Early warning method and device for wind power plant, electronic equipment and storage medium

Abstract

The invention provides an early warning method and device for a wind power plant, electronic equipment and a storage medium. The early warning method for the wind power plant comprises the steps of obtaining severe convection weather real-time data of a target area where a target wind power plant is located; based on the severe convection weather real-time data, a severe convection strong wind identification model is utilized to determine a severe convection strong wind area in the target area; according to the severe convection weather extrapolation forecast data of the target area, predicting a movement track of the severe convection strong wind area; and performing strong convection strong wind early warning of the target wind power plant in response to the condition that the predicted moving track meets a preset condition.

Description

Technical Field

[0001] This disclosure generally relates to the field of wind power generation technology, and more specifically, to a method, apparatus, electronic device, and storage medium for early warning of wind farms. Background Technology

[0002] Severe convective weather refers to processes accompanied by strong localized vertical atmospheric movements, resulting in severe weather phenomena such as thunderstorms, strong winds, and hail. It is characterized by its strong locality, short duration, sudden occurrence, intense intensity, and high destructiveness. For wind farms, severe convective weather poses dangers to on-site maintenance personnel. Furthermore, the accompanying downbursts and other weather processes can rapidly increase the load on wind turbine generators (hereinafter referred to as turbines) within a short period, posing a threat to the turbines. Summary of the Invention

[0003] Exemplary embodiments of this disclosure provide a wind farm early warning method, apparatus, electronic device, and storage medium that can improve the accuracy of early warnings for strong convection and strong winds in wind farms.

[0004] According to a first aspect of the present disclosure, a method for early warning of wind farms is provided, comprising: acquiring real-time data of severe convective weather in a target area where the target wind farm is located; determining a severe convective wind area in the target area based on the real-time data of severe convective weather and using a severe convective wind identification model; predicting the movement trajectory of the severe convective wind area based on extrapolated forecast data of severe convective weather in the target area; and issuing a severe convective wind warning for the target wind farm in response to the predicted movement trajectory meeting preset conditions.

[0005] Optionally, the step of determining the strong convective and strong wind areas in the target area based on the real-time data of severe convective weather and using the severe convective and strong wind identification model includes: determining the severe convective areas in the target area based on the real-time data of severe convective weather; inputting the real-time data of severe convective weather in the severe convective areas, or the severe convective weather features obtained by feature extraction from the real-time data of severe convective weather in the severe convective areas, into the severe convective and strong wind identification model to obtain the strong wind level of the severe convective areas; and identifying the severe convective areas with strong wind levels higher than a preset level as severe convective and strong wind areas.

[0006] Optionally, the severe convective weather characteristics include at least one of the following: the maximum radar echo value of the severe convective region, the area of ​​the region in the severe convective region where the radar echo value is higher than a first preset threshold, the aspect ratio of the severe convective region, the moving speed of the severe convective region, the top height of the convective system in the severe convective region, and the radar-observed atmospheric liquid water content in the severe convective region.

[0007] Optionally, the method further includes: acquiring a historical sample set comprising multiple historical samples; constructing or training the severe convective wind identification model based on the historical sample set; wherein each historical sample includes: severe convective weather data at the wind farm during the time period in which severe convective weather occurs, and a severe wind level label obtained based on the wind speed data of the wind farm during the time period; wherein the wind speed data is data obtained by observing the wind speed at a preset height range, and the preset height range is determined based on the hub height and / or rotor surface height of the wind turbine generator set of the wind farm.

[0008] Optionally, the step of constructing or training the severe convective and strong wind identification model based on the historical sample set includes: training the severe convective and strong wind identification model based on the historical sample set using a machine learning algorithm; or, extracting features from the severe convective weather data in each historical sample to obtain the corresponding severe convective weather features, and constructing the severe convective and strong wind identification model based on the correlation between the severe convective weather features corresponding to each historical sample and the strong wind level label.

[0009] Optionally, the step of determining the severe convective region in the target area based on the real-time severe convective weather data includes: determining the region in the target area with a radar combined reflectivity greater than a second preset threshold as a severe convective region.

[0010] Optionally, the step of predicting the movement trajectory of the severe convective and strong wind region based on the severe convective weather extrapolation forecast data of the target region includes: using optical flow method to match and track the movement of the severe convective and strong wind region based on the real-time severe convective weather data and the severe convective weather extrapolation forecast data of the target region.

[0011] Optionally, the step of issuing a strong convective wind warning for the target wind farm includes: notifying the wind turbine generators in the target wind farm to perform protective actions; wherein the protective actions include at least one of the following: retracting the propellers, limiting the maximum pitch angle, and yawing to a designated position.

[0012] Optionally, the real-time data of severe convective weather includes at least one of the following: radar combined reflectivity, radar echo top height, radar-observed atmospheric liquid water content, convective system top height, satellite-observed cloud top height, satellite-observed cloud top brightness temperature, satellite-observed cloud top pressure, and satellite-observed cloud cover; and / or, the extrapolated forecast data of severe convective weather includes at least one of the following: extrapolated forecast value of radar combined reflectivity, extrapolated forecast value of radar-observed atmospheric liquid water content, extrapolated forecast value of radar echo top height, and extrapolated forecast value of satellite-observed cloud top height.

[0013] According to a second aspect of the present disclosure, a wind farm early warning device is provided, comprising: a data acquisition unit configured to acquire real-time data of severe convective weather in a target area where the target wind farm is located; a region determination unit configured to determine a severe convective wind region in the target area based on the real-time data of severe convective weather and using a severe convective wind identification model; a movement prediction unit configured to predict the movement trajectory of the severe convective wind region based on the extrapolated forecast data of severe convective weather in the target area; and an early warning unit configured to issue a severe convective wind early warning for the target wind farm in response to the predicted movement trajectory meeting preset conditions.

[0014] According to a third aspect of the present disclosure, a computer-readable storage medium storing a computer program is provided, which, when executed by a processor, causes the processor to perform the wind farm early warning method as described above.

[0015] According to a fourth aspect of the present disclosure, an electronic device is provided, the electronic device comprising: a processor; and a memory storing a computer program, wherein when the computer program is executed by the processor, it causes the processor to perform the wind farm early warning method as described above.

[0016] According to a fifth aspect of the present disclosure, a computer program product is provided, including a computer program that, when executed by a processor, implements the wind farm early warning method as described above.

[0017] The wind farm early warning method, apparatus, electronic device, and storage medium according to exemplary embodiments of the present disclosure utilize real-time data and extrapolated forecast data of severe convective weather to identify and provide short-term forecasts of areas affected by short-term strong winds during severe convective weather, thereby accurately providing on-site workers and on-site turbine units with early warning information related to short-term strong winds.

[0018] In the following description, some aspects and / or advantages of the general concept of this disclosure will be set forth, and other aspects and / or advantages will become apparent from the following description or from practice of the general concept of this disclosure. Attached Figure Description

[0019] These and / or other aspects and advantages of this application will become clearer and more readily understood from the following detailed description of embodiments of this application taken in conjunction with the accompanying drawings, wherein:

[0020] Figure 1 A flowchart illustrating an early warning method for a wind farm according to an exemplary embodiment of the present disclosure is provided.

[0021] Figure 2 A flowchart illustrating a method for determining a region of strong convection and strong winds according to an exemplary embodiment of the present disclosure;

[0022] Figure 3 A flowchart illustrating a method for obtaining a strong convective wind identification model according to an exemplary embodiment of the present disclosure;

[0023] Figure 4 An example of a statistically based model for identifying strong convective winds according to an exemplary embodiment of the present disclosure is shown;

[0024] Figure 5 A flowchart illustrating an early warning method for a wind farm according to another exemplary embodiment of the present disclosure is provided.

[0025] Figure 6 A structural block diagram of a wind farm early warning device according to an exemplary embodiment of the present disclosure is shown. Detailed Implementation

[0026] The embodiments of this disclosure will now be described in detail with reference to the accompanying drawings, examples of which are illustrated in the drawings, wherein the same reference numerals always refer to the same parts. The embodiments will now be described with reference to the accompanying drawings in order to explain this disclosure.

[0027] It should be noted that the terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims.

[0028] It should be noted that the phrase "at least one of several items" in this disclosure refers to three parallel cases: "any one of the several items", "a combination of any number of the several items", and "all of the several items". For example, "including at least one of A and B" includes the following three parallel cases: (1) including A; (2) including B; (3) including A and B. As another example, "performing at least one of step one and step two" indicates the following three parallel cases: (1) performing step one; (2) performing step two; (3) performing both step one and step two.

[0029] For ease of understanding, some terms used in the exemplary embodiments of this disclosure will be explained below.

[0030] Severe convective weather: Severe convective weather refers to weather phenomena accompanied by thunderstorms, including strong convective winds, hail, and short-duration heavy rainfall.

[0031] Severe convection and strong winds: Short-term strong winds that accompany severe convective weather.

[0032] Radar composite reflectivity: The ratio of radar waves reflected by clouds at different altitudes within a certain range received by a weather radar. It reflects the size and number density of precipitation particles inside the meteorological target and is often used to represent the intensity of meteorological targets such as raindrops.

[0033] Radar echo top height: The highest altitude above the ground reached by a strong radar echo region within a cloud (e.g., equivalent reflectivity factor > 36 dBz).

[0034] Satellite-observed cloud top height: The highest cloud height retrieved from satellite observation data.

[0035] Figure 1 A flowchart illustrating an early warning method for a wind farm according to an exemplary embodiment of the present disclosure is shown.

[0036] As an example, the wind farm early warning method according to the exemplary embodiments of this disclosure can be executed by an electronic device with data processing capabilities, such as a terminal (e.g., a personal laptop, desktop computer, etc.), a server (e.g., a standalone server, server cluster, cloud platform, etc.), or a wind farm-level controller. This disclosure does not limit the scope of the method.

[0037] Reference Figure 1 In step S101, real-time data of severe convective weather in the target area where the target wind farm is located is obtained.

[0038] The target area includes the target wind farm. For example, the target area can be a certain range outside the target wind farm.

[0039] As an exemplary embodiment, the real-time severe convective weather data may include, but is not limited to, at least one of the following: real-time observed severe convective weather data, and severe convective weather data obtained by inversion from real-time observed severe convective weather data.

[0040] As an exemplary embodiment, the type of real-time severe convective weather data may include, but is not limited to, at least one of the following: weather radar monitoring data, data obtained by inversion from weather radar monitoring data, satellite monitoring data, and data obtained by inversion from satellite monitoring data.

[0041] As an exemplary embodiment, real-time data on severe convective weather may include, but is not limited to, at least one of the following: radar combined reflectivity, radar-observed atmospheric liquid water content, convective system top height, radar echo top height, satellite-observed cloud top height, satellite-observed cloud top brightness temperature, satellite-observed cloud top pressure, and satellite-observed cloud cover.

[0042] In step S102, based on real-time data of severe convective weather in the target area, the severe convective and strong wind areas in the target area are determined using a severe convective and strong wind identification model.

[0043] The area of ​​strong convection and strong winds is the area where strong convective weather and strong winds occur, also known as the area of ​​strong convection and strong winds.

[0044] As an exemplary embodiment, step S102 may include: determining the strong convective region (i.e., the region where strong convective weather occurs) within the target region based on real-time data of severe convective weather in the target region; and then, based on the real-time data of severe convective weather in the strong convective region, using a strong convective and strong wind identification model to determine the strong convective and strong wind region from the strong convective region (i.e., extracting the strong convective region for identification). Alternatively, the real-time data of severe convective weather in the target region may be directly input into the strong convective and strong wind identification model to obtain the strong convective and strong wind region (i.e., directly identifying the entire target region). It should be understood that the strong convective and strong wind identification models used in the above two methods are different.

[0045] The following will combine Figure 2 An exemplary embodiment of step S102 will be described below, but will not be elaborated here.

[0046] In step S103, the movement trajectory of the severe convective and strong wind area is predicted based on the extrapolation forecast data of severe convective weather in the target area.

[0047] Extrapolation forecasting of weather systems refers to extrapolating the past evolutionary trends of weather systems to predict future conditions.

[0048] As an exemplary embodiment, the extrapolation forecast data for severe convective weather may include, but is not limited to, at least one of the following: extrapolated radar combined reflectivity, extrapolated radar observation of atmospheric liquid water content, extrapolated radar echo top height, and extrapolated satellite observation of cloud top height. It should be understood that the extrapolation variables used may be the same as or different from those used in identifying the areas affected by severe convection and strong winds.

[0049] As an exemplary embodiment, step S103 may include: using optical flow to match and track the movement of severe convective and strong wind regions based on real-time data and extrapolated forecast data of severe convective weather in the target area. Alternatively, a severe convective and strong wind region tracking model may be used to track the movement of these regions, enabling short-term forecasts of their movement.

[0050] According to an exemplary embodiment of this disclosure, real-time data and short-term extrapolation data of severe convective weather are fused to achieve short-term forecasts of the areas affected by severe convection and strong winds. Preferably, the areas affected by severe convection and strong winds can be identified first using real-time data of severe convective weather, and then the movement of these areas can be tracked based on AI extrapolation forecast results.

[0051] In step S104, in response to the predicted movement trajectory of the strong convection and strong wind area meeting the preset conditions, a strong convection and strong wind warning is issued for the target wind farm.

[0052] As an exemplary embodiment, the predicted trajectory of the severe convective wind region can indicate the location where the region will move next. Furthermore, the predicted trajectory can also indicate the direction and speed of movement of the severe convective wind region.

[0053] As an exemplary embodiment, the preset condition may include: the distance between the predicted location to which the strong convective wind area will move and the target wind farm is less than a first preset distance. It should be understood that the specific value of the first preset distance can be freely set according to actual needs.

[0054] As another exemplary embodiment, the preset conditions may include: the distance from the predicted location to which the strong convective wind region will move, as indicated by its predicted trajectory, to the target wind farm is less than a second preset distance; and the direction of movement indicated by the predicted trajectory of the strong convective wind region is towards the target wind farm. It should be understood that the specific value of the second preset distance can be freely set according to actual needs, and the first preset distance may be the same as or different from the second preset distance.

[0055] As an exemplary embodiment, the step of issuing a strong convective wind warning for a target wind farm may include: notifying the wind turbine generators in the target wind farm to perform protective actions. As an example, protective actions may include, but are not limited to, at least one of the following: retracting the pitch (e.g., retracting the pitch in light winds), limiting the maximum pitch angle (e.g., limiting the pitch angle in strong winds), and yawing to a designated position. Upon receiving the warning, the generators may perform protective actions via automatic commands.

[0056] Regarding retracting the propeller in light wind, if the propeller is in the open state in light wind, and the propeller jams during a sudden increase in wind, it will cause a serious accident to the unit. Therefore, the propeller can be retracted in advance in light wind.

[0057] High wind limit pitch angle: Limits the maximum pitch angle to reduce the windward force on the blades and achieve load reduction.

[0058] Yaw to the designated position: Take yaw action in advance to reduce the load on the crew when encountering strong winds.

[0059] Figure 5 A flowchart illustrating an early warning method for a wind farm according to another exemplary embodiment of the present disclosure is shown.

[0060] According to an exemplary embodiment of this disclosure, the area affected by severe convection and strong winds is first identified using real-time observation data, and then the movement of the area affected by severe convection and strong winds is tracked using extrapolation forecast results. This avoids the problem of rapid decrease in extreme values ​​in extrapolation forecasts, reduces the requirements for extrapolation forecast data sources, and improves the reliability, stability, and accuracy of the early warning.

[0061] Figure 2 A flowchart illustrating a method for determining a region of strong convection and strong winds according to an exemplary embodiment of the present disclosure is shown.

[0062] Reference Figure 2 In step S201, based on real-time data of severe convective weather in the target area, the severe convective area in the target area is determined.

[0063] As an exemplary embodiment, step S201 may include: identifying areas in the target region with a combined radar reflectivity greater than a second preset threshold as strong convection areas. The second preset threshold can be set according to actual conditions and specific needs; for example, the second preset threshold may be between 40 and 60 dBz, preferably between 45 and 55 dBz. It should be understood that strong convection areas can also be determined by other appropriate methods.

[0064] As an exemplary embodiment, the severe convective region can be extracted by image segmentation based on real-time data of severe convective weather in the target area.

[0065] In step S202, the real-time data of severe convective weather in the severe convective region, or the severe convective weather features obtained by feature extraction from the real-time data of severe convective weather in the severe convective region, are input into the severe convective wind identification model to obtain the wind level of the severe convective region.

[0066] As an exemplary embodiment, the strong convection and strong wind identification model can be used to identify the wind intensity in a strong convection area to determine the probability of different instantaneous wind speeds or strong wind levels. For example, the strong convection and strong wind identification model can be used to identify the wind intensity in a strong convection area at the turbine hub height / impeller surface height. Furthermore, the strong convection and strong wind identification model can also be used to identify specific wind speed values ​​in a strong convection area, which can then be used to determine the wind intensity.

[0067] As an exemplary embodiment, the type of strong convection and strong wind identification model can be a statistical model or a machine learning model.

[0068] As an exemplary embodiment, real-time data of severe convective weather in a severe convective region can be input into a severe convective and strong wind identification model to obtain the strong wind level of the severe convective region output by the model. This severe convective and strong wind identification model can be a machine learning model. Since machine learning models can also perform feature extraction, real-time data of severe convective weather in a severe convective region can be directly input into the model.

[0069] As another exemplary embodiment, the severe convective weather features obtained by extracting features from real-time data of severe convective weather in a severe convective region can be input into a severe convective and strong wind identification model to obtain the strong wind level of the severe convective region output by the model. Here, the severe convective and strong wind identification model can be a machine learning model or a statistical model.

[0070] As an example, severe convective weather characteristics may include, but are not limited to, at least one of the following: the maximum radar echo value of the severe convective region, the area of ​​the region within the severe convective region where the radar echo value is higher than a first preset threshold, the aspect ratio of the severe convective region, the movement speed of the severe convective region, the top height of the convective system in the severe convective region, and the radar-observed atmospheric liquid water content in the severe convective region. The first preset threshold can be set according to actual conditions and specific needs; as an example, the first preset threshold can be 60 dBz.

[0071] In step S203, areas with strong winds exceeding a preset level are defined as areas with strong convection and strong winds.

[0072] According to an exemplary embodiment of this disclosure, a strong convection and strong wind identification model is used to convert real-time data of strong convective weather into areas of strong convection and strong wind in real time.

[0073] The following will combine Figure 3 An exemplary embodiment of the method for obtaining a strong convection and strong wind identification model is described below.

[0074] Reference Figure 3 In step S301, a historical sample set (i.e., a strong convection and strong wind dataset) including multiple historical samples is obtained.

[0075] Each historical sample includes: severe convective weather data for a wind farm (including but not limited to the target wind farm) during a specific time period in which severe convective weather occurred, and a strong wind level label derived from the wind speed data of the wind farm during that time period. The wind speed data for the wind farm during that time period is obtained by observing wind speeds within a preset height range, which is determined based on the hub height and / or rotor surface height of the wind turbine generators at the wind farm. For example, the preset height range can be either a hub height range or a rotor surface height range. For example, the wind speed data for the wind farm during that time period can be observation data from the turbine generators within the wind farm or data from a meteorological tower.

[0076] As an exemplary embodiment, measured data from wind farm turbines or anemometer data can be collected as a true data source for strong winds; strong convective weather data observed by weather radar, satellites, etc., in the corresponding area during the corresponding time period can be collected as a modeling data source. Next, time matching is performed to extract measured wind speed characteristic values ​​within the strong convective weather observation coverage period (for example, if strong convective weather observations are updated every 6 minutes, then the maximum wind speed value over 6 minutes is preferably extracted), constructing a historical sample set. This process may include the following sub-steps:

[0077] (a) Region segmentation: Segment the region where the radar combined reflectivity is greater than a specific value (i.e., the second preset threshold mentioned above). For example, the specific value is between 40 and 60 dBz, preferably between 45 and 55 dBz.

[0078] (b) Expand the segmented region to obtain a strong convection segmented region: preferably, expand it to within 30 km, more preferably, expand it to 5 to 20 km;

[0079] (c) Severe Convection-Strong Wind Observation Matching: Using the severe convection segmentation area as the sample, the wind speed at the corresponding time moment is used as the ground truth data source. Preferably, the maximum wind speed during the passage of the strong wind system is taken as the maximum wind speed of the severe convection system. For example, if a location is affected by the severe convection system for a total of 2 hours, during which the wind speed first increases and then decreases, then the maximum wind speed within these 2 hours is preferably used as the wind speed label of the severe convection system.

[0080] According to the exemplary embodiments of this disclosure, using data that is closer to the hub height of the turbine, such as turbine nacelle wind measurement data, ground-based radar wind measurement data, and wind measurement tower data, as the true data source, rather than surface wind speed, is more suitable for wind farm application scenarios.

[0081] In step S302, a strong convection and strong wind identification model is constructed or trained based on the historical sample set.

[0082] As an exemplary embodiment, a strong convection and strong wind identification model is obtained by training a machine learning algorithm based on a historical sample set. The type of the strong convection and strong wind identification model obtained here is a machine learning model.

[0083] As an example, object detection and pattern recognition algorithms such as Unet can be used to build the model.

[0084] As an example, training-validation set construction: samples are randomly grouped. Preferably, 60-80% of the samples are selected as the training set and directly used for training; the remaining 20-40% are used as the validation set to verify the accuracy of the training.

[0085] As an example, machine learning model training: Machine learning can be performed using target recognition-related machine learning algorithms such as CNN, Unet, and YOLO to estimate the short-term strong wind level.

[0086] As another exemplary embodiment, features are extracted from the severe convective weather data in each historical sample to obtain the corresponding severe convective weather features. Based on the correlation between the severe convective weather features corresponding to each historical sample and the strong wind level label, a severe convective strong wind identification model is constructed. The type of the severe convective strong wind identification model obtained here is a statistical model.

[0087] As an example, such as Figure 4 As shown, the characteristics of severe convective weather during periods of strong winds can be statistically analyzed to establish patterns. For example, if radar combined reflectivity is used as sample data, the analysis can be performed according to the following logic:

[0088] (a) Statistical analysis of the maximum echo value within the strong convection cut-off region, the area of ​​regions with values ​​>60 dBz (i.e., the first preset threshold mentioned above), and the aspect ratio of the region shape, etc.

[0089] (b) Match the maximum wind speed within the corresponding time period, and use machine learning methods such as logistic statistics or random forest to clarify the relationship between strong wind state and various features, and build a model.

[0090] This disclosure uses wind farm data as the ground truth for constructing a strong convection and strong wind identification model, making this disclosure more suitable for early warning of short-term strong winds caused by strong convection in wind farms.

[0091] Short-term strong convection warning for wind farms is an important part of wind farm meteorological early warning guarantee. According to the exemplary embodiment of this disclosure, a method for short-term strong wind warning of strong convection in wind farms is provided. It uses one or more data from weather radar, satellite data, etc. as forecast data sources and uses algorithms such as image segmentation and pattern recognition to realize the method for short-term strong wind warning of strong convection. It can segment the strong convection system and combine short-term extrapolation to track the strong convection strong wind landing area to achieve accurate early warning of the strong convection strong wind landing area. It has the following advantages: (1) When constructing or training the strong convection strong wind recognition model, the wind speed at the hub height or impeller surface height of the unit is used as the true data source. Compared with the surface wind speed, it can better represent the wind speed of the environment where the unit is located. For example, the hub height of the unit can reach 100m or even higher. The wind measurement data of the surface automatic station cannot truly represent the situation faced by the unit. Therefore, this disclosure has better adaptability to wind farms. (2) Emphasizing the method of first using real-time observation data to identify the area of ​​strong convection and strong winds, and then using extrapolation forecast results to track the displacement of the area of ​​strong convection and strong winds, is beneficial to improving the accuracy of early warning. This disclosure considers that since existing short-term extrapolation algorithms are generally inaccurate in predicting the intensity, especially the intensity of the center of strong convection, the early warning results that directly refer to AI extrapolation forecasts during the early warning process are often inaccurate due to inaccurate learning information or inaccurate extrapolation results. (3) Emphasizing not relying on real-time automatic surface station temperature, pressure, humidity, wind and rain monitoring, and not relying on mesoscale model forecast data. This disclosure considers that if the forecast needs to combine measured radar and regional mesoscale model forecasts, on the one hand, it increases the cost of short-term strong wind warnings for strong convection, and on the other hand, mesoscale model forecasts may not contain information directly related to strong convection. If relying on ground observation and regional numerical model data, ground observation is highly localized and may not be able to monitor the occurrence of strong convection in real time, and real-time monitoring data from automatic surface stations is difficult to obtain for non-meteorological bureau commercial scenarios; the temporal and spatial resolution of regional numerical model results are relatively coarse and cannot accurately match the strong convection process.

[0092] Figure 6 A structural block diagram of a wind farm early warning device according to an exemplary embodiment of the present disclosure is shown.

[0093] Reference Figure 6 The early warning device for a wind farm according to an exemplary embodiment of the present disclosure includes: a data acquisition unit 100, a region determination unit 200, a movement prediction unit 300, and an early warning unit 400.

[0094] Specifically, the data acquisition unit 100 is configured to acquire real-time data on severe convective weather in the target area where the target wind farm is located.

[0095] The region determination unit 200 is configured to determine the strong convection and strong wind regions in the target region based on the real-time data of the severe convective weather and using a severe convection and strong wind identification model.

[0096] The movement prediction unit 300 is configured to predict the movement trajectory of the severe convective and strong wind area based on the severe convective weather extrapolation forecast data of the target area.

[0097] The early warning unit 400 is configured to issue a strong convective wind warning for the target wind farm in response to the predicted movement trajectory meeting preset conditions.

[0098] As an exemplary embodiment, the region determination unit 200 may be configured to: determine a severe convective region in the target region based on the real-time severe convective weather data; input the real-time severe convective weather data of the severe convective region, or the severe convective weather features obtained by feature extraction from the real-time severe convective weather data of the severe convective region, into the severe convective wind identification model to obtain the wind level of the severe convective region; and determine the severe convective region with a wind level higher than a preset level as a severe convective wind region.

[0099] As an exemplary embodiment, the severe convective weather characteristics may include, but are not limited to, at least one of the following: the maximum radar echo value of the severe convective region, the area of ​​the region in the severe convective region where the radar echo value is higher than a first preset threshold, the aspect ratio of the severe convective region, the moving speed of the severe convective region, the top height of the convective system in the severe convective region, and the radar-observed atmospheric liquid water content in the severe convective region.

[0100] As an exemplary embodiment, the early warning device for a wind farm according to an exemplary embodiment of this disclosure may further include: a model acquisition unit (not shown), the model acquisition unit being configured to: acquire a historical sample set including multiple historical samples; and construct or train the strong convective wind identification model based on the historical sample set; wherein each historical sample includes: strong convective weather data at the wind farm during the time period in which strong convective weather occurs at the wind farm, and a strong wind level label obtained based on the wind speed data of the wind farm during the time period; wherein the wind speed data is data obtained by observing the wind speed at a preset height range, the preset height range being determined based on the hub height and / or rotor surface height of the wind turbine generator set of the wind farm.

[0101] As an exemplary embodiment, the model acquisition unit may be configured to: train the severe convective and strong wind identification model based on the historical sample set using a machine learning algorithm; or, extract features from the severe convective weather data in each historical sample to obtain the corresponding severe convective weather features, and construct the severe convective and strong wind identification model based on the correlation between the severe convective weather features corresponding to each historical sample and the strong wind level label.

[0102] As an exemplary embodiment, the region determination unit 200 may be configured to: determine the region in the target region whose radar combined reflectivity is greater than a second preset threshold as a strong convection region.

[0103] As an exemplary embodiment, the region determination unit 200 may be configured to: use optical flow method to match and track the movement of the severe convective and strong wind regions based on real-time data of severe convective weather and extrapolated forecast data of severe convective weather in the target region.

[0104] As an exemplary embodiment, the early warning unit 400 may be configured to: notify the wind turbine generators in the target wind farm to perform protective actions; wherein the protective actions include at least one of the following: retracting the propeller, limiting the maximum pitch angle, and yawing to a specified position.

[0105] As an exemplary embodiment, the real-time data of severe convective weather may include, but is not limited to, at least one of the following: radar combined reflectivity, radar echo top height, radar-observed atmospheric liquid water content, convective system top height, satellite-observed cloud top height, satellite-observed cloud top brightness temperature, satellite-observed cloud top pressure, and satellite-observed cloud cover; and / or, the extrapolated forecast data of severe convective weather may include, but is not limited to, at least one of the following: extrapolated forecast value of radar combined reflectivity, extrapolated forecast value of radar-observed atmospheric liquid water content, extrapolated forecast value of radar echo top height, and extrapolated forecast value of satellite-observed cloud top height.

[0106] It should be understood that the specific processing performed by the early warning device for a wind farm according to the exemplary embodiments of this disclosure has been referenced. Figures 1 to 5 A detailed description has been provided, and the relevant details will not be repeated here.

[0107] It should be understood that the various units in the wind farm early warning device according to the exemplary embodiments of this disclosure may be implemented as hardware components and / or software components. Those skilled in the art may implement the various units, for example, using field-programmable gate arrays (FPGAs) or application-specific integrated circuits (ASICs), based on the processes performed by the defined various units.

[0108] An electronic device according to an exemplary embodiment of the present disclosure includes a processor (not shown) and a memory (not shown), wherein the memory stores a computer program that, when executed by the processor, causes the processor to perform a wind farm early warning method as described in the exemplary embodiment above.

[0109] As an example, the electronic device may be an electronic device with data processing capabilities. For example, the electronic device may be a terminal (such as a personal laptop, desktop computer, etc.), a server (such as a standalone server, server cluster, cloud platform, etc.), or a wind farm-level controller. This disclosure does not limit this.

[0110] According to exemplary embodiments of the present disclosure, a computer-readable storage medium storing instructions may also be provided, wherein when the instructions are executed by at least one processor, they cause at least one processor to perform the wind farm early warning method as described in the exemplary embodiments above. Examples of computer-readable storage media herein include: read-only memory (ROM), random access programmable read-only memory (PROM), electrically erasable programmable read-only memory (EEPROM), random access memory (RAM), dynamic random access memory (DRAM), static random access memory (SRAM), flash memory, non-volatile memory, CD-ROM, CD-R, CD+R, CD-RW, CD+RW, DVD-ROM, DVD-R, DVD+R, DVD-RW, DVD+RW, DVD-RAM, BD-ROM, BD-R, BD-R LTH, BD-RE, Blu-ray or optical disc storage, hard disk drive (HDD), solid-state drive (SSD), card storage (such as multimedia cards, secure digital (SD) cards, or ultra-fast digital (XD) cards), magnetic tape, floppy disk, magneto-optical data storage device, optical data storage device, hard disk, solid-state drive, and any other device configured to store a computer program and any associated data, data files, and data structures in a non-transitory manner and to provide the computer program and any associated data, data files, and data structures to a processor or computer so that the processor or computer can execute the computer program. The computer program in the aforementioned computer-readable storage medium can run in an environment deployed in computer devices such as clients, hosts, agent devices, servers, etc. Furthermore, in one example, the computer program and any associated data, data files, and data structures are distributed across a networked computer system, such that the computer program and any associated data, data files, and data structures are stored, accessed, and executed in a distributed manner through one or more processors or computers.

[0111] According to exemplary embodiments of the present disclosure, a computer program product may also be provided, wherein the instructions in the computer program product are executable by at least one processor to perform the wind farm early warning method as described in the exemplary embodiments above.

[0112] Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the claims.

[0113] It should be understood that this disclosure is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this disclosure is limited only by the appended claims.

Claims

1. A method for early warning of wind farms, characterized in that, include: Acquire real-time data on severe convective weather in the target area where the target wind farm is located; Based on the real-time data of severe convective weather, the severe convective and strong wind identification model is used to determine the areas of severe convective and strong wind in the target area. Based on the extrapolated forecast data of severe convective weather in the target area, predict the movement trajectory of the severe convective and strong wind area; In response to the predicted movement trajectory meeting preset conditions, a strong convective wind warning is issued for the target wind farm.

2. The early warning method according to claim 1, characterized in that, The step of determining the area of ​​severe convection and strong wind in the target area based on the real-time data of severe convective weather and using the severe convection and strong wind identification model includes: Based on the real-time data of the severe convective weather, the severe convective areas in the target area are determined; The severe convective weather real-time data of the severe convective region, or the severe convective weather features obtained by feature extraction from the severe convective weather real-time data of the severe convective region, are input into the severe convective wind identification model to obtain the wind level of the severe convective region. Areas with strong convective winds exceeding the preset level are designated as areas with strong convective winds.

3. The early warning method according to claim 2, characterized in that, The severe convective weather characteristics include at least one of the following: The maximum radar echo value of the strong convection region, the area of ​​the region in the strong convection region where the radar echo value is higher than the first preset threshold, the aspect ratio of the strong convection region, the moving speed of the strong convection region, the top height of the convection system in the strong convection region, and the radar-observed atmospheric liquid water content in the strong convection region.

4. The early warning method according to claim 1, characterized in that, Also includes: Obtain a historical sample set that includes multiple historical samples; Based on the historical sample set, the strong convection and strong wind identification model is constructed or trained. Each historical sample includes: severe convective weather data at the wind farm during the period in which severe convective weather occurred, and a strong wind level label obtained based on the wind speed data of the wind farm during the period in which severe convective weather occurred; The wind speed data is obtained by observing the wind speed at a preset height range, which is determined based on the hub height and / or rotor surface height of the wind turbine generator set in the wind farm.

5. The early warning method according to claim 4, characterized in that, The steps for constructing or training the strong convective wind identification model based on the historical sample set include: Based on the historical sample set, the strong convection and strong wind identification model is trained using machine learning algorithms. or, Feature extraction is performed on the severe convective weather data in each historical sample to obtain the corresponding severe convective weather features. Based on the correlation between the severe convective weather features corresponding to each historical sample and the strong wind level label, the severe convective strong wind identification model is constructed.

6. The early warning method according to claim 2, characterized in that, The step of determining the severe convective region in the target area based on the real-time severe convective weather data includes: The region in the target area with a radar combined reflectivity greater than a second preset threshold is identified as a strong convection region.

7. The early warning method according to claim 1, characterized in that, The step of predicting the movement trajectory of the severe convective and strong wind area based on the severe convective weather extrapolation forecast data of the target area includes: Based on real-time data and extrapolated forecast data of severe convective weather in the target area, optical flow method is used to match and track the movement of the severe convective and strong wind areas.

8. The early warning method according to claim 1, characterized in that, The steps for issuing a severe convective wind warning for the target wind farm include: The wind turbine generators in the target wind farm are notified to perform protection actions. The protective actions include at least one of the following: retracting the propeller, limiting the maximum pitch angle, and yawing to a specified position.

9. The early warning method according to claim 1, characterized in that, The real-time data on severe convective weather includes at least one of the following: radar combined reflectivity, radar echo top height, radar-observed atmospheric liquid water content, convective system top height, satellite-observed cloud top height, satellite-observed cloud top brightness temperature, satellite-observed cloud top pressure, and satellite-observed cloud cover. And / or, The extrapolated forecast data for severe convective weather includes at least one of the following: extrapolated forecast values ​​of radar combined reflectivity, extrapolated forecast values ​​of atmospheric liquid water content observed by radar, extrapolated forecast values ​​of radar echo top height, and extrapolated forecast values ​​of cloud top height observed by satellite.

10. A wind farm early warning device, characterized in that, include: The data acquisition unit is configured to acquire real-time data on severe convective weather in the target area where the target wind farm is located. The region determination unit is configured to determine the strong convection and strong wind regions in the target region based on the real-time data of the severe convective weather and using a severe convection and strong wind identification model. The movement prediction unit is configured to predict the movement trajectory of the severe convective and strong wind area based on the severe convective weather extrapolation forecast data of the target area. The early warning unit is configured to issue a strong convective wind warning for the target wind farm in response to the predicted movement trajectory meeting preset conditions.

11. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it causes the processor to perform the wind farm early warning method as described in any one of claims 1 to 9.

12. An electronic device, characterized in that, The electronic device includes: processor; A memory storing a computer program that, when executed by a processor, causes the processor to perform the wind farm early warning method as described in any one of claims 1 to 9.

13. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the early warning method for wind farms as described in any one of claims 1 to 9.

Images