Wind power output prediction method, electronic device, storage medium and system

Abstract

The present disclosure provides a wind power output prediction method, electronic device, storage medium and system, which relate to the field of wind power technology. The method includes: periodically obtaining an initial meteorological data set corresponding to each receiving time node, and the initial meteorological data set includes initial meteorological sub-data of at least one dimension of at least one meteorological element; after obtaining the latest initial meteorological data set, identifying and smoothing the abnormal sub-data to obtain a smoothed meteorological data set; determining the average wind energy density within the target time period; using the smoothed meteorological data set and the average wind energy density within the target time period as model input features, and obtaining a wind power output prediction value through the model. In the present disclosure, the initial meteorological sub-data of each dimension of each meteorological element, as well as the average wind energy density that has a greater impact on wind power output, can be used as model input features, so that the model can predict wind power output based on a large number of features, and can obtain a higher prediction accuracy.

Description

Technical Field

[0001] The present disclosure relates to the field of wind power technology, and in particular to a wind power output prediction method, electronic equipment, storage medium, and system. Background Art

[0002] As wind power technology matures, the capacity of individual wind turbines and the scale of wind farms continue to expand, and the proportion of wind power in the total power generation of the power system is also increasing year by year. The increasing penetration power of wind farms is causing increasingly prominent problems for the power system, hindering its safe, stable, economical, and reliable operation. Timely and accurate forecasting of wind power output can enhance the safety, stability, economic efficiency, and controllability of the power system. Summary of the Invention

[0003] The present disclosure provides a wind power output prediction method, the method comprising:

[0004] Periodically acquiring an initial meteorological data set corresponding to each receiving time node; each of the initial meteorological data sets includes initial meteorological data corresponding to at least one meteorological element, and the initial meteorological data includes initial meteorological sub-data of at least one dimension of the corresponding meteorological element;

[0005] After obtaining the latest initial meteorological data set corresponding to the latest receiving time node, identifying abnormal sub-data from each latest initial meteorological sub-data;

[0006] When abnormal sub-data are identified, smoothing is performed on the identified abnormal sub-data to obtain a smoothed meteorological data set; when no abnormal sub-data are identified, the latest initial meteorological data set is used as a smoothed meteorological data set;

[0007] Determining the instantaneous wind energy density corresponding to the latest receiving time node;

[0008] Performing a rolling mean calculation on the instantaneous wind energy density to obtain an average wind energy density within a target time period; the target time period includes the latest receiving time node;

[0009] The smoothed meteorological data set and the average wind energy density within the target time period are used as input features and input into a wind power output prediction model, so that the wind power output prediction model outputs a wind power output prediction value for the target time period.

[0010] Optionally, performing a rolling mean calculation on the instantaneous wind energy density to obtain an average wind energy density within a target time period; the target time period includes the latest receiving time node, including:

[0011] Scrolling the first time window along the time axis to align the first time window with the target time period;

[0012] An average of the multiple instantaneous wind energy densities within the first time window is calculated to obtain an average wind energy density within the target time period.

[0013] Optionally, the smoothed meteorological sub-data includes at least smoothed wind speed and smoothed air density at the wind turbine hub, and determining the instantaneous wind energy density corresponding to the latest receiving time node includes:

[0014] The instantaneous wind energy density corresponding to the latest receiving time node is determined according to the smoothed wind speed at the wind turbine hub and the smoothed air density corresponding to the latest receiving time node.

[0015] Optionally, after obtaining the latest initial meteorological data set corresponding to the latest receiving time node, identifying abnormal sub-data from each latest initial meteorological sub-data includes:

[0016] After obtaining the latest initial meteorological data set corresponding to the latest receiving time node, scrolling the second time window along the time axis so that the second time window includes the latest receiving time node;

[0017] Performing normalization processing on the initial meteorological sub-data of the same dimension and non-null values ​​belonging to the same meteorological element within the second time window to obtain a normalized normalized value corresponding to each latest initial meteorological sub-data;

[0018] The latest initial meteorological sub-data whose normalized value is not within a preset numerical range is determined as the abnormal sub-data.

[0019] Optionally, after acquiring the latest initial meteorological data set corresponding to the latest receiving time node, scrolling the second time window along the time axis so that the second time window includes the latest receiving time node further includes:

[0020] The null value in each of the latest initial meteorological sub-data within the second time window is determined as the abnormal sub-data.

[0021] Optionally, when abnormal sub-data are identified, smoothing is performed on each of the identified abnormal sub-data to obtain a smoothed meteorological data set, including:

[0022] For any of the identified abnormal sub-data, scroll a third time window along the time axis so that the third time window includes the receiving time node corresponding to the abnormal sub-data and multiple receiving time nodes prior to the receiving time node corresponding to the abnormal sub-data;

[0023] performing mean calculation on the initial meteorological sub-data belonging to the same meteorological element and the same dimension as the abnormal sub-data within the third time window to obtain a new value corresponding to the abnormal sub-data;

[0024] Each abnormal sub-data is replaced with the corresponding new value to obtain a smoothed meteorological data set.

[0025] Optionally, the periodically acquiring the initial meteorological data set corresponding to each receiving time node includes:

[0026] An initial meteorological data prediction set corresponding to each receiving time node is periodically obtained; the initial meteorological data prediction set is predicted based on a historical initial meteorological data true value set corresponding to at least one historical receiving time node.

[0027] Optionally, the periodically acquiring the initial meteorological data set corresponding to each receiving time node includes:

[0028] Periodically obtain the initial meteorological data truth value set corresponding to each receiving time node.

[0029] Optionally, before periodically acquiring the initial meteorological data set corresponding to each receiving time node, the method further includes:

[0030] Acquire multiple first sample sets; the first sample sets include smoothed meteorological data set samples corresponding to each historical receiving time node in a historical time period, average wind energy density samples in the historical time period, and wind power output true value samples in the historical time period;

[0031] Build an improvement model;

[0032] Training the boosting model according to the plurality of first sample sets to obtain a trained boosting model;

[0033] The wind power output prediction model is generated according to the trained improved model.

[0034] Optionally, the training of the boosting model based on the plurality of first sample sets to obtain a trained boosting model includes:

[0035] The boosting model is trained according to the plurality of first sample sets by a cross-validation method to obtain a trained boosting model.

