turbine by utilizing a broadcast message containing all necessary information to evaluate whether a critical predetermined event is relevant to the operation of the second wind turbine itself. In particular, this enables an earlier reaction to critical events. A reduction in component failures, an improvement in performance or an increase in lifetime will result from the adapted control strategy. As a further advantage, the levelized costs of electricity (LCOE) can be reduced.

[0018] The probability value for determining whether the event information is true or not enhances the correctness of the decision. If the environmental data acquired by the first wind turbine is incorrect, for example due to a fault in a sensor, this will lead to erroneous and undesirable control commands. By introducing probability values, significant deficiencies in the environmental data can be taken into account.

[0019] In a preferred embodiment, the message contains wind direction and wind speed as environmental data.Knowledge of wind direction and wind speed enables in conjunction with the timestamp to assess whether a predetermined event at a first wind turbine will affect a second wind turbine in the near future.

[0020] According to a further preferred embodiment, determining the state information is performed by a computing unit of the first wind turbine. The computing unit of the first wind turbine processes the environmental data and the stress data to determine the state information, which indicates whether a predetermined event was present when the environmental data and the stress data were acquired. This corresponds to the usual behavior of a wind turbine, which processes the available information to control the wind turbine by itself. In addition to this usual behavior, the first wind turbine broadcasts a message into the wind farm. The information contained in the broadcast message enables a receiver to assess whether a predetermined event at the first wind turbine will affect a second wind turbine.

[0021] In an alternative embodiment, determining the state information is performed by a central computing unit. According to this embodiment, the acquired environmental data and stress data are transmitted to the central computing unit by a broadcast message or a unicast message, and the central computing unit determines the state information based on the environmental data and the stress data. The central computing unit may be a computing unit of the wind farm or any other computing unit in the cloud.

[0022] In another preferred embodiment, broadcasting the message into the wind park is performed by a sending unit of the first wind turbine.The sending unit may be a communication unit configured to only send data or to send and receive data, ie configured as a transceiver.

[0023] In another preferred embodiment, the evaluation of the event information is performed by the computing unit of the second wind turbine. The broadcast message is received by the second wind turbine, which evaluates the event information of the broadcast message itself and, if the evaluation indicates that the predetermined event will affect the second wind turbine, generates a control command with control parameters for controlling the second wind turbine. According to this embodiment, the analysis of the available data is performed directly at the second wind turbine. It defines a reasonable response by itself. This may be beneficial because real-time reactions are required. In addition, this embodiment does not require a network connection to the cloud and processing power in the cloud. Instead, the second wind turbine acquires all information, processes it and defines the best response. Real-time computing is available because powerful hardware is available today to perform so-called edge computing. Performing the evaluation of the event information by a central computing unit has the advantage of powerful processing units of these control units. However, it must be ensured that all information is exchanged between the first wind turbine and the second wind turbine in real time.

[0024] In an alternative embodiment, evaluating the event information is performed by a computing unit of the central computing unit. According to this embodiment, the event information is received and evaluated by the central computing unit. In case the evaluation indicates that the predetermined event will affect the second wind turbine, a control command with control parameters for controlling the second wind turbine is generated by the central computing unit. According to this embodiment, the computing power of the central computing unit can be used.

[0025] According to another preferred embodiment, evaluating the event information consists of or includes determining a time period starting from the time when the predetermined event at the first wind turbine has been determined until the predetermined event will affect the second wind turbine. Depending on the determined time period, appropriate countermeasures can be taken. The determination of the time period until the predetermined event will affect the second wind turbine can be calculated based on the timestamp, wind direction, wind speed information and the known distance between the first wind turbine and the second wind turbine.

[0026] This probability is based on "historical" information of the turbine credibility. For example, sensor readings over time can be analyzed by the turbine that sent the event and this information can be part of the broadcast message. For example: if there is a sudden (non-physical) jump in the sensor reading, it can send the event message together with information that it may be a sensor error and therefore not relevant to the receiver (i.e. the second wind turbine or the central computing unit). And it can also be achieved that the receiver tracks the events and the actual importance of the events. If the broadcasting turbine (i.e. the first wind turbine) sends a signal that is proven to be irrelevant to the receiver, the receiver reduces the credibility (probability) of the transmitter.

[0027] According to another preferred embodiment, evaluating the event information and / or generating the control commands is based on processing the event information according to a predetermined mapping. Using the predetermined mapping is based on the knowledge of whether and how the second wind turbine reacts to the environmental information within the broadcast message. The reaction is of a "hard-coded" type in the mapping. In other words, the control commands are defined in advance with respect to each possible situation (event). This is a classic way of defining a control strategy.

[0028] According to an alternative embodiment, evaluating the event information is based on processing the event information by a trained data driven model, wherein the event information is fed as a digital input to the trained data driven model and the trained data driven model provides as a digital output information whether a predetermined event of the first wind turbine will affect the second wind turbine. Any known data driven model learned by machine learning may be used in the method according to the invention.

[0029] In a particularly preferred embodiment, the trained data-driven model is a neural network, preferably a recursive neural network. However, other trained data-driven models, such as reinforcement learning, may also be implemented in the method of the invention. Each time the second wind turbine receives a broadcast message, it has to make a decision whether the information in the broadcast message has to be taken into account for its control strategy. The second wind turbine stores all its reactions or non-reactions and knows retrospectively whether the decision was good or bad. Thus, the second wind turbine can learn from its experience of previous events and take this into account for further operations. Such reinforcement learning may also be applied to the central computing unit.

[0030] According to another preferred embodiment, the control command consists of or includes at least one of the following countermeasures: shutdown; soft shutdown; adjusting the yaw angle; adjusting the pitch angle. In addition, a combination of countermeasures can be applied. Shutdown and soft shutdown mainly differ based on the time taken from triggering the stop command until the turbine stops completely. Soft shutdown allows a smoother ramp down than hard shutdown. These countermeasures can be initiated by corresponding control commands generated by the central computing unit or the second wind turbine itself. Taking into account the time period until the predetermined event at the first wind turbine will affect the second wind turbine, those countermeasures that have the least impact on the negative load can be selected. Such countermeasures are more "friendly" to the second wind turbine and its components, so that the life of the components can be extended.

[0031] According to another preferred embodiment, the control command is generated by the second wind turbine.Alternatively, the control command is generated by a central computing unit, wherein the control command is sent to the second wind turbine via a broadcast message or a unicast message.

[0032] In addition to the above method, the invention relates to a computer implemented system for controlling one or more wind turbines in a wind park, wherein the system is configured to perform the method according to the invention or one or more embodiments of the method according to the invention.

[0033] Furthermore, the invention relates to a computer program product with a program code stored on a non-transitory machine-readable carrier, the computer program product being configured for performing the method according to the invention or one or more preferred embodiments thereof when the program code is executed on a computer.

[0034] Furthermore, the invention relates to a computer program having a program code for executing the method according to the invention or one or more preferred embodiments thereof, when the program code is executed on a computer.

[0035] The present invention also includes the following scheme:

[0036] Solution 1. A computer-implemented method for controlling one or more wind turbines 1, 2 in a wind farm, wherein the wind farm comprises a first wind turbine 1 upstream and a second wind turbine 2 downstream, wherein at each of one or more time points during operation of the wind farm, the following steps are performed:

[0037] i) obtaining environment data ED and stress data SD of the first wind turbine 1, wherein the environment data ED and the stress data SD are current data acquired by one or more sensors 13 installed at the first wind turbine 1;

[0038] ii) determining, based on the environment data ED and the stress data SD, state information STI indicating whether a predetermined event existed when acquiring the environment data ED and the stress data SD, wherein the predetermined event requires immediate control of the first wind turbine 1;

[0039] iii) in case the status information STI indicates the predetermined event, broadcasting a message M to the wind farm, wherein the message M contains at least some of the environmental data ED and a timestamp TS as event information EVI;

[0040] iv) upon receiving the broadcast message M, evaluating said event information EVI, said event information EVI being whether said predetermined event at said first wind turbine 1 will affect said second wind turbine 2;

[0041] v) generating a control command CO for controlling the second wind turbine 2 to counteract the predetermined event if the evaluation indicates that the predetermined event will affect the second wind turbine 2,

[0042] The evaluating of the event information EVI includes determining a probability value of whether the event information EVI is true, wherein the generation of the control command CO is initiated only when the probability value exceeds a predetermined probability threshold.

[0043] Solution 2. The method according to Solution 1, wherein the message includes wind direction WD and wind speed WS as the environmental data ED.

[0044] Solution 3. The method according to solution 1 or 2, wherein determining the state information STI is performed by a central computing unit CCU or a computing unit 14 of the first wind turbine 1 .

[0045] Solution 4. The method according to one of the preceding solutions, wherein broadcasting the message M into the wind farm is performed by a sending unit 15 of the first wind turbine 1 .

[0046] Aspect 5. The method according to one of the preceding aspects, wherein evaluating the event information EVI is performed by the central computing unit CCU or the computing unit 24 of the second wind turbine 2 .

[0047] Aspect 6. The method according to one of the preceding aspects, wherein evaluating the event information EVI comprises determining a time period starting from the time when the predetermined event has been determined at the first wind turbine 1 until the predetermined event will affect the second wind turbine 2 .

[0048] Solution 7. The method according to one of the preceding solutions, wherein evaluating the event information EVI and / or generating the control command CO is based on processing the event information EVI according to a predetermined mapping.

[0049] Option 8. A method according to one of the preceding options, wherein evaluating the event information EVI is based on processing the event information EVI through a trained data-driven model MO, wherein the event information EVI is fed to the trained data-driven model MO as a digital input, and the trained data-driven model MO provides information as a digital output whether the predetermined event at the first wind turbine 1 will affect the second wind turbine 2.

[0050] Solution 9. The method according to one of the preceding solutions, wherein the control command CO comprises at least one of the following countermeasures:

[0051] - downtime;

[0052] -Soft shutdown;

[0053] -Adjust the yaw angle;

[0054] -Adjust the pitch angle,

[0055] In particular, wherein the stop and the soft stop differ primarily based on the time taken from triggering the stop command until the received turbine 1, 2 is completely stopped,

[0056] In particular, wherein the soft stop allows a smoother ramp down than a hard stop,

[0057] In particular, countermeasures are selected such that the impact on negative loads is minimized, and / or

[0058] In particular, wherein selecting a countermeasure comprises considering a time period until a predetermined event at the first wind turbine affects the second wind turbine.

[0059] Aspect 10. The method according to any one of aspects 1 to 9, wherein the control command CO is generated by the second wind turbine 2 .

[0060] Solution 11. The method according to one of the above-mentioned solutions 1 to 9, wherein the control command CO is generated by the central computing unit CCU, wherein the control command CO is sent to the second wind turbine via a broadcast message or a unicast message M2.

[0061] Solution 12. A system for computer-implemented monitoring of one or more wind turbines 1, 2 in a wind farm, wherein the wind farm comprises a first wind turbine 1 upstream and a second wind turbine 2 downstream, wherein the system is configured to perform the following steps at each of one or more time points during operation of the wind farm:

[0062] i) obtaining environment data ED and stress data SD of the first wind turbine 1, wherein the environment data ED and the stress data SD are current data acquired by one or more sensors 13 installed at the first wind turbine 1;

[0063] ii) determining, based on the environment data ED and the stress data SD, state information STI indicating whether a predetermined event exists when acquiring the environment data ED and the stress data SD, wherein the predetermined event requires immediate control of the first wind turbine 1;

[0064] iii) in case the status information STI indicates the predetermined event, broadcasting a message M to the wind farm, wherein the message M contains at least some of the environmental data ED and a timestamp TS as event information EVI;

[0065] iv) upon receiving the broadcast message M, evaluating said event information EVI, said event information EVI being whether said predetermined event at said first wind turbine 1 will affect said second wind turbine 2;

[0066] v) generating a control command CO for controlling the second wind turbine 2 to counteract the predetermined event if the evaluation indicates that the predetermined event will affect the second wind turbine 2,

[0067] Wherein, in the step of evaluating the event information EVI, the system is further configured to determine a probability value of whether the event information EVI is true, wherein the control command CO is generated only when the probability value exceeds a predetermined probability threshold.

[0068] Option 13. A system according to Option 12, wherein the system is configured to perform a method according to one of Option 2 to Option 11.

[0069] Solution 14. A wind farm, comprising a first wind turbine 1 and a second wind turbine 2, wherein the wind farm comprises the system according to Solution 12 or 13.

[0070] Solution 15. A computer program product having a program code, wherein the program code is stored on a non-transitory machine-readable carrier, wherein the computer program product is configured to perform the method according to one of Solution 1 to Solution 11 when the program code is executed on a computer. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0072] Figure 1 A schematic diagram of a wind farm for implementing a first embodiment of the present invention is shown.

[0073] Figure 2 A schematic diagram of a method performed by a wind farm according to a first embodiment of the present invention is shown.

[0074] Figure 3 A schematic diagram of a wind farm for implementing a second embodiment of the present invention is shown.

[0075] Figure 4 A schematic diagram of a method performed by a wind farm according to a second embodiment of the present invention is shown. DETAILED DESCRIPTION

[0076] Figure 1 A wind farm comprising an upstream first wind turbine 1 and a downstream second wind turbine 2 is shown in a view from above. The wind farm may have more than these two wind turbines. The method described herein is applied to the first and second wind turbines 1 and 2. However, the method may also be applied to other pairs of wind turbines that are part of the wind farm, or one or more upstream first wind turbines and one or more downstream second wind turbines.

[0077] The wind turbines 1, 2 are shown in a plan view from above. The 3D coordinate system CS for indicating the spatial arrangement of the wind turbines is Figure 1 The vertical direction is indicated by the z-axis of the coordinate system CS, and the directions parallel to the horizontal direction are indicated by the x-axis and y-axis of the coordinate system CS. The wind direction is along the x-axis of the coordinate system CS.

[0078] The wind turbine 1, which is an upstream turbine with respect to the wind direction, comprises an upper section located on top of a tower (not shown) extending in the vertical z direction. The upper section comprises a nacelle 12 accommodating a generator for generating electricity. Furthermore, the upper section comprises a rotor 11R having three rotor blades 11B with an angle of 120° between the rotor blades, wherein Figure 1 Only two of these blades are shown. The wind causes the rotor 11R to rotate about a horizontal rotor axis, thereby causing a generator to generate electricity within the nacelle 12. The upper section of the wind turbine 1 is pivotable about a vertical yaw axis extending in the z-direction.

[0079] The wind turbine 1 is equipped with a plurality of sensors 13, of which only one sensor 13 is shown in the schematic diagram. The plurality of sensors 13 is constituted by sensors for acquiring environmental data ED (such as temperature, wind speed, wind direction) and stress data SD (such as mechanical load, strain, vibration of the tower, vibration of the nacelle, etc.). In addition, the wind turbine 1 comprises a calculation unit 14 and a transmission unit 15. The transmission unit 15 may be a transceiver configured to also be able to receive data.

[0080] The wind turbine 2 located downstream with respect to the wind direction has the same structure as the wind turbine 1. That is, the wind turbine 2 includes an upper section located at the top of a vertical tower (not shown), which includes a nacelle 22 accommodating a generator and a rotor 21R attached to the nacelle 22. As the wind rotates the rotor 21R around a horizontal rotor axis, the generator generates electricity in the nacelle 22. Similar to the wind turbine 1, the wind turbine 2 can pivot around a vertical yaw axis.

[0081] The wind turbine 2 is equipped with a plurality of sensors 23, of which only one sensor 23 is shown in the schematic diagram. The plurality of sensors 23 is composed of sensors for acquiring environmental data ED (such as temperature, wind speed, wind direction) and stress data SD (such as mechanical load, strain, vibration of the tower, vibration of the nacelle, etc.). In addition, the wind turbine 2 includes a calculation unit 24 and a receiving unit 25. The receiving unit 25 may be a transceiver configured to also be able to send data.

[0082] The data exchange from the transmitting unit 15 of the wind turbine 1 to the receiving unit 25 of the wind turbine 2 may be wired or wireless.

[0083] In case of an environmental event such as a gust of wind or a wave, first the upstream wind turbine 1 is affected (hit) by the wind and after a time period depending on the wind direction, wind speed and the distance between the wind turbines 1, 2, the wind turbine 2 is also affected by the environmental event. The environmental event may be such that each of the computing units 14, 24 of the first and second wind turbines 1, 2 may have to adapt their control strategies to avoid excessive loads and component failures of the wind turbines 1, 2. An immediate reaction may be an immediate hard shutdown after one or more sensors 13, 23 of the respective wind turbine 1, 2 have obtained the environmental data ED and the stress data SD.

[0084] The method described below provides an easy way to control a downstream second wind turbine 2 using acquired sensor data obtained from one or more sensors 13 of an upstream wind turbine 1 , thereby enabling the second wind turbine 2 to avoid excessive loading due to environmental events and hard control strategies.

[0085] To this end, environmental data ED and stress data SD are obtained, wherein the environmental data ED and the stress data SD are current data acquired by one or more sensors 13 installed at the first wind turbine 1. The environmental data ED and the stress data SD are fed as digital inputs to a calculation unit 14 of the first wind turbine 1. The calculation unit 14 determines state information STI based on the environmental data ED and the stress data SD, which indicates whether a predetermined event exists when the environmental data ED and the stress data SD are acquired. The term "predetermined event" corresponds to an environmental event that requires immediate control of the first wind turbine 1 to avoid extreme loads and resulting component failures, such as gusts or waves.

[0086] In the case where the status information STI indicates a predetermined event, a message M is sent from the sending unit 15 to the wind farm as a broadcast message. The message M contains event information EVI. The event information EVI is composed of or includes a timestamp TS when the environmental data ED and the stress data SD representing the predetermined event are obtained and at least some of the environmental data ED. Specifically, the wind speed WS and the wind direction are used as the event information EVI.

[0087] The broadcast message M is received by the receiving unit 25 of the second wind turbine 2. The message is forwarded from the receiving unit 25 to the computing unit 24, where it is evaluated. The evaluation consists of or includes event information processing whether the predetermined event at the first wind turbine 1 will also affect the second wind turbine 2. This information can be obtained from the wind speed, the wind direction and the known distance between the first wind turbine 1 and the second wind turbine 2, which enable the determination of a time period starting from the time when the predetermined event at the first wind turbine 1 has been determined until the predetermined event will affect the second wind turbine 2.

[0088] In case it is assessed that the predetermined event will affect the second wind turbine 2, a control command CO for controlling the second wind turbine 2 is generated so that the predetermined event can be counteracted. The control command may consist of or may include control parameters suitable for initiating a shutdown, a soft shutdown, an adjustment of the yaw angle and / or an adjustment of the pitch angle. In addition, further countermeasures or combinations thereof may be taken. If the determined time period until the predetermined event affects the second wind turbine is long enough so that a soft shutdown is possible, the control command CO may be initiated before the predetermined event reaches the second wind turbine 2. For example, a soft shutdown is much friendlier to the wind turbine and its components, thereby extending the life of the components.

[0089] If the evaluation of the event information has the result that the predetermined event will not cause extreme loads but can be used to improve performance, a control command CO can be generated such that performance is improved, for example by early adjustment of the pitch angle.

[0090] The generation of the control command CO may be hard-coded, i.e. the evaluation of the event information and / or the generation of the control command is based on processing the event information according to a predetermined mapping. Such a mapping consists of or comprises a possible reaction to each possible situation. Alternatively, the evaluation may be based on processing the event information by a trained data-driven model, wherein the event information is fed as a digital input to the trained data-driven model and the trained data-driven model provides information as a digital output whether a predetermined event at the first wind turbine will affect the second wind turbine. The trained data-driven model may be based on, for example, reinforcement learning or deep learning.

[0091] For reinforcement learning, each time a turbine reads a broadcast message, it must decide whether to take the event information into account for its control. The second wind turbine stores all its reactions or non-reactions and knows retrospectively whether the decision was good or bad. It can therefore learn from the experience of previous events and take it into account in further future operations.

[0092] Alternatively, the second wind turbine may acquire information about previous events, its reactions and results, i.e. the control commands and strategies used. This information may be fed to a machine learning algorithm, such as a neural network. The second wind turbine learns the best behavior for certain events. This learning may have to be repeated once in a while in order to update the algorithm with the latest information.

[0093] although Figure 1 and Figure 2 The first example shown in is based on direct analysis of information at the first and second wind turbines 1, 2, where the appropriate response to a predetermined event is defined by the second wind turbine 2 itself, cloud-based or centralized processes are also possible. Figure 1 and Figure 2 In the embodiment of , no network connection to the cloud is required, nor is any processing in the cloud required. Thus, real-time reaction can be guaranteed. The second wind turbine 2 acquires all information, processes it, and defines the best response for itself.

[0094] In accordance with Figure 3 and Figure 4 The second embodiment shows a cloud-based process. Figure 3 Corresponding to Figure 1 The first embodiment of the invention is similar to that of the present invention, except that there is no direct communication between the first and second wind turbines 1 , 2 . Instead, the central computing unit CCU is involved in the communication from the wind turbine 1 to the wind turbine 2 .

[0095] According to a second embodiment, a message M1 containing a time stamp TS and environmental information EVI is sent from the sending unit 15 of the first wind turbine 1 to the central computing unit CCU ( Figure 4 ), the environmental information EVI consists of or comprises the wind speed WS and the wind direction WD of the environmental data ED. The central computing unit CCU determines status information STI indicating whether a predetermined event is present when acquiring the environmental data ED and the stress data SD of the first wind turbine 1. Alternatively, the status information STI may be determined by the first wind turbine 1 and may also be part of the message M1.

[0096] Furthermore, the central computing unit CCU evaluates whether the predetermined event at the first wind turbine 1 will affect the event information EVI of the second wind turbine 2. In case the evaluation shows that the predetermined event will affect the second wind turbine 2, a control command CO is generated and sent via a message M2 to the receiving unit 25 of the second wind turbine 2. Thus, the wind turbine 2 does not have to evaluate the event information EVI itself, but instead directly receives the control command CO.

[0097] Alternatively, in case the evaluation indicates that the predetermined event will affect the second wind turbine 2, the central computing unit CU may forward the event information EVI to the second wind turbine 2. The computing unit 24 of the second wind turbine 2 will then generate a control command CO for controlling the second wind turbine to counteract the predetermined event based on the event information EVI.

Claims

1. A computer-implemented method for controlling one or more wind turbines (1, 2) in a wind farm, wherein: The wind farm comprises a first wind turbine (1) upstream and a second wind turbine (2) downstream, wherein at each of one or more time points during operation of the wind farm the following steps are performed: i) obtaining environmental data (ED) and stress data (SD) of the first wind turbine (1), the environmental data (ED) and the stress data (SD) being current data acquired by one or more sensors (13) installed at the first wind turbine (1); ii) determining status information (STI) based on said environmental data (ED) and said stress data (SD), said status information indicating whether a predetermined event existed when acquiring said environmental data (ED) and said stress data (SD), wherein said predetermined event requires immediate control of said first wind turbine (1); iii) in case the status information (STI) indicates the predetermined event, broadcasting a message (M) into the wind farm, wherein the message (M) contains at least some of the environmental data (ED) and a timestamp (TS) as event information (EVI); iv) upon receiving the broadcast message (M), evaluating said event information (EVI), said event information (EVI) being whether said predetermined event at said first wind turbine (1) will affect said second wind turbine (2); v) generating a control command (CO) for controlling the second wind turbine (2) to counteract the predetermined event if the evaluation indicates that the predetermined event will affect the second wind turbine (2), Wherein, evaluating the event information (EVI) comprises determining a probability value of whether the event information (EVI) is true, wherein generating the control command (CO) is initiated only if the probability value exceeds a predetermined probability threshold.

2. The method according to claim 1, wherein: The message comprises wind direction (WD) and wind speed (WS) as the environmental data (ED).

3. The method according to claim 1 or 2, wherein: The determination of the status information (STI) is performed by a central computing unit (CCU) or a computing unit (14) of the first wind turbine (1).

4. The method according to claim 1 or 2, wherein: Broadcasting the message (M) into the wind farm is performed by a sending unit (15) of the first wind turbine (1).

5. The method according to claim 3, wherein: Evaluating the event information (EVI) is performed by the central computing unit (CCU) or a computing unit (24) of the second wind turbine (2).

6. The method according to claim 1 or 2, wherein: Evaluating the event information (EVI) comprises determining a time period from the time at which the predetermined event has been determined at the first wind turbine (1) until the predetermined event will affect the second wind turbine (2).

7. The method according to claim 1 or 2, wherein: Evaluating the event information (EVI) and / or generating the control command (CO) is based on processing the event information (EVI) according to a predetermined mapping.

8. The method according to claim 1 or 2, wherein: Evaluating the event information (EVI) is based on processing the event information (EVI) by a trained data driven model (MO), wherein the event information (EVI) is fed as a digital input to the trained data driven model (MO), and the trained data driven model (MO) provides information as a digital output whether the predetermined event at the first wind turbine (1) will affect the second wind turbine (2).

9. The method according to claim 1 or 2, wherein: The control command (CO) includes at least one of the following countermeasures: Hard shutdown; Soft shutdown; Adjust the yaw angle; Adjust the pitch angle.

10. The method according to claim 9, wherein: The hard stop and the soft stop differ primarily based on the time taken from triggering a stop command until the received turbine (1, 2) is completely stopped.

11. The method according to claim 9, wherein: The soft stop allows for a smoother ramp down than the hard stop.

12. The method according to claim 9, wherein: Select countermeasures to minimize the impact of negative loads.

13. The method according to claim 9, wherein: Selecting a countermeasure includes considering a time period until a predetermined event at the first wind turbine affects the second wind turbine.

14. The method according to claim 9, wherein: The hard stop and the soft stop differ primarily based on the time taken from triggering a stop command until the received turbines (1, 2) are completely stopped; the soft stop allows a smoother ramp down than a hard stop; a countermeasure is selected so that the impact on negative loads is minimized; and the selection of the countermeasure includes taking into account the time period until a predetermined event at the first wind turbine affects the second wind turbine.

15. The method according to claim 1 or 2, wherein: The control command (CO) is generated by the second wind turbine (2).

16. The method according to claim 3, wherein: The control command (CO) is generated by the central computing unit (CCU), wherein the control command (CO) is sent to the second wind turbine via a broadcast message or a unicast message (M2).

17. A system for computer-implemented monitoring of one or more wind turbines (1, 2) in a wind farm, wherein: The wind farm comprises a first wind turbine (1) upstream and a second wind turbine (2) downstream, wherein the system is configured to perform the following steps at each of one or more time points during operation of the wind farm: i) obtaining environmental data (ED) and stress data (SD) of the first wind turbine (1), the environmental data (ED) and the stress data (SD) being current data acquired by one or more sensors (13) installed at the first wind turbine (1); ii) determining, based on said environmental data (ED) and said stress data (SD), state information (STI) indicating whether a predetermined event existed when acquiring said environmental data (ED) and said stress data (SD), wherein said predetermined event requires immediate control of said first wind turbine (1); iii) in case the status information (STI) indicates the predetermined event, broadcasting a message (M) into the wind farm, wherein the message (M) contains at least some of the environmental data (ED) and a timestamp (TS) as event information (EVI); iv) upon receiving the broadcast message (M), evaluating said event information (EVI), said event information (EVI) being whether said predetermined event at said first wind turbine (1) will affect said second wind turbine (2); v) generating a control command (CO) for controlling the second wind turbine (2) to counteract the predetermined event if the evaluation indicates that the predetermined event will affect the second wind turbine (2), Wherein, in step iv) of evaluating the event information (EVI), the system is further configured to determine a probability value of whether the event information (EVI) is true, wherein the generation of the control command (CO) is initiated only when the probability value exceeds a predetermined probability threshold.

18. The system of claim 17, wherein: The system is configured to perform the method according to any one of claims 2 to 16.

19. A wind farm, comprising a first wind turbine (1) and a second wind turbine (2), wherein: The wind farm comprises a system according to claim 17 or 18.

20. A non-transitory computer-readable storage medium having a program code which, when executed on a computer, performs the method according to one of claims 1 to 16.

21. A computer program product, comprising a computer program, which, when executed by a processor, implements the steps of the method according to any one of claims 1 to 16.

Citation Information

Patent Citations

  • Windpark control system

    EP1790851A2

  • Methods and systems for warning a wind turbine generator in a wind park of an extreme wind event

    EP2788620B1

  • Yaw control method and device for wind power farm

    CN107882679A

  • Detection method for judging influence of wake flow of upstream draught fan on downstream draught fan

    CN109973330A

  • Windpark turbine control system and method for wind condition estimation and performance optimization

    CN1975155A