Monitoring device, wind power generation system, and monitoring method

Abstract

To enable abnormality determination for an anemometer in a wind power generation site including two or more wind power generators mounted with an anemometer.SOLUTION: A monitoring device 100 includes: an operation information collection part 111 for acquiring operation information on a windmill including a wind direction and a wind speed from a plurality of windmills equipped with an anemoscope and an anemometer; a prediction part 113 for predicting a wind direction and a wind speed in a diagnosis target windmill on the basis of the operation information on a reference windmill in the periphery of the diagnosis target windmill, which is a windmill equipped with the anemoscope and anemometer to be diagnosed; and an abnormality determination part 114 for calculating a degree of abnormality indicating the magnitude of differences between the predicted wind direction and wind speed, and the wind direction and wind speed included in the operation information on the diagnosis target windmill on the basis of the predicted wind direction and wind speed, and the wind direction and wind speed included in the operation information on the diagnosis target windmill.SELECTED DRAWING: Figure 4

Description

[Technical field]

[0001] The present invention relates to a monitoring device for monitoring an anemometer provided in a wind power generation facility, a wind power generation system, and a monitoring method. [Background technology]

[0002] Wind power generation equipment (wind turbines), which generate electricity using wind power, a natural energy source, includes a rotor with rotor blades attached to a nacelle installed on a pole, and a generator connected to the rotor shaft.Wind power generation equipment uses the power of the wind to rotate the rotor, generating electricity with the generator.

[0003] The control device (wind turbine control device) installed on the wind turbine measures the wind direction with an anemometer, and controls the orientation of the nacelle so that the rotors always face the wind and receive the maximum wind force (yaw control). The control device also detects abnormalities in which the yaw error, which is the difference between the orientation of the nacelle and the wind direction, becomes excessive, and controls the rotors to stop safely. Furthermore, the control device measures the wind speed with an anemometer installed on the nacelle, and in order to protect the wind turbine from damage due to strong winds, etc., it releases yaw control when the wind speed is too high, and controls the turbine so that excessive load is not placed on it.

[0004] In this way, in controlling a wind turbine, the wind vane, which measures wind direction, and the anemometer, which measures wind speed, are important for the safe and stable operation of the wind turbine. Therefore, if an abnormality occurs in the wind vane or anemometer, it is necessary to detect this early and safely stop the wind turbine. Note that there are also wind vanes and anemometers that are integrated, but these are collectively referred to as anemometers.

[0005] Patent Document 1 proposes a method for detecting anomalies in an anemometer, as described in which, when a calm state in which wind speed data obtained by the anemometer is below a predetermined value continues for a predetermined period of time or more, an anemometer abnormality is determined based on the number of wind direction changes in the wind direction data within the predetermined period of time, and an abnormality warning is output.

[0006] Patent Document 2 also proposes an operational status evaluation method for evaluating the operational status of wind power equipment including at least one wind turbine, the operational status evaluation method including the steps of: acquiring operational conditions for the wind power equipment or the at least one wind turbine; calculating estimated values ​​corresponding to the operational conditions for measurable physical quantities related to the at least one wind turbine; acquiring actual measured values ​​corresponding to the physical quantities; and determining the presence or absence of an abnormality in the at least one wind turbine by comparing the estimated values ​​with the actual measured values. [Prior art documents] [Patent documents]

[0007] [Patent Document 1] Japanese Patent Application Publication No. 11-072502 [Patent Document 2] JP 2020-133616 A Summary of the Invention [Problem to be solved by the invention]

[0008] However, in the method for determining an anomaly in an anemometer / wind direction meter described in Patent Document 1, an anomaly is determined based on the presence or absence of a windless state, such as a continuation of a constant numerical value for the wind speed. For this reason, it is not possible to determine an anomaly in which the wind speed value fluctuates and the wind speed includes an offset. Furthermore, in determining an anomaly in wind direction based on the number of wind direction changes, it is not possible to determine an anomaly in which the wind direction includes an offset and the number of changes is the same. Furthermore, the operational status evaluation method described in Patent Document 2 is related to diagnosing the operational status of the entire wind turbine, and does not take into consideration diagnosing an anemometer anomaly.

[0009] For the safe and stable operation of wind turbines, it is necessary to measure wind direction and speed correctly, and it is necessary to detect anomalies in the anemometer early. There are no restrictions on wind direction and speed conditions, and it is necessary to accurately determine anomalies in the anemometer. The present invention has been made in consideration of the above background, and has an objective of providing a monitoring device, a wind power generation system, and a monitoring method that enable determination of abnormalities in an anemometer at a wind power generation site that includes two or more wind power generation devices equipped with an anemometer. [Means for solving the problem]

[0010] In order to solve the above-mentioned problems, the monitoring device of the present invention comprises an operation information collection unit that acquires operation information of a plurality of wind turbines equipped with anemometers and anemometers, including wind direction and wind speed, a prediction unit that predicts the wind direction and wind speed at a target wind turbine, which is a wind turbine equipped with the anemometers and anemometers and is the subject of diagnosis, based on operation information of a reference wind turbine located in the vicinity of the target wind turbine, and an abnormality determination unit that calculates a degree of abnormality that indicates the magnitude of the difference between the predicted wind direction and wind speed and the wind direction and wind speed included in the operation information of the target wind turbine, based on the predicted wind direction and wind speed and the wind direction and wind speed included in the operation information of the target wind turbine. Effect of the Invention

[0011] According to the present invention, it is possible to provide a monitoring device, a wind power generation system, and a monitoring method that enable abnormality determination of an anemometer mounted on a wind power generation site including two or more wind power generation devices. Problems, configurations, and effects other than those described above will become clear from the description of the following embodiments. [Brief description of the drawings]

[0012] [Figure 1] FIG. 1 is a diagram showing a wind power generation site where a wind turbine according to an embodiment of the present invention is installed. [Diagram 2] FIG. 1 is a diagram showing a configuration of a wind turbine according to an embodiment of the present invention. [Diagram 3] FIG. 2 is a functional block diagram of the wind turbine control device according to the present embodiment. [Figure 4] 1 is a functional block diagram of a monitoring device according to an embodiment of the present invention. [Diagram 5]4 is a diagram for explaining an abnormality determination (diagnosis) method for the anemometer according to the present embodiment. FIG. [Figure 6] 11 is a flowchart of a prediction model generation process according to the present embodiment. [Figure 7] 4 is a flowchart of an abnormality determination process (diagnosis process) according to the present embodiment. [Figure 8] FIG. 4 is a diagram showing the configuration of a diagnosis screen according to the embodiment. [Figure 9] FIG. 13 is a diagram showing the configuration of a monitoring screen according to a modified example of the present embodiment. [Figure 10] 10 is a flowchart of an abnormality determination process (diagnosis process) according to a modified example of the present embodiment. [Figure 11] FIG. 11 is a functional block diagram of a monitoring device according to a modified example of the embodiment. [Figure 12] FIG. 11 is a sequence diagram of a control process for a wind turbine according to a modified example of the embodiment. [Figure 13] FIG. 11 is a functional block diagram of a monitoring device according to a modified example of the embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0013] <Wind power generation site configuration> Before describing a monitoring device as a mode (embodiment) for carrying out the present invention, a wind turbine (wind power generation facility) equipped with an anemometer to be monitored will be described. Fig. 1 is a diagram showing a wind power generation site 300 on which a wind turbine 380 according to this embodiment is installed. Sixteen wind turbines 380 (see Fig. 2 described later) are installed on the wind power generation site 300. Note that in Fig. 1, only one wind turbine 380 at the bottom right is marked with a reference numeral.

[0014] <Windmill configuration> 2 is a diagram showing the configuration of a wind turbine 380 according to this embodiment. The wind turbine 380 includes rotor blades 381, a hub 382, ​​an anemometer 383, a nacelle 384, and a support 385. The anemometer 383 is installed on the nacelle 384.

[0015] 3 is a functional block diagram of a wind turbine control device 400 according to this embodiment. The wind turbine control device 400 is assumed to be located inside a wind turbine 380. The wind turbine control device 400 is a computer or microcomputer, and includes a control unit 410 and a communication unit 480. The control unit 410 includes a microprocessor. The communication unit 480 includes a communication device, and may be integrated with the control unit 410. The control unit 410 includes a wind turbine information acquisition unit 411, a wind turbine control unit 412, and a wind turbine information transmission unit 413.

[0016] The wind turbine information acquisition unit 411 acquires operation information of the wind turbine 380. The operation information includes the wind direction and wind speed measured by the anemometer 383, as well as the rotor rotation speed (rotation speed of the rotor 381), the pitch angle of the rotor 381, the amount of power generation, etc. The wind turbine information acquisition unit 411 acquires the operation information periodically (at a predetermined interval). The wind turbine control unit 412 controls the entire wind turbine 380 based on the information acquired by the wind turbine information acquisition unit 411. For example, the wind turbine control unit 412 orients the nacelle 384 upwind based on the wind direction. Furthermore, the wind turbine control unit 412 adjusts the pitch angle of the rotor 381 to be parallel to the wind direction when the wind speed exceeds a predetermined value. The wind turbine information transmission unit 413 periodically transmits the operation information via the communication unit 480 to a monitoring device 100 (see FIG. 4, which will be described later).

[0017] <Overview of the monitoring device> Monitoring device 100 for an anemometer 383 predicts the wind direction and wind speed at the wind turbine 310 (see FIG. 1) that is the subject of monitoring, based on operation information of wind turbines 320 located around (for example, adjacent to) the wind turbine 310. Note that the operation information is not limited to wind direction and wind speed, but may also include rotor rotation speed, pitch angle of rotor blades 381, power generation amount, etc. Monitoring device 100 compares the predicted wind direction and wind speed with the wind direction and wind speed measured by the anemometer 383 of the wind turbine 310, to determine whether or not there is an abnormality in the anemometer 383. Such monitoring device 100 is able to monitor the anemometer 383.

[0018] <Configuration of monitoring device> 4 is a functional block diagram of the monitoring device 100 according to this embodiment. The monitoring device 100 is a computer, and includes a control unit 110, a storage unit 120, and an input / output unit 180. User interface devices such as a display, keyboard, and mouse are connected to the input / output unit 180. The input / output unit 180 includes a communication device, and is capable of transmitting and receiving data to and from the wind turbine control device 400.

[0019] ≪Monitoring device: storage unit≫ The storage unit 120 includes storage devices such as a read only memory (ROM), a random access memory (RAM), a solid state drive (SSD), etc. The storage unit 120 stores an operation information database 130, a prediction model database 140, and a program 128.

[0020] The program 128 includes descriptions of a prediction model generation process (see FIG. 6) and an abnormality determination process (diagnosis process, see FIG. 7), which will be described later. The operation information database 130 stores operation information for each wind turbine 380. The operation information is time-series data, and is stored in association with the wind turbine 380 and the time.

[0021] Prediction model database 140 stores a prediction model, which is a machine learning model. This prediction model is used to predict the wind direction or wind speed at wind turbine 310 (see FIG. 1) equipped with an anemometer 383 that is the subject of anomaly determination (diagnosis), based on operation information of wind turbine 320 located in the vicinity of wind turbine 310. Wind turbine 320 is also referred to as a reference wind turbine in comparison with wind turbine 310. The reference wind turbine is, for example, wind turbine 320 adjacent to wind turbine 310. Note that there are eight reference wind turbines in FIG. 1. Generally, the reference wind turbine is one or more adjacent wind turbines 380, and in order to improve the prediction accuracy of the prediction model, it is desirable to adjust the number of reference wind turbines depending on the prediction accuracy.

[0022] The explanatory variables of the prediction model used when predicting wind direction desirably include wind speed in addition to the wind direction included in the operation information of the reference wind turbine, and desirably further include other operation information such as rotor rotation speed, etc. The same applies to the explanatory variables of the prediction model used when predicting wind speed, and may include operation information such as wind direction, rotor rotation speed, etc. in addition to the wind speed of the reference wind turbine.

[0023] Note that although anemometer 383 is the subject of abnormality determination (diagnosis), anemometer 383 and windmill 380 equipped with anemometer 383 may be regarded as the same, and windmill 380 may be described as the subject of abnormality determination. For example, "windmill 380 equipped with anemometer 383 to be diagnosed" may be simply described as "windmill 380 to be diagnosed." A prediction model is provided for each wind turbine 380 that is the subject of anomaly judgment (diagnosis subject), and is stored in the prediction model database 140 in association with the objective variable (wind direction or wind speed) and the reference wind turbine.

[0024] <Monitoring device: control unit> The control unit 110 includes a CPU (Central Processing Unit) and is equipped with an operation information collection unit 111, a prediction model generation unit 112, a prediction unit 113, and an abnormality determination unit 114. The control unit 110 may include a GPU (Graphics Processing Unit), an FPGA (Field Programmable Gate Array), an ASIC (Application Specific Integrated Circuit), etc.

[0025] The operation information collector 111 collects operation information from the wind turbine control devices 400 provided in each of the wind turbines 380 and stores the information in the operation information database 130 . The prediction model generation unit 112 generates a prediction model and stores it in the prediction model database 140. The prediction model generation unit 112 generates a prediction model for wind direction and wind speed for each wind turbine 380 that is to be subjected to anomaly judgment. The prediction unit 113 predicts the wind direction or wind speed at the wind turbine 310 using a prediction model based on the operation information of the reference wind turbine (surrounding wind turbine 320).

[0026] The abnormality determination unit 114 calculates the degree of abnormality by comparing the predicted wind direction or wind speed for a predetermined period of time with the wind direction or wind speed included in the operation information, which is the actual measured value at the wind turbine 310 for that period. The abnormality determination unit 114 also outputs a determination result, which is the result of the abnormality determination (diagnosis) based on the degree of abnormality.

[0027] The degree of anomaly is an index showing the difference between the predicted value and the actual measured value, and is calculated based on, for example, the root mean squared error (RMSE), the mean squared error (MSE), or the coefficient of determination (R2). In what follows, the degree of anomaly is calculated by subtracting the coefficient of determination from 1 (degree of anomaly = 1 - coefficient of determination). The coefficient of determination is a value between 0 and 1, and the closer it is to 1, the closer the predicted value is to the actual measured value. The higher the degree of anomaly, the greater the difference between the predicted value and the actual measured value, and it is estimated that there is a high possibility that there is an abnormality in the anemometer 383 of the wind turbine 310.

[0028] If the degree of abnormality is greater than a predetermined value (a predetermined value serving as a threshold value for determining whether or not there is an abnormality), the abnormality determination unit 114 determines that there is an abnormality in the anemometer 383. The abnormality determination unit 114 may output a three-level determination such as normal / caution / abnormal using multiple threshold values, rather than normal / abnormal. Note that the predetermined value is appropriately set by simulation or experiment, but if the predetermined value is set to a small value, the frequency of determining that there is an abnormality increases.

[0029] 5 is a diagram for explaining an anomaly determination (diagnosis) method for the anemometer 383 according to this embodiment. The prediction unit 113 calculates a predicted value of the wind direction or wind speed at the wind turbine 310 based on operation information of a reference wind turbine (see wind turbine 320 in FIG. 1) located in the vicinity of the wind turbine 310 to be diagnosed. The operation information of the reference wind turbine that is input to the prediction unit 113 is not limited to the wind direction or wind speed, and desirably includes the rotor rotation speed, the pitch angle of the rotor blades 381, and the amount of power generation. The abnormality determination unit 114 calculates the degree of abnormality based on the predicted value and the wind direction or wind speed of the wind turbine 310, which is the wind turbine to be diagnosed, and determines whether or not there is an abnormality.

[0030] As described above, the monitoring device 100 includes an operation information collector 111 that acquires operation information of the wind turbines 380, including wind direction and wind speed, from a plurality of wind turbines 380 equipped with an anemometer and an anemometer (see anemometer 383). The monitoring device 100 also includes a prediction unit 113 that predicts the wind direction and wind speed at the diagnosis target wind turbine (see wind turbine 310), which is a wind turbine equipped with a wind vane and anemometer to be diagnosed, based on operation information of a reference wind turbine (see wind turbine 320) located in the vicinity of the diagnosis target wind turbine. The prediction unit 113 predicts the wind direction and wind speed in the diagnosis target wind turbine using a machine learning model (prediction model) that uses the operation information of the reference wind turbine as an explanatory variable and the wind direction or wind speed in the diagnosis target wind turbine as a response variable. The monitoring device 100 is equipped with an abnormality determination unit 114 that calculates a degree of abnormality indicating the magnitude of the difference between the predicted wind direction and wind speed and the wind direction and wind speed included in the operation information of the wind turbine to be diagnosed, based on the predicted wind direction and wind speed and the wind direction and wind speed included in the operation information of the wind turbine to be diagnosed.

[0031] <Prediction model generation process> 6 is a flowchart of a prediction model generation process according to this embodiment. The process of generating a prediction model will be described with reference to FIG. In step S11, the prediction model generation unit 112 starts the process of repeating steps S12 to S14 for each wind turbine 380. Hereinafter, the wind turbine 380 that is the target of this repeated process is referred to as the processing target wind turbine.

[0032] In step S12, the prediction model generation unit 112 identifies a reference wind turbine for the wind turbine to be treated. The reference wind turbine is a wind turbine adjacent to the wind turbine to be treated (see wind turbine 320 for wind turbine 310 in FIG. 1). Note that if the wind turbine to be treated is on the edge or in a corner of the wind power generation site 300, the number of reference wind turbines will be less than eight.

[0033] In step S13, the prediction model generation unit 112 generates learning data for generating (training) a prediction model. The explanatory variables of the learning data are operation information of the reference wind turbine over a predetermined period of time. The objective variable is the wind direction or wind speed included in the operation information of the processing target wind turbine over that period. Note that during that period, the processing target wind turbine and the reference wind turbine, including the anemometer 383, are assumed to be normal. If there is an abnormal wind turbine among the reference wind turbines, that wind turbine is excluded from the reference wind turbines, and the operation information of that wind turbine is not included in the learning data.

[0034] In step S14, the prediction model generation unit 112 generates a prediction model using the learning data generated in step S13, and stores it in the prediction model database 140 in association with the objective variable (wind direction or wind speed), the target wind turbine, and the reference wind turbine (identification information). There may be a plurality of reference wind turbines (combinations of wind turbines 380) for one processing target wind turbine. In this case, learning data is generated for each reference wind turbine, and a prediction model is generated.

[0035] <Abnormality determination process (diagnosis process)> Fig. 7 is a flowchart of anomaly determination processing (diagnosis processing) according to this embodiment. Processing for diagnosing anemometer 383 provided in designated wind turbine 380 will be described with reference to Fig. 7. It is assumed that wind turbine 380 has already been designated as the diagnosis target at the start of the anomaly determination processing. Hereinafter, wind turbine 380 to be diagnosed will be referred to as the diagnosis target wind turbine.

[0036] In step S21, the abnormality determination unit 114 acquires the wind direction and wind speed for a prescribed length of time that is included in the operation information of the diagnosis target wind turbine. The wind direction and wind speed are actual measurements taken by the anemometer 383 provided in the diagnosis target wind turbine. In step S22, the abnormality determination unit 114 acquires operation data of the reference wind turbine for the diagnosis target wind turbine for the same period as in step S21. The reference wind turbine for the diagnosis target wind turbine can be acquired by referring to the prediction model database 140.

[0037] In step S23, prediction section 113 predicts the wind direction and wind speed using a prediction model corresponding to the diagnosis target wind turbine based on the operation information of the reference wind turbine. In step S24, the abnormality determination unit 114 calculates the degree of abnormality for each of the wind direction and wind speed based on the predicted values ​​of the wind direction and wind speed obtained in step S23 and the actual measured values ​​of the wind direction and wind speed acquired in step S21. Next, the abnormality determination unit 114 compares the calculated values ​​with a predetermined value (a predetermined value serving as a threshold value for determining whether or not there is an abnormality) to determine whether or not there is an abnormality (abnormal / normal).

[0038] <Diagnosis screen> Fig. 8 is a diagram showing the configuration of diagnosis screen 510 according to this embodiment. Buttons corresponding to wind turbines 380 (see Fig. 1) are arranged in display area 511 for specifying the wind turbine 380 to be diagnosed. In Fig. 8, button 512 is pressed and displays "Diagnosis" on button 512, indicating that the wind turbine 380 corresponding to button 512 is the one to be diagnosed. Furthermore, the button corresponding to the reference wind turbine displays "Reference". This allows the user of monitoring device 100 to easily grasp the positional relationship between wind turbine 380 to be diagnosed and wind turbine 380 serving as the reference wind turbine.

[0039] The wind direction is also displayed in display area 515. The user of monitoring device 100 can refer to the wind direction and specify wind turbine 380 that will be the reference wind turbine for wind turbine 380 that is the diagnosis target. For example, when wind turbine 380 upwind of wind turbine 380 that is the diagnosis target is set as the reference wind turbine, the user can refer to this display of wind direction. When a “Judge” button 513 arranged at the bottom of the diagnosis screen 510 is pressed, an abnormality judgment process (see FIG. 7) is executed, and the judgment result and the degree of abnormality are displayed in an area 514 .

[0040] <Features of the monitoring device> Monitoring device 100 predicts the wind direction and wind speed at wind turbine 310 (see FIG. 1) to be diagnosed, based on operation information of reference wind turbine, which is wind turbine 320 located in the vicinity of wind turbine 310. Monitoring device 100 compares the predicted wind direction and wind speed with the wind direction and wind speed measured by an anemometer 383 of wind turbine 310 to determine the presence or absence of an abnormality in anemometer 383. Monitoring device 100 can diagnose (determine an abnormality) the anemometer 383 to be diagnosed based on the wind direction and wind speed in the vicinity, enabling highly accurate and reliable diagnosis.

[0041] <<Variation: Constant Surveillance>> In the above embodiment, the monitoring device 100 diagnoses the specified wind turbine 380 (see Figs. 7 and 8). All the wind turbines 380 at the wind power generation site 300 may be repeatedly diagnosed in order for constant monitoring. Note that repeated diagnosis may also be performed on all wind turbines 380 except for those that are stopped for repair or inspection.

[0042] 9 is a diagram showing the configuration of a monitoring screen 520 according to a modified example of this embodiment. In a display area 521, the diagnosis results of the anemometer 383 provided on each wind turbine 380 are displayed. In this modified example, the diagnosis results (abnormality determination results) are in three stages: normal, caution, and abnormal. If the abnormality level is less than 0.6, it is normal. If the abnormality level is 0.6 or more and less than 0.8, it is caution. If the abnormality level is 0.8 or more, it is abnormal. By repeatedly performing the diagnosis, the user of the monitoring device 100 can know the state of the anemometer 383 at all times.

[0043] <<Modification: Standard Windmill>> Considering the possibility that the reference wind turbine that serves as the input for the prediction model at the time of diagnosis is stopped and operation information cannot be obtained, the reference wind turbine for the wind turbine 380 to be diagnosed may be some of the surrounding wind turbines 380 rather than all of the surrounding wind turbines 380. In this way, prediction models for wind direction and wind speed are generated for the number of selected reference wind turbines (combinations of reference wind turbines) and stored in association with the reference wind turbines in the prediction model database 140. In this way, prediction models corresponding to a plurality of reference wind turbines may be generated in advance, or may be generated dynamically. An embodiment in which prediction models are generated dynamically will be described below.

[0044] 10 is a flowchart of an abnormality determination process (diagnosis process) according to a modified example of this embodiment. An abnormality determination process that replaces the abnormality determination process shown in FIG. Step S31 is the same process as step S21 (see FIG. 7). In step S32, the abnormality determination unit 114 selects a wind turbine 380 that is operating normally from among the wind turbines 380 around the wind turbine to be diagnosed as a reference wind turbine, and acquires operation data of the reference wind turbine. A wind turbine 380 that is not operating normally (is abnormal) is, for example, a wind turbine 380 that has been stopped for repair or inspection. Furthermore, a wind turbine 380 whose wind direction or wind speed is significantly different (different by more than a predetermined value) from adjacent wind turbines 380 may be determined to be not operating normally and not included in the reference wind turbine. Note that this predetermined value is a threshold value related to wind direction and wind speed for determining whether or not the wind turbine is operating normally.

[0045] In step S33, if there is a prediction model corresponding to the reference wind turbine selected in step S32 (step S33→YES), the abnormality determination unit 114 proceeds to step S36, and if there is not (step S33→NO), the abnormality determination unit 114 proceeds to step S34. Steps S34 and S35 are similar processes to steps S13 and S14 (see FIG. 6). Note that the operation information of the reference wind turbine that serves as the explanatory variable of the learning data is the operation information of the reference wind turbine selected in step S32. Steps S36 and S37 are similar processes to steps S23 and S24 (see FIG. 7).

[0046] 10, even if there is no prediction model corresponding to a normally operating reference wind turbine, a prediction model corresponding to a new normally operating reference wind turbine can be generated and anemometer 383 can be diagnosed. Even if there is an abnormality in the reference wind turbine of an existing prediction model in prediction model database 140, a prediction model can be generated and diagnosed in which wind turbines 380 operating normally, excluding the abnormal wind turbine 380, are used as the reference wind turbine. An abnormal wind turbine 380 is, for example, a wind turbine 380 that is stopped for repair or inspection. In addition, a wind turbine 380 whose wind direction or speed is significantly different from that of adjacent wind turbines 380 (different by a predetermined value or more) may also be considered to be abnormal.

[0047] As described above, the monitoring device 100 is equipped with a prediction model generation unit 112 that generates a machine learning model in which, when the reference wind turbine related to the operation information included in the explanatory variables of the machine learning model (prediction model) is abnormal, the explanatory variables are operation information excluding the operation information of the reference wind turbine (see steps S34 and S35). The prediction unit 113 predicts the wind direction and wind speed in the diagnosis target wind turbine using the generated machine learning model.

[0048] <<Variation: Two Anemometers>> In the above embodiment, the wind turbine 380 is equipped with one anemometer 383. There are also wind turbines 380 equipped with two anemometers 383. When diagnosing one of the anemometers 383 of such a wind turbine 380, the wind direction and wind speed of the other anemometer 383 may be added to the operation information of the reference wind turbine. In this way, it is expected that the accuracy of the predicted value calculated by the prediction unit 113 will improve, and anomaly determination will be made more accurately. Furthermore, when the reference wind turbine is equipped with two anemometers 383, the objective variable of the prediction model may be the wind direction or wind speed of both anemometers 383, or the average value of the measurements of the two anemometers 383, or just one of them.

[0049] <<Modification: Prediction model for wind speed prediction>> A prediction model may be prepared for each wind direction to be used in predicting wind speed. For example, the wind direction at the reference wind turbine, which is an explanatory variable, is divided into multiple ranges, and a prediction model is generated for each range. The ranges may overlap. At the time of diagnosis, the wind speed is predicted using a prediction model corresponding to the wind direction at the reference wind turbine. For example, four prediction models may be generated using the respective learning data for four wind direction ranges that are ±45 degrees in each of the east-west-north-south directions.

[0050] By predicting wind speed using a prediction model for each wind direction in this way, the effects of wake turbulence and terrain can be avoided, improving prediction accuracy and ultimately improving diagnostic accuracy. In addition, a prediction model used for predicting the wind direction may be prepared for each wind direction.

[0051] <<Variation: Learning Data>> Although no conditions are set for the learning data used to generate the prediction model in the above embodiment, conditions may be set for the wind direction and wind speed. For example, the learning data related to the prediction model used to predict the wind speed may be limited to data for a period in which the wind speed, which is the objective variable, is the wind speed between cut-in and cut-out (operation information of the wind turbine to be predicted and the reference wind turbine). In addition, when the wind direction at the wind power generation site 300 is concentrated in a specific direction, the learning data related to the prediction model used to predict the wind direction may be limited to data for a period in which the wind direction, which is the objective variable, is included in that direction. For example, when there are many northwesterly winds, the prediction model may be generated using learning data limited to a direction of ±60 degrees northwest. By generating a prediction model using learning data with such restrictions, it is expected that the accuracy of prediction will improve, and in turn, the accuracy of diagnosis will also improve.

[0052] As described above, the machine learning model (prediction model) is configured to include a plurality of prediction models for each range of wind direction included in the operation information that serves as an explanatory variable. The machine learning model is a model that has been generated using training data in which the wind speed, which is the objective variable, is between the cut-in and cut-out of the wind turbine to be diagnosed. The machine learning model is a model that has been generated using training data in which the wind direction in the wind turbine to be diagnosed is within a predetermined range.

[0053] <<Modification: Wind turbine operation based on predicted wind direction and speed when an abnormality is detected>> If an anemometer 383 is diagnosed as abnormal, it is desirable to stop operation of the wind turbine 380 equipped with that anemometer 383. This is because the wind direction and wind speed measured by the anemometer 383 differ from the actual wind direction and wind speed, causing the orientation of the nacelle 384 and the pitch angle of the rotor 381 to become inappropriate, increasing the risk of the wind turbine 380 breaking down or being damaged.

[0054] When the anemometer 383 is diagnosed as abnormal, the wind turbine control device 400 (see FIG. 3) may control the wind turbine 380 based on predicted wind direction and wind speed calculated based on operation information of the reference wind turbine, instead of the measured values ​​of the anemometer 383. Below, a wind power generation system 10 (see FIG. 12 described later) configured including a monitoring device 100A and a wind turbine control device 400A for realizing such an operation will be described.

[0055] 11 is a functional block diagram of a monitoring device 100A according to a modification of this embodiment. Compared to the monitoring device 100 (see FIG. 4), the control unit 110 includes a wind direction and speed transmission unit 115, and a prediction unit 113A is different. The prediction unit 113A repeatedly calculates predicted values ​​of wind direction and wind speed for a wind turbine 380 that has been diagnosed as abnormal, based on the operation information of the reference wind turbine for that wind turbine 380. The wind direction and speed transmitting unit 115 repeatedly transmits this predicted value to the wind turbine 380.

[0056] A wind turbine control unit 412 (see FIG. 3) of the wind turbine control device 400A (not shown) controls the wind turbine 380 based on the predicted wind direction and wind speed received from the monitoring device 100A. For example, the wind turbine control unit 412 controls the direction of the nacelle 384 and the pitch angle of the rotor blades 381.

[0057] Fig. 12 is a sequence diagram of a control process for the wind turbine 380 according to a modified example of this embodiment. Fig. 12 shows the process that is repeatedly executed after the wind turbine 380 is diagnosed as having an abnormality. Steps S51 and S52 are similar processes to steps S22 and S23 (see FIG. 7). In step S53, the wind direction and wind speed transmitting unit 115 transmits to the wind turbine 380 the wind direction and wind speed predicted in step S52. In step S54, the wind turbine control unit 412 of the wind turbine control device 400A provided in the wind turbine 380 controls the wind turbine 380 based on the wind direction and wind speed received in step S53.

[0058] According to such monitoring device 100A, even if an abnormality occurs in the anemometer 383 provided in the monitoring device 100A itself, the wind turbine 380 can continue to operate based on the predicted wind direction and wind speed.

[0059] As described above, the wind power generation system 10 includes the wind turbine control device 400A provided in the wind turbine 380 and the monitoring device 100A. The wind turbine control device 400 includes a wind turbine control unit 412 . The monitoring device 100A includes an operation information collector 111 that acquires operation information of the wind turbines 380, including wind direction and wind speed, from a plurality of wind turbines 380 equipped with a wind vane and an anemometer (see an anemometer 383).

[0060] The monitoring device 100A comprises a prediction unit 113A that repeatedly predicts the wind direction and wind speed in the diagnosis target wind turbine, which is a wind turbine 380 equipped with a wind vane and anemometer to be diagnosed, based on operation information of a reference wind turbine located in the vicinity of the diagnosis target wind turbine. The monitoring device 100A includes an abnormality determination unit 114 that calculates an abnormality level indicating the magnitude of the difference based on the difference between the predicted wind direction and wind speed and the wind direction and wind speed included in the operation information of the diagnosis target wind turbine. The monitoring device 100A is equipped with a wind direction and wind speed transmission unit 115 that repeatedly transmits predicted wind direction and wind speed to the wind turbine to be diagnosed when the degree of abnormality of the wind turbine to be diagnosed exceeds a predetermined value (a predetermined value serving as a threshold for determining whether or not there is an abnormality). A wind turbine control unit 412 provided in the diagnosis target wind turbine controls the wind turbine 380 based on the received wind direction and wind speed.

[0061] <<Variation: Operation of the wind turbine when an abnormality is detected>> It is assumed that the wind turbine 380 is equipped with multiple anemometers 383. In this case, if one anemometer 383 is determined to be abnormal, the wind turbine control unit 412 (see FIG. 3) of the wind turbine 380 may be configured to control the wind turbine 380 based on the wind direction and wind speed measured by the normal anemometer 383. In this way, even if an abnormality occurs in the anemometer 383, the operation of the wind turbine 380 can be continued based on the measurement values ​​of the other anemometers 383.

[0062] Fig. 13 is a functional block diagram of a monitoring device 100B according to a modified example of this embodiment. Compared to monitoring device 100 (see Fig. 4), control unit 110 is provided with a control change command transmission unit 116. When abnormality determination unit 114 diagnoses an abnormality in anemometer 383 as exceeding a predetermined value, control change command transmission unit 116 transmits identification information of anemometer 383 to wind turbine 380, and commands (instructs) the wind turbine 380 to change the control of the wind turbine 380.

[0063] A wind turbine control unit 412 (see FIG. 3) of the wind turbine control device 400B (not shown) controls the wind turbine 380 based on the wind direction and wind speed measured by an anemometer 383 different from the anemometer 383 identified by the identification information received from the monitoring device 100B. For example, the wind turbine control unit 412 controls the direction of the nacelle 384 and the pitch angle of the rotor 381.

[0064] Furthermore, if the degree of abnormality falls below a predetermined value after the anemometer 383 is diagnosed as abnormal, the control change command transmitter 116 transmits the identification information of the anemometer 383 to the wind turbine 380, and transmits a command to restore the control change of the wind turbine 380. The wind turbine control unit 412 that receives this command controls the wind turbine 380 based on the wind direction and wind speed measured by the anemometer 383 identified by the identification information received from the monitoring device 100B and the other anemometers 383.

[0065] The above-described modified example is an operation method when the wind turbine 380 is equipped with multiple anemometers 383. The case where the wind turbine 380 is equipped with one anemometer 383 will be described below. When the anemometer 383 is diagnosed as abnormal and a command to change control is received, the wind turbine control unit 412 controls the wind turbine 380 to stop operation (power generation). Furthermore, when a command to return the control change to its original state is received, the wind turbine control unit 412 controls the wind turbine 380 to return to normal operation. In this way, it is possible to avoid breakdowns or damage to the wind turbine 380 caused by an abnormality in the anemometer 383. Furthermore, when the abnormality is recovered from, the wind turbine 380 will resume operation.

[0066] As described above, the monitoring device 100B comprises a control change command transmitting unit 116 that transmits a command to change the control of the diagnosis target wind turbine when the degree of abnormality of the diagnosis target wind turbine exceeds a predetermined value. When the degree of abnormality of the wind turbine to be diagnosed exceeds a predetermined value and then falls below the predetermined value, the control change command transmitting unit 116 transmits a command to the wind turbine to be diagnosed to revert the control change of the wind turbine to be diagnosed. The wind turbine control unit 412 changes the control of the wind turbine 380 when a command to change the control is received.

[0067] It is assumed that the wind turbine is equipped with multiple anemometers and anemometers (wind vane and anemometer 383). When the degree of abnormality of the anemometers and anemometers of the wind turbine to be diagnosed exceeds a predetermined value, the control change command transmitting unit 116 transmits a command to change control, including identification information of the anemometers and anemometers, to the wind turbine to be diagnosed. When the wind turbine control unit 412 receives the command to change control, it controls the wind turbine 380 based on the wind direction and wind speed measured by anemometers and anemometers other than the anemometer and anemometer identified by the identification information.

[0068] <Modification: Prediction model generation unit> The monitoring device 100 generates a prediction model and predicts wind direction and wind speed using the prediction model, but the generating device and the predicting device may be different devices. For example, the monitoring device 100 may not include the prediction model generating unit 112 and may execute the anomaly determination process (see FIG. 7) using a prediction model generated by another device.

[0069] Other variations Although several embodiments of the present invention have been described above, these embodiments are merely illustrative and do not limit the technical scope of the present invention. For example, the prediction model is a machine learning model for predicting wind direction or wind speed, but it may be a machine learning model for predicting both. Furthermore, while the reference wind turbine in the above-described embodiments is a wind turbine in the vicinity of the wind turbine to be diagnosed (an adjacent wind turbine), it may be a wind turbine other than the wind turbine to be diagnosed at the wind power generation site 300.

[0070] The present invention can take various other embodiments, and furthermore, various modifications such as omissions and substitutions can be made without departing from the gist of the present invention. These embodiments and modifications are included in the scope and gist of the invention described in this specification, etc., and are included in the scope of the invention described in the claims and their equivalents. [Explanation of symbols]

[0071] 10. Wind power generation system 100,100A,100B monitoring device 111 Operational Information Collection Department 112 Prediction model generation unit 113,113A Prediction Section 114 Abnormality determination section 115 Wind direction and speed transmitter 116 Control change command transmission unit 128 Programs 130 Operational Information Database 140 Prediction Model Database 380 Windmill 383 Wind direction anemometer 400,400A Wind Turbine Control Device 411 Wind turbine information acquisition section 412 Wind Turbine Control Unit 413 Wind Turbine Information Transmitter

Claims

1. an operation information collection unit that acquires operation information of a plurality of wind turbines, including wind direction and wind speed, from the plurality of wind turbines each having a wind vane and an anemometer; a prediction unit that predicts the wind direction and wind speed in the diagnosis target wind turbine, which is a wind turbine that is the subject of diagnosis and is equipped with the anemometer and anemometer, based on operation information of a reference wind turbine that is located in the vicinity of the diagnosis target wind turbine; an anomaly determination unit that calculates, based on the predicted wind direction and wind speed and the wind direction and wind speed included in the operation information of the diagnosis target wind turbine, a degree of anomaly that indicates the magnitude of the difference between the predicted wind direction and wind speed and the wind direction and wind speed included in the operation information of the diagnosis target wind turbine. monitoring equipment.

2. The prediction unit is The wind direction and wind speed in the diagnosis target wind turbine are predicted using a machine learning model that uses the operation information of the reference wind turbine as an explanatory variable and the wind direction or wind speed in the diagnosis target wind turbine as an objective variable. The monitoring device according to claim 1.

3. a prediction model generation unit that generates a machine learning model in which, when a reference wind turbine related to operation information included in an explanatory variable of the machine learning model is abnormal, operation information excluding the operation information of the reference wind turbine is used as an explanatory variable; The prediction unit is The generated machine learning model is used to predict the wind direction and wind speed in the wind turbine to be diagnosed. The monitoring device according to claim 2.

4. The machine learning model is The predictive model includes a plurality of predictive models for each range of wind direction included in the operation information that is the explanatory variable. The monitoring device according to claim 2.

5. The machine learning model is A machine learning model generated using learning data in which the wind speed, which is the objective variable, is between the cut-in and cut-out of the diagnosis target wind turbine. The monitoring device according to claim 2.

6. The machine learning model is A machine learning model generated using learning data in which the wind direction in the diagnosis target wind turbine is within a predetermined range. The monitoring device according to claim 2.

7. and a wind direction and wind speed transmitting unit that repeatedly transmits predicted wind direction and wind speed to the wind turbine to be diagnosed when the degree of abnormality of the wind turbine to be diagnosed exceeds a predetermined value. The monitoring device according to claim 1.

8. and a control change command transmission unit that transmits to the diagnosis target wind turbine a command to change control of the diagnosis target wind turbine when the degree of abnormality of the diagnosis target wind turbine exceeds a predetermined value. The monitoring device according to claim 1.

9. The control change command transmission unit is When the degree of abnormality of the diagnosis target wind turbine exceeds a predetermined value and then falls below the predetermined value, a command to restore the control change of the diagnosis target wind turbine is transmitted to the diagnosis target wind turbine.

9. The monitoring device according to claim 8.

10. The wind turbine includes a wind turbine control device and a monitoring device, The wind turbine control device includes a wind turbine control unit, The monitoring device includes: an operation information collection unit that acquires operation information of a plurality of wind turbines, including wind direction and wind speed, from the plurality of wind turbines each having a wind vane and an anemometer; a prediction unit that repeatedly predicts the wind direction and wind speed at the diagnosis target wind turbine, which is a wind turbine that is the subject of diagnosis and is equipped with the anemometer and anemometer, based on operation information of a reference wind turbine that is located in the vicinity of the diagnosis target wind turbine; an anomaly determination unit that calculates a degree of anomaly that indicates the magnitude of the difference between the predicted wind direction and wind speed and the wind direction and wind speed included in the operation information of the diagnosis target wind turbine, based on the predicted wind direction and wind speed and the wind direction and wind speed included in the operation information of the diagnosis target wind turbine; a control change command transmission unit that transmits to the diagnosis target wind turbine a command to change control of the diagnosis target wind turbine when the degree of abnormality of the diagnosis target wind turbine exceeds a predetermined value, A wind turbine control unit provided in the diagnosis target wind turbine, When a command to change the control is received, the control of the wind turbine is changed. Wind power generation system.

11. The monitoring device includes: a wind direction and wind speed transmitting unit that repeatedly transmits predicted wind direction and wind speed to the diagnosis target wind turbine when the degree of abnormality of the diagnosis target wind turbine exceeds a predetermined value, The wind turbine control unit is The wind turbine is controlled based on the received wind direction and wind speed. The wind power generation system according to claim 10.

12. The wind turbine is Equipped with a plurality of wind vanes and anemometers, The control change command transmission unit is when the degree of abnormality of the anemometer and anemometer of the diagnosis target wind turbine exceeds a predetermined value, transmitting a command to change the control, the command including identification information of the anemometer and anemometer, to the diagnosis target wind turbine; The wind turbine control unit is When a command to change the control is received, the wind turbine is controlled based on the wind direction and wind speed measured by a wind vane and anemometer different from the wind vane and anemometer identified by the identification information. The wind power generation system according to claim 10.

13. The monitoring device, acquiring operation information of a plurality of wind turbines, the operation information including wind direction and wind speed, from the plurality of wind turbines each equipped with an anemometer and an anemometer; predicting the wind direction and wind speed in the diagnosis target wind turbine, which is a wind turbine equipped with the anemometer and anemometer to be diagnosed, based on operation information of a reference wind turbine located in the vicinity of the diagnosis target wind turbine; and calculating, based on the predicted wind direction and wind speed and the wind direction and wind speed included in the operation information of the diagnosis target wind turbine, a degree of anomaly indicating the magnitude of the difference between the predicted wind direction and wind speed and the wind direction and wind speed included in the operation information of the diagnosis target wind turbine. Monitoring method.

Images