[0036] Optionally, before inputting the smoothed meteorological dataset and the average wind energy density within the target time period as input features into a wind power output prediction model so that the wind power output prediction model outputs a predicted wind power output value for the target time period, the method further includes:

[0037] Obtaining a historical wind power output true value corresponding to each receiving time node within the target time period;

[0038] The step of inputting the smoothed meteorological data set and the average wind energy density within the target time period as input features into a wind power output prediction model so that the wind power output prediction model outputs a predicted wind power output value for the target time period, including:

[0039] The true values ​​of the historical wind power outputs within the target time period, the smoothed meteorological data set, and the average wind energy density are input as input features into a wind power output prediction model, so that the wind power output prediction model outputs a predicted wind power output value for the target time period.

[0040] Optionally, before generating the wind power output prediction model based on the trained improved model, the method further includes:

[0041] Acquire multiple second sample sets; the second sample sets include historical wind power output true value samples corresponding to each of the historical receiving time nodes within the historical time period;

[0042] Build time series models;

[0043] Training the time series model according to the plurality of second sample sets to obtain a trained time series model;

[0044] Generating the wind power output prediction model according to the trained improved model includes:

[0045] The trained boosting model and the trained time series model are stacked and fused to obtain the wind power output prediction model.

[0046] Optionally, the method further includes:

[0047] After obtaining multiple new first sample sets and multiple new second sample sets, the wind power output prediction model is retrained according to the multiple new first sample sets and the multiple new second sample sets to update the wind power output prediction model.

[0048] Optionally, the meteorological elements include wind speed, gas density, air pressure and temperature.

[0049] Optionally, the latest initial meteorological sub-data includes at least the latest initial wind speed at the wind turbine hub, and the method further includes:

[0050] After obtaining the latest initial meteorological data set corresponding to the latest receiving time node, when the latest initial wind speed at the wind turbine hub is less than a first preset wind speed, or greater than a second preset wind speed, monitoring the initial wind speed at the wind turbine hub obtained each time from the latest receiving time node; the first preset wind speed is less than the second preset wind speed;

[0051] An early warning is issued based on the initial wind speed at the wind turbine hub obtained each time during monitoring within a preset time period.

[0052] Optionally, the providing of an early warning prompt based on the initial wind speed at the wind turbine hub acquired each time during monitoring within a preset time period includes:

[0053] When the initial wind speed at the wind turbine hub acquired each time during the preset time period is less than the first preset wind speed, outputting a warning prompt for recommending shutting down the wind turbines in the wind farm; or

[0054] Counting the number of wind speed data that are less than the first preset wind speed among the initial wind turbine hub wind speeds monitored each time within the preset time period; when the number of wind speed data is greater than the first preset number, outputting a warning prompt for recommending shutting down the wind turbines in the wind farm.

[0055] Optionally, the providing of an early warning prompt based on the initial wind speed at the wind turbine hub acquired each time during monitoring within a preset time period includes:

[0056] When the initial wind speed at the wind turbine hub acquired each time during the preset time period is greater than the second preset wind speed, outputting a warning prompt for recommending shutting down the wind turbines in the wind farm; or,

[0057] Counting the number of wind speed data that are greater than the second preset wind speed among the initial wind turbine hub wind speeds monitored each time within the preset time period; when the number of wind speed data is greater than the second preset number, outputting a warning prompt for recommending shutting down the wind turbines in the wind farm.

[0058] The present disclosure also provides an electronic device, including a processor, a memory, and a program stored in the memory and executable on the processor, wherein the program, when executed by the processor, implements the steps of the wind power output prediction method described above.

[0059] The present disclosure also provides a computer non-transitory readable storage medium. When instructions in the storage medium are executed by a processor of an electronic device, the electronic device is enabled to execute the wind power output prediction method as described above.

[0060] The present disclosure further provides a wind power control system, comprising a plurality of data acquisition devices, a control device, and the electronic device described above, wherein the data acquisition device is arranged in a wind farm, the data acquisition device is communicatively connected to the control device, and the control device is communicatively connected to the electronic device;

[0061] The data acquisition device is configured to collect raw meteorological sub-data in the wind farm and transmit the raw meteorological sub-data to the control device;

[0062] The control device is configured to generate an initial meteorological data set based on each of the original meteorological sub-data, and transmit the initial meteorological data set to the electronic device at a preset time interval so that the electronic device can perform wind power output forecasting; each of the initial meteorological data sets corresponds to a receiving time node received by the electronic device.

[0063] The above description is only an overview of the technical solution of the present disclosure. In order to more clearly understand the technical means of the present disclosure, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present disclosure more obvious and easy to understand, the specific implementation methods of the present disclosure are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] In order to more clearly illustrate the technical solutions in the embodiments of the present disclosure or related technologies, the following is a brief introduction to the drawings required for use in the embodiments or related technical descriptions. Obviously, the drawings described below are some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0065] Figure 1 A flow chart showing the steps of a method for predicting wind power output according to an embodiment of the present disclosure is shown;

[0066] Figure 2 A flowchart showing the steps of another method for predicting wind power output according to an embodiment of the present disclosure is shown;

[0067] Figure 3 A flowchart showing the steps of training a wind power output prediction model according to an embodiment of the present disclosure is shown;

[0068] Figure 4 A flowchart showing another step of training a wind power output prediction model according to an embodiment of the present disclosure is shown;

[0069] Figure 5 A flowchart showing the steps of a wind farm early warning according to an embodiment of the present disclosure is shown;

[0070] Figure 6 A structural block diagram of a wind power control system according to an embodiment of the present disclosure is shown. Specific embodiments

[0071] To make the objectives, technical solutions, and advantages of the embodiments of the present disclosure more clear, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present disclosure, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present disclosure without making any creative efforts shall fall within the scope of protection of the present disclosure.

[0072] Unless otherwise defined, the technical or scientific terms used in this disclosure should have the usual meanings understood by people with ordinary skills in the field to which this disclosure belongs. The words "first", "second" and similar words used in this disclosure do not indicate any order, quantity or importance, but are only used to distinguish different components. Similarly, words such as "one", "an" or "the" do not indicate a quantity limitation, but rather indicate the presence of at least one. Words such as "include" or "comprise" mean that the elements or objects preceding the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. Words such as "connect" or "connected" are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. Directional words such as "upper", "lower", "left" and "right" are only used to indicate relative positional relationships based on the accompanying drawings. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0073] Figure 1 A flow chart showing the steps of a wind power output prediction method according to an embodiment of the present disclosure is shown. The method is used to predict the wind power output of a wind farm. Figure 1 , the method comprises the following steps:

[0074] Step 101: Periodically obtain an initial meteorological data set corresponding to each receiving time node; each initial meteorological data set includes initial meteorological data corresponding to at least one meteorological element, and the initial meteorological data includes initial meteorological sub-data of at least one dimension of the corresponding meteorological element.

[0075] In this step, the electronic device obtains an initial meteorological dataset from the wind farm at regular intervals. The reception time node corresponding to each initial meteorological dataset may be the time when the initial meteorological dataset is received by the electronic device. For each initial meteorological sub-data in the initial meteorological dataset, the electronic device may store the initial meteorological sub-data according to the corresponding relationship between the reception time node and the initial meteorological sub-data, as shown in Table 1 below.

[0076] Table 1

[0077]

[0078]

[0079] It should be understood that the data in Table 1 above is only an example and does not constitute a limitation to the present disclosure.

[0080] In specific applications, a wind farm may collect data for at least one meteorological element. Therefore, each initial meteorological data set may include multiple initial meteorological data sets, each corresponding to a meteorological element. In some optional embodiments, meteorological elements may include wind speed, gas density, air pressure, and temperature. Of course, other meteorological elements may also include wind direction, although this disclosure is not intended to limit these elements.

[0081] Furthermore, for each meteorological element, the wind farm can also collect data from at least one dimension. Different dimensions may specifically refer to different locations, different heights, and different objects.

[0082] Taking the meteorological element of wind speed as an example, wind speeds at different heights from the ground can be collected, such as the wind speed at 100 meters from the ground, the wind speed at 70 meters from the ground, and the wind speed at 30 meters from the ground. Wind speeds at different locations can also be collected, such as the wind speed between wind turbines and the wind speed at the wind turbine hub.

[0083] Taking the meteorological element gas density as an example, the density of different gas objects can be collected, such as air density.

[0084] Taking the meteorological element of air pressure as an example, the air pressure at different locations can be collected, such as surface pressure, sea level pressure, etc.

[0085] Taking the meteorological element of temperature as an example, the temperature at different heights from the ground can be collected, such as the temperature at 30 meters from the ground, the temperature at 2 meters from the ground, etc.

[0086] After the electronic equipment obtains these initial data from the wind farm, it can use these initial data as model input features to predict wind power output.

[0087] Step 102: After obtaining the latest initial meteorological data set corresponding to the latest receiving time node, identify abnormal sub-data from each latest initial meteorological sub-data.

[0088] In actual applications, the initial meteorological sub-data obtained may have abnormal conditions such as missing data and excessive deviation of values ​​from the normal range due to abnormalities in the data acquisition equipment. Therefore, in this step, whenever the electronic device obtains the latest initial meteorological data set corresponding to the latest receiving time node, the electronic device can identify the abnormal sub-data in the latest initial meteorological sub-data.

[0089] Step 103: When abnormal sub-data are identified, the identified abnormal sub-data are smoothed to obtain a smoothed meteorological data set. When no abnormal sub-data are identified, the latest initial meteorological data set is used as the smoothed meteorological data set.

[0090] In this step, when the electronic device identifies abnormal sub-data from each of the latest initial meteorological sub-data, it can smooth the identified abnormal sub-data, thereby preventing the abnormal sub-data from interfering with the prediction results and ensuring data integrity. After all abnormal sub-data in the latest initial meteorological dataset are processed, a smoothed meteorological dataset can be obtained.

[0091] When the electronic device does not identify abnormal sub-data from each latest initial meteorological sub-data, it means that each latest initial meteorological sub-data itself is relatively smooth data. Therefore, the electronic device can directly use the latest initial meteorological data set as the smoothed meteorological data set.

[0092] Step 104: Determine the instantaneous wind energy density corresponding to the latest receiving time node.

[0093] Wind energy density is the amount of wind energy flowing vertically through a unit area per unit time, measured in watts per square meter. It is the most convenient and valuable parameter for describing a location's wind energy potential and a key factor influencing wind power output. Therefore, in the disclosed embodiments, given its importance to wind power output, wind energy density can be used as a model input feature, thereby improving the accuracy of wind power output predictions.

[0094] After acquiring the latest initial meteorological data set each time, the electronic device may first determine the instantaneous wind energy density corresponding to the latest receiving time node.

[0095] Step 105: Calculate the rolling mean of the instantaneous wind energy density to obtain the average wind energy density within the target time period; the target time period includes the latest receiving time node.

[0096] The calculation of wind energy density requires wind speed data. However, due to the high randomness of wind speed, the wind energy potential of a wind farm cannot be accurately assessed using instantaneous wind energy density. Therefore, in this step, the electronic device can perform a rolling mean calculation on the instantaneous wind energy densities determined within the target time period to obtain the average wind energy density for the target time period. This average wind energy density can more accurately reflect the wind energy conditions of the wind farm over a period of time, thereby improving the accuracy of wind power output forecasts.

[0097] Step 106: The smoothed meteorological data set and the average wind energy density in the target time period are input as input features to the wind power output prediction model, so that the wind power output prediction model outputs the wind power output prediction value in the target time period.

[0098] In this step, a wind power output prediction model can be pre-deployed in the electronic device. After the electronic device obtains the smoothed meteorological data set and the average wind energy density within the target time period, these data can be used as input features of the model and input into the wind power output prediction model, so that the wind power output prediction model can output the wind power output prediction value for the target time period.

[0099] The electronic device can use the initial meteorological sub-data for each dimension of each meteorological element, as well as the average wind energy density, which has a significant impact on wind power output, as input features for the model. This allows the wind power output prediction model to output predicted wind power output values ​​based on a large number of features, thus enabling wind power output prediction. Furthermore, compared to methods based on time series models that predict wind power output using only a single feature—the historical wind power generation true value—the wind power output prediction method provided by the disclosed embodiments can achieve higher prediction accuracy.

[0100] In an embodiment of the present disclosure, an electronic device can periodically obtain an initial meteorological data set corresponding to each receiving time node, wherein the initial meteorological data set includes initial meteorological sub-data of at least one dimension of at least one meteorological element; after obtaining the latest initial meteorological data set, the abnormal sub-data is identified, and when identified, smoothing is performed to obtain a smoothed meteorological data set, and when not identified, the latest initial meteorological data set is used as the smoothed meteorological data set; then, the average wind energy density within the target time period is determined; and the smoothed meteorological data set and the average wind energy density within the target time period are used as input features of the model, and the wind power output prediction value for the target time period is obtained through the wind power output prediction model. In an embodiment of the present disclosure, the electronic device can use the initial meteorological sub-data of each dimension of each meteorological element, as well as the average wind energy density that has a greater impact on wind power output, as input features of the model, so that the wind power output prediction model can output the wind power output prediction value based on a large number of features, thereby realizing the prediction of wind power output and achieving a higher prediction accuracy.

[0101] Optionally, in some embodiments, step 101 may specifically include: periodically obtaining an initial meteorological data prediction set corresponding to each receiving time node; the initial meteorological data prediction set is predicted based on a historical initial meteorological data true value set corresponding to at least one historical receiving time node.

[0102] Among them, in some scenarios, wind farms can provide electronic equipment with meteorological data for a period of time in the future, that is, weather forecast data. The weather forecast data is a predicted value obtained by predicting historical meteorological data, rather than a true value. Then, electronic equipment can predict the wind power output for a period of time in the future (that is, the target time period) based on the weather forecast data for a period of time in the future.

[0103] For example, the current time is 7:50. Before 7:50, the electronic device had already obtained initial meteorological data forecast sets 1, 2, 3, and 4 at 7:00, 7:15, 7:30, and 7:45, respectively. This means the electronic device can obtain an initial meteorological data forecast set every 15 minutes. Initial meteorological data forecast sets 1, 2, 3, and 4 are the meteorological forecast data for 8:00, 8:15, 8:30, and 8:45, respectively. At this point, the electronic device can predict the wind power output for the target time period of 8:00-9:00 based on the meteorological forecast data for 8:00-9:00.

[0104] In other embodiments, step 101 may specifically include: periodically obtaining an initial meteorological data true value set corresponding to each receiving time node.

[0105] Among them, in other scenarios, the wind farm can provide electronic equipment with meteorological data for a period of time in the past, that is, historical meteorological data. The historical meteorological data is the true value of the meteorological data, not the predicted value. Then, the electronic equipment can predict the wind power output for a period of time in the future (that is, the target time period) based on the historical meteorological data for a period of time in the past.

[0106] For example, the current time is 7:50. Before 7:50, the electronic device had already obtained initial meteorological data truth value sets 1, 2, 3, and 4 at 7:00, 7:15, 7:30, and 7:45, respectively. This means that the electronic device can obtain an initial meteorological data truth value set every 15 minutes. Initial meteorological data prediction sets 1, 2, 3, and 4 can be the historical meteorological data corresponding to 7:00, 7:15, 7:30, and 7:45. In this case, the electronic device can predict the wind power output for the target time period of 8:00-9:00 based on the historical meteorological data from 7:00-8:00.

[0107] Optionally, in some embodiments, step 102 may specifically include:

[0108] S11: After obtaining the latest initial meteorological data set corresponding to the latest receiving time node, scrolling the second time window along the time axis so that the second time window includes the latest receiving time node;

[0109] S12: performing normal normalization processing on the initial meteorological sub-data of the same dimension and non-null values ​​belonging to the same meteorological element in the second time window to obtain the normal normalized value corresponding to each latest initial meteorological sub-data;

[0110] S13: The latest initial meteorological sub-data whose normalized values ​​are not within the preset numerical range is determined as abnormal sub-data.

[0111] The Z-score method can be used to calculate the mean and standard deviation of meteorological data that is close to a normal distribution and filter out data that exceeds multiple standard deviations. This reduces the data range and reduces the impact of non-null outliers on forecast accuracy. The Z-score method uses standard deviation as a unit to measure the distance between the original data and its mean.

[0112] Specifically, the electronic device can identify abnormal sub-data in a rolling manner. First, the second time window can be rolled along the time axis of the receiving time node so that the second time window includes the current latest receiving time node, wherein the last time node in the rolled second time window is the current latest receiving time node. Then, for the latest initial meteorological sub-data with non-null values ​​in each dimension, normal standardization processing can be performed, that is, Z-score calculation is performed to obtain the Z-score, which is the normal standardization value.

[0113] For the latest initial meteorological sub-data a of dimension A belonging to meteorological element Y, the Z score of a can be calculated by the formula Z = (x-μ) / σ, where x is the latest initial meteorological sub-data a, μ is the mean of multiple initial meteorological sub-data of dimension A belonging to meteorological element Y with the latest initial meteorological sub-data a, and σ is the standard deviation of multiple initial meteorological sub-data of dimension A belonging to meteorological element Y with the latest initial meteorological sub-data a.

[0114] Then, the latest initial meteorological sub-data whose corresponding Z-score is not within the preset value range can be determined as abnormal sub-data. For example, the preset value range can be [-3, 3], and the electronic device can filter out non-null value data that exceeds 3 times the standard deviation.

[0115] Further optionally, in some embodiments, after S11, step 102 may further include the following steps:

[0116] S14: Determine the null values ​​in each latest initial meteorological sub-data within the second time window as abnormal sub-data.

[0117] Among them, abnormal data includes not only non-null values ​​with relatively unreasonable values, but also null values. Electronic devices can also determine null values ​​as abnormal sub-data, so that both null values ​​and unreasonable non-null values ​​can be smoothed later instead of being deleted directly, so that the smoothed data can have temporal continuity, which is conducive to further improving the accuracy of the prediction.

[0118] In specific applications, for abnormal sub-data with non-null values, the electronic device can set them to null values ​​after identification. In this way, before subsequent smoothing processing, the abnormal sub-data in the latest initial meteorological data set are all null value data. In this way, when smoothing processing is performed later, you only need to pay attention to the null value identifier (such as the nan value) without paying attention to the specific row and column position of the abnormal sub-data, which improves the efficiency of the smoothing processing to a certain extent.

[0119] Optionally, in some embodiments, when abnormal sub-data are identified in step 103, the step of smoothing each of the identified abnormal sub-data to obtain a smoothed meteorological data set may specifically include:

[0120] S21: For any identified abnormal sub-data, scroll the third time window along the time axis so that the third time window includes the receiving time node corresponding to the abnormal sub-data and multiple receiving time nodes prior to the receiving time node corresponding to the abnormal sub-data;

[0121] S22: performing mean calculation on the initial meteorological sub-data of the same dimension and the same meteorological element as the abnormal sub-data within the third time window to obtain a new value corresponding to the abnormal sub-data;

[0122] S23: Replace each abnormal sub-data with the corresponding new value to obtain a smoothed meteorological data set.

[0123] Among them, the electronic device can perform smoothing of abnormal sub-data in a rolling manner. First, for dimension A of meteorological element Y, the third time window can be rolled along the time axis of the receiving time node, so that the third time window contains the first abnormal sub-data of dimension A belonging to meteorological element Y in the current latest initial meteorological data set, wherein the last time node in the rolled third time window is the receiving time node t1 corresponding to the first abnormal sub-data of dimension A belonging to meteorological element Y, and the rolled third time window also includes multiple receiving time nodes within a period of time before t1. Then, the mean calculation can be performed on each initial meteorological sub-data of dimension A belonging to meteorological element Y in the third time window to obtain a new value corresponding to the first abnormal sub-data of dimension A belonging to meteorological element Y. Afterwards, the first abnormal sub-data of dimension A belonging to meteorological element Y can be replaced with the corresponding new value.

[0124] Similarly, the third time window is scrolled again, and the above method can be repeated to determine the new values ​​corresponding to the second, third, ..., mth abnormal sub-data of dimension A belonging to meteorological element Y. Similarly, the new value of each abnormal sub-data of each dimension belonging to other meteorological elements is determined in the same way as the new value of each abnormal sub-data of dimension A belonging to meteorological element Y. This process continues until all abnormal sub-data of all dimensions of all meteorological elements are replaced with the corresponding new values, completing the smoothing process and obtaining a smoothed meteorological dataset.

[0125] For example, the electronic device can replace (or fill) the abnormal sub-data b with the initial meteorological sub-data within an hour before the abnormal sub-data b and belonging to the same meteorological element and the same dimension as the abnormal sub-data b. In this way, while ensuring the relative accuracy of the data, the continuity of the data in time series is also taken into account.

[0126] Optionally, in some embodiments, the smoothed meteorological sub-data in the smoothed meteorological data set include at least the smoothed wind speed at the wind turbine hub and the smoothed air density. Accordingly, step 104 may specifically include: determining the instantaneous wind energy density corresponding to the latest receiving time node based on the smoothed wind speed and smoothed air density at the wind turbine hub corresponding to the latest receiving time node.

[0127] If the latest wind turbine hub wind speed is an anomalous sub-data and has been smoothed, the smoothed wind turbine hub wind speed is the latest wind turbine hub wind speed after smoothing. If the latest wind turbine hub wind speed is not an anomalous sub-data, the smoothed wind turbine hub wind speed is the latest wind turbine hub wind speed. Similarly, if the latest air density is an anomalous sub-data and has been smoothed, the smoothed air density is the latest air density after smoothing. If the latest air density is not an anomalous sub-data, the smoothed air density is the latest air density.

[0128] Further optionally, in some embodiments, step 105 may specifically include:

[0129] S31: Scrolling the first time window along the time axis to align the first time window with the target time period;

[0130] S32: Calculate the average of multiple instantaneous wind energy densities within the first time window to obtain the average wind energy density within the target time period.

[0131] Generally, the wind energy density W within a period of time (t1-t2) can be calculated by the following formula (1), where ρ t is the air density corresponding to time t, V t is the wind speed at the hub of the wind turbine at time t.

[0132]

[0133] In the above formula (1), a period of time is used as the calculation standard. However, in actual applications, it is impossible to obtain the air density and wind speed at the fan hub at continuous moments. What is obtained is the air density and wind speed at the fan hub at discrete moments. Therefore, the embodiment of the present disclosure can determine the average wind energy density W within a period of time (target time period) through the following formula (2).

[0134]

[0135] In the above formula (2), n is the number of receiving time nodes included in the target time period, w i is the smooth instantaneous wind energy density corresponding to the receiving time node i, w i =0.5·ρ i ·(V i ) 3 ,ρ i is the smoothed air density corresponding to the receiving time node i, V i is the smoothed wind speed at the wind turbine hub corresponding to the receiving time node i.

[0136] In practical applications, the electronic device can first determine the smoothed instantaneous wind energy density w corresponding to each receiving time node i i The electronic device then sums and averages the instantaneous wind energy densities to obtain the average wind energy density for the target time period.

[0137] The electronic device can calculate the average wind energy density in a rolling manner. First, the first time window can be rolled along the time axis of the receiving time node to align the first time window with the target time period. Then, the average of the multiple instantaneous wind energy densities determined in the first time window can be calculated using the above formula (2) to obtain the average wind energy density in the target time period.

[0138] In addition, in practical applications, the rolling maximum and rolling minimum values ​​of the instantaneous wind energy density can also be calculated.

[0139] Optionally, in some embodiments, the wind power output prediction model may adopt a boosting model. In this case, the input features of the wind power output prediction model may only include a smoothed meteorological data set and an average wind energy density within a target time period.

[0140] Since the boosting model can focus more on multi-feature mining, after practical application, compared with using only the time series model, the use of the boosting model for wind power output forecasting can achieve higher prediction accuracy.

[0141] Alternatively, in other embodiments, the wind power output forecasting model may utilize a fusion model of a boosting model and a time series model. In this case, the input features of the wind power output forecasting model may include, in addition to the smoothed meteorological dataset and average wind energy density within the target time period, the historical wind power output true value corresponding to each receiving time node within the target time period.

[0142] Accordingly, refer to Figure 2 Before step 106, the following steps may also be included:

[0143] Step 107: Obtain the historical wind power output true value corresponding to each receiving time node within the target time period.

[0144] Accordingly, refer to Figure 2 , step 106 may specifically include:

[0145] Step 1061: The historical wind power output true values, smoothed meteorological data sets, and average wind energy density within the target time period are input as input features to the wind power output prediction model, so that the wind power output prediction model outputs the wind power output prediction value for the target time period.

[0146] Since the boosting model can focus more on multi-feature mining, and the time series model can focus more on the temporal relationship of features, after practical application, compared with using only the boosting model or only the time series model, the fusion model of the boosting model and the time series model for wind power output forecasting can achieve higher prediction accuracy.

[0147] Before step 101, the method may further include a model training process. Figure 3 ,For the case of adopting the boost model for the wind power output prediction model, the model training process may specifically include:

[0148] Step 201: Acquire multiple first sample sets; the first sample sets include smoothed meteorological data set samples corresponding to each historical receiving time node in the historical time period, average wind energy density samples in the historical time period, and wind power output true value samples in the historical time period.

[0149] In this step, the electronic device can obtain multiple first sample sets for training the boosting model. The first sample sets are labeled datasets, and the true wind power output samples for a historical time period are labels corresponding to a set of samples consisting of smoothed meteorological dataset samples and average wind energy density samples within the historical time period. Different first sample sets correspond to different historical time periods, and the historical time periods corresponding to different first sample sets can be continuous.

[0150] Step 202: Build a lifting model.

[0151] Among them, the boosting model is a model obtained by training using a boosting ensemble learning mechanism, such as a gradient boosting model. Furthermore, the gradient boosting model can adopt a GBDT (gradient boosting decision tree) model, and the GBDT can specifically adopt an XGBoost (eXtreme Gradient Boosting) model or a lightGBM (Light Gradient Boosting Machine) model, etc. The embodiments of the present disclosure do not specifically limit this.

[0152] In this step, a suitable gradient boosting model, such as the lightGBM model, can be selected according to the needs. In practical applications, the electronic device can obtain the boosting model from other platforms and configure it in the local model pool of the electronic device, and then select the boosting model from the model pool to complete the model construction. Of course, the boosting model can also be directly constructed locally on the electronic device and configured in the model pool, and then the boosting model can be selected from the model pool to complete the model construction.

[0153] Step 203: training the boosting model based on the multiple first sample sets to obtain a trained boosting model.

[0154] In this step, multiple first sample sets can be divided into training sets and test sets, and the training sets can be input into the improved model in sequence. After each input, the parameters in the improved model can be adjusted until all training sets are input. Then the test set is input into the improved model for model verification. When the verification result reaches a certain accuracy, the model parameter adjustment is completed and the trained improved model is obtained.

[0155] In practical applications, you can use automatic parameter tuning tools to tune the model parameters, such as the hyperopt tool.

[0156] Further optionally, step 203 may specifically include: training the boosting model by a cross-validation method according to the multiple first sample sets to obtain a trained boosting model.

[0157] In order to make the final model parameters not overly dependent on the division method of the training set and the test set, and to make full use of the existing first sample set, the cross-validation method can be used to train the boosting model so that each first sample set has a chance to be used as a test set alone, thereby optimizing the wind power output prediction model and further improving the prediction accuracy.

[0158] Step 204: Generate a wind power output prediction model based on the trained improved model.

[0159] In this step, the electronic device can deploy the trained enhanced model locally to obtain a usable wind power output prediction model.

[0160] Reference Figure 4 In the case where the wind power output prediction model adopts a fusion model of the boosting model and the time series model, before step 204, the model training process may further include:

[0161] Step 205: Acquire multiple second sample sets; the second sample sets include historical wind power output true value samples corresponding to each historical receiving time node within the historical time period.

[0162] In this step, the electronic device may obtain multiple second sample sets for training the time series model, wherein the second sample sets are unlabeled data sets. Different second sample sets correspond to different historical time periods, and the historical time periods corresponding to different second sample sets may be continuous.

[0163] Step 206: Build a time series model.

[0164] In this step, you can select an appropriate time series model based on your needs, such as the ARIMA model (Autoregressive Integrated Moving Average model), the SARIMA model (Seasonal Autoregressive Integrated Moving Average model), the LSTM model (Long short-term memory model), or variants of these models.

[0165] Step 207: Train the time series model based on the multiple second sample sets to obtain a trained time series model.

[0166] In this step, multiple second sample sets can be divided into training sets and test sets, and the training sets can be input into the time series model in sequence. After each input, the parameters in the time series model can be adjusted until all training sets are input. The test set is then input into the time series model for model verification. When the verification result reaches a certain accuracy, the model parameter adjustment is completed and the trained time series model is obtained.

[0167] It should be noted that the embodiment of the present disclosure does not limit the order of training the boosting model and the time series model. The boosting model can be trained first through steps 201-203, and then the time series model can be trained through steps 205-207. Alternatively, the time series model can be trained first through steps 205-207, and then the boosting model can be trained through steps 201-203.

[0168] Accordingly, in the case where the wind power output prediction model adopts a fusion model of the boosting model and the time series model, step 204 may specifically include:

[0169] Step 2041: stack and fuse the trained boosting model and the trained time series model to obtain a wind power output prediction model.

[0170] After separately training the trained boosted model and the trained time series model, the trained gradient boosted model and the trained time series model can be fused through stacking. The trained gradient boosted model and the trained time series model serve as base models, and a meta-model is trained through the stacking ensemble learning mechanism to combine these base models.

[0171] Optionally, the method may further include the following steps:

[0172] After obtaining the multiple new first sample sets and the multiple new second sample sets, the wind power output prediction model is retrained according to the multiple new first sample sets and the multiple new second sample sets to update the wind power output prediction model.

[0173] In actual applications, the situation of wind farms is not static, so the deployed wind power output prediction model may not be able to achieve a high prediction accuracy after a period of time. Therefore, after the wind power output prediction model is deployed, the electronic equipment can subsequently obtain or generate a lot of new data, and then use this data as a new first sample set and a new second sample set to retrain the old model and obtain a new model, thereby realizing the update of the model, so that the model can adapt to the changes in the wind farm, and thus maintain a high prediction accuracy most of the time.

[0174] Optionally, in some embodiments, the electronic device may also have an early warning mechanism. Specifically, the electronic device may determine whether the wind turbines in the current wind farm are suitable for continuing to work based on the data in the latest initial meteorological data set, and may issue an early warning prompt if it is not suitable for continuing to work.

[0175] Specifically, the latest initial meteorological sub-data includes at least the latest initial wind speed at the wind turbine hub. Figure 5, the method may further comprise the following steps:

[0176] Step 301: After obtaining the latest initial meteorological data set corresponding to the latest receiving time node, when the latest initial wind speed at the wind turbine hub is less than the first preset wind speed, or greater than the second preset wind speed, monitor the initial wind speed at the wind turbine hub obtained each time starting from the latest receiving time node; the first preset wind speed is less than the second preset wind speed.

[0177] Each time the electronic device obtains the latest initial meteorological data set, it can determine the initial wind speed at the wind turbine hub. If the initial wind speed at the wind turbine hub is less than a first preset wind speed, it indicates that the wind speed at the wind farm is currently very low. Considering losses in power transmission and other processes, the power generated by the wind is insufficient and the cost is high. If the initial wind speed at the wind turbine hub is greater than a second preset wind speed, it indicates that the wind speed at the wind farm is currently very high, which may damage the wind turbines and affect the stability of the power system.

[0178] However, due to the randomness of instantaneous wind speed, relying solely on the instantaneous wind speed at the wind turbine hub will result in frequent warnings, and manual judgment is required as to whether the wind turbine is truly unsuitable for continued operation, rendering the warning mechanism ineffective. Therefore, in the disclosed embodiment, when the latest initial wind speed at the wind turbine hub is less than a first preset wind speed, or greater than a second preset wind speed, the electronic device may begin monitoring the initial wind speed at the wind turbine hub acquired over a period of time, and then determine whether a warning is necessary based on the initial wind speed at the wind turbine hub acquired over a period of time.

[0179] Step 302: issuing an early warning based on the initial wind speed at the wind turbine hub acquired each time during the preset monitoring period.

[0180] Electronic equipment can provide early warnings when wind turbine operating costs are high, and can also provide early warnings when the power system is unstable.

[0181] In the case of issuing an early warning when the wind turbine operating cost is high, step 302 may specifically include:

[0182] S41: When the initial wind speed at the hub of the wind turbine monitored each time within a preset time period is less than a first preset wind speed, outputting an early warning prompt for recommending shutting down the wind turbines in the wind farm; or

[0183] S42: Counting the number of wind speed data that are less than a first preset wind speed in the initial wind turbine hub wind speeds acquired each time during a preset time period; when the number of wind speed data is greater than the first preset number, outputting an early warning prompt for recommending shutting down the wind turbines in the wind farm.

[0184] In an optional embodiment, after monitoring the initial wind speed at the wind turbine hub for a preset period of time, if the wind speed at each initial wind turbine hub within the predicted period is less than the first preset wind speed, an early warning prompt can be output to advise relevant personnel to shut down the wind turbines in the wind farm to reduce the operating costs of the wind turbines.

[0185] In another optional embodiment, after monitoring the initial wind speed at the wind turbine hub for a preset period of time, when the initial wind speed at the wind turbine hub that is less than the first preset wind speed reaches a large amount of data within the predicted period, an early warning prompt can be output to advise relevant personnel to shut down the wind turbines in the wind farm to reduce the operating costs of the wind turbines.

[0186] In the case of providing an early warning when the power system is unstable, step 302 may specifically include:

[0187] S51: When the initial wind speed at the hub of the wind turbine monitored each time within the preset time period is greater than the second preset wind speed, outputting an early warning prompt for recommending shutting down the wind turbines in the wind farm; or

[0188] S52: Counting the number of wind speed data that are greater than a second preset wind speed among the initial wind speed at the wind turbine hub acquired each time within a preset time period; when the number of wind speed data is greater than the second preset number, outputting a warning prompt for recommending shutting down the wind turbines in the wind farm.

[0189] In an optional embodiment, after monitoring the initial wind speed at the wind turbine hub for a preset period of time, if the wind speed at each initial wind turbine hub within the predicted period is greater than a second preset wind speed, an early warning prompt can be output to advise relevant personnel to shut down the wind turbines in the wind farm to avoid instability of the power system due to damage to the wind turbines.

[0190] In another optional embodiment, after monitoring the initial wind speed at the wind turbine hub for a preset period of time, when the initial wind speed at the wind turbine hub that is less than the first preset wind speed reaches a large amount of data within the predicted period, an early warning prompt can be output to advise relevant personnel to shut down the wind turbines in the wind farm to avoid instability of the power system due to damage to the wind turbines.

[0191] Optionally, the warning prompt may be provided by playing specific audio (such as a special warning sound) through an audio playback device such as a speaker, or generating a warning light through a lighting device, etc., which is not specifically limited in the embodiment of the present disclosure.

[0192] In addition, in actual applications, the first preset wind speed and the second preset wind speed can be set in combination with data such as the power generation cost of a specific wind farm and historical damage status of wind turbines.

[0193] The embodiment of the present disclosure also discloses an electronic device, including a processor, a memory, and a program stored in the memory and executable on the processor, wherein the program, when executed by the processor, implements the steps of the wind power output prediction method described above.

[0194] The embodiment of the present disclosure further discloses a computer non-transitory readable storage medium. When instructions in the storage medium are executed by a processor of an electronic device, the electronic device is able to execute the wind power output prediction method as described above.

[0195] Reference Figure 6 The present disclosure also discloses a wind power control system 1000, including a plurality of data acquisition devices 100, a control device 200, and the electronic device 300 described above. The data acquisition devices 100 are arranged in a wind farm Q, which includes a plurality of wind turbines q. The data acquisition devices 100 are communicatively connected to the control devices 200, and the control devices 200 are communicatively connected to the electronic device 300.

[0196] The data acquisition device 100 is configured to collect raw meteorological sub-data in the wind farm Q and transmit the raw meteorological sub-data to the control device 200;

[0197] The control device 200 is configured to generate an initial meteorological data set based on each of the original meteorological sub-data, and transmit the initial meteorological data set to the electronic device 300 at a preset time interval so that the electronic device 300 can perform wind power output forecasting; each of the initial meteorological data sets corresponds to a receiving time node received by the electronic device 300.

[0198] Among them, the control device can be a main control device corresponding to the data acquisition device. The control device can perform preliminary processing on each original meteorological sub-data to obtain initial meteorological sub-data, and then obtain an initial meteorological data set. Among them, the original meteorological sub-data of each dimension of each meteorological data can be collected by at least one data acquisition device. For example, the dimension of the meteorological element wind speed at 30 meters from the ground surface can be obtained by multiple wind speed acquisition devices at 30 meters from the ground surface to obtain multiple original wind speeds at 30 meters from the ground surface (that is, original meteorological sub-data). Then, the control device can calculate the average of the multiple original wind speeds at 30 meters from the ground surface to obtain the wind speed at 30 meters from the ground surface (that is, initial meteorological sub-data) to be sent to the electronic device.

[0199] Of course, in practical applications, the preliminary processing includes but is not limited to mean calculation.

[0200] Furthermore, in specific applications, the control device may provide an initial meteorological data set for more than one electronic device.

[0201] References herein to "one embodiment," "an embodiment," or "one or more embodiments" mean that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the present disclosure. Furthermore, please note that instances of the phrase "in one embodiment" do not necessarily all refer to the same embodiment.

[0202] In the description provided herein, numerous specific details are described. However, it is understood that embodiments of the present disclosure may be practiced without these specific details. In some instances, well-known methods, structures, and techniques are not shown in detail so as not to obscure the understanding of this description.

[0203] In the claims, any reference signs placed between parentheses shall not be construed as limiting the claim. The word "comprising" does not exclude the presence of elements or steps not listed in the claim. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The present disclosure may be implemented by means of hardware comprising several different elements and by means of a suitably programmed computer. In a unit claim enumerating several means, several of these means may be embodied by the same item of hardware. The use of the words first, second, and third etc. does not indicate any order. These words may be interpreted as names.

[0204] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present disclosure, rather than to limit them. Although the present disclosure has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present disclosure.

Claims

1. A wind power output prediction method, characterized in that: The method comprises: Periodically acquiring an initial meteorological data set corresponding to each receiving time node; each of the initial meteorological data sets includes initial meteorological data corresponding to at least one meteorological element, and the initial meteorological data includes initial meteorological sub-data of at least one dimension of the corresponding meteorological element; After obtaining the latest initial meteorological data set corresponding to the latest receiving time node, identifying abnormal sub-data from each latest initial meteorological sub-data; When abnormal sub-data are identified, smoothing is performed on the identified abnormal sub-data to obtain a smoothed meteorological data set; when no abnormal sub-data are identified, the latest initial meteorological data set is used as a smoothed meteorological data set; Determining the instantaneous wind energy density corresponding to the latest receiving time node; Performing a rolling mean calculation on the instantaneous wind energy density to obtain an average wind energy density within a target time period; the target time period includes the latest receiving time node; The smoothed meteorological data set and the average wind energy density within the target time period are used as input features and input into a wind power output prediction model, so that the wind power output prediction model outputs a wind power output prediction value for the target time period.

2. The method according to claim 1, characterized in that The step of performing a rolling mean calculation on the instantaneous wind energy density to obtain an average wind energy density within a target time period includes: Scrolling the first time window along the time axis to align the first time window with the target time period; An average of the multiple instantaneous wind energy densities within the first time window is calculated to obtain an average wind energy density within the target time period.

3. The method according to claim 1, characterized in that The smoothed meteorological sub-data in the smoothed meteorological data set includes at least smoothed wind speed and smoothed air density at the wind turbine hub, and determining the instantaneous wind energy density corresponding to the latest receiving time node includes: The instantaneous wind energy density corresponding to the latest receiving time node is determined according to the smoothed wind speed at the wind turbine hub and the smoothed air density corresponding to the latest receiving time node.

4. The method according to claim 1, wherein After obtaining the latest initial meteorological data set corresponding to the latest receiving time node, identifying abnormal sub-data from each latest initial meteorological sub-data includes: After obtaining the latest initial meteorological data set corresponding to the latest receiving time node, scrolling the second time window along the time axis so that the second time window includes the latest receiving time node; Performing normalization processing on the initial meteorological sub-data of the same dimension and non-null values ​​belonging to the same meteorological element within the second time window to obtain a normalized normalized value corresponding to each latest initial meteorological sub-data; The latest initial meteorological sub-data whose normalized value is not within a preset numerical range is determined as the abnormal sub-data.

5. The method according to claim 4, characterized in that After acquiring the latest initial meteorological data set corresponding to the latest receiving time node, scrolling the second time window along the time axis so that the second time window includes the latest receiving time node, further comprising: The null value in each of the latest initial meteorological sub-data within the second time window is determined as the abnormal sub-data.

6. The method according to claim 1, characterized in that When abnormal sub-data are identified, smoothing is performed on each of the identified abnormal sub-data to obtain a smoothed meteorological data set, including: For any of the identified abnormal sub-data, scroll a third time window along the time axis so that the third time window includes the receiving time node corresponding to the abnormal sub-data and multiple receiving time nodes prior to the receiving time node corresponding to the abnormal sub-data; performing mean calculation on the initial meteorological sub-data belonging to the same meteorological element and the same dimension as the abnormal sub-data within the third time window to obtain a new value corresponding to the abnormal sub-data; Each abnormal sub-data is replaced with the corresponding new value to obtain a smoothed meteorological data set.

7. The method according to claim 1, characterized in that The periodic acquisition of the initial meteorological data set corresponding to each receiving time node includes: An initial meteorological data prediction set corresponding to each receiving time node is periodically obtained; the initial meteorological data prediction set is predicted based on a historical initial meteorological data true value set corresponding to at least one historical receiving time node.

8. The method according to claim 1, characterized in that The periodic acquisition of the initial meteorological data set corresponding to each receiving time node includes: Periodically obtain the initial meteorological data truth value set corresponding to each receiving time node.

9. The method according to claim 1, characterized in that Before periodically acquiring the initial meteorological data set corresponding to each receiving time node, the method further includes: Acquire multiple first sample sets; the first sample sets include smoothed meteorological data set samples corresponding to each historical receiving time node in a historical time period, average wind energy density samples in the historical time period, and wind power output true value samples in the historical time period; Build an improvement model; Training the boosting model according to the plurality of first sample sets to obtain a trained boosting model; The wind power output prediction model is generated based on the trained improved model.

10. The method according to claim 9, characterized in that The step of training the boost model according to the plurality of first sample sets to obtain a trained boost model includes: The boosting model is trained according to the plurality of first sample sets by a cross-validation method to obtain a trained boosting model.

11. The method according to claim 9 or 10, characterized in that Before inputting the smoothed meteorological data set and the average wind energy density within the target time period as input features into a wind power output prediction model so that the wind power output prediction model outputs a wind power output prediction value for the target time period, the method further includes: Obtaining a historical wind power output true value corresponding to each receiving time node within the target time period; The step of inputting the smoothed meteorological data set and the average wind energy density within the target time period as input features into a wind power output prediction model so that the wind power output prediction model outputs a predicted wind power output value for the target time period, including: The true values ​​of the historical wind power outputs within the target time period, the smoothed meteorological data set, and the average wind energy density are input as input features into a wind power output prediction model, so that the wind power output prediction model outputs a predicted wind power output value for the target time period.

12. The method according to claim 11, characterized in that Before generating the wind power output prediction model according to the trained improved model, the method further includes: Acquire multiple second sample sets; the second sample sets include historical wind power output true value samples corresponding to each of the historical receiving time nodes within the historical time period; Build time series models; Training the time series model according to the plurality of second sample sets to obtain a trained time series model; Generating the wind power output prediction model according to the trained improved model includes: The trained boosting model and the trained time series model are stacked and fused to obtain the wind power output prediction model.

13. The method according to claim 12, characterized in that The method further comprises: After obtaining multiple new first sample sets and multiple new second sample sets, the wind power output prediction model is retrained according to the multiple new first sample sets and the multiple new second sample sets to update the wind power output prediction model.

14. The method according to claim 1, wherein The meteorological elements include wind speed, gas density, air pressure and temperature.

15. The method according to claim 1, wherein The latest initial meteorological sub-data includes at least the latest initial wind speed at the wind turbine hub, and the method further includes: After obtaining the latest initial meteorological data set corresponding to the latest receiving time node, when the latest initial wind speed at the wind turbine hub is less than a first preset wind speed, or greater than a second preset wind speed, monitoring the initial wind speed at the wind turbine hub obtained each time from the latest receiving time node; the first preset wind speed is less than the second preset wind speed; An early warning is issued based on the initial wind speed at the wind turbine hub obtained each time during monitoring within a preset time period.

16. The method according to claim 15, characterized in that The step of providing an early warning based on the initial wind speed at the wind turbine hub acquired each time during the preset monitoring period includes: When the initial wind speed at the wind turbine hub acquired each time during the preset time period is less than the first preset wind speed, outputting a warning prompt for recommending shutting down the wind turbines in the wind farm; or Counting the number of wind speed data that are less than the first preset wind speed among the initial wind turbine hub wind speeds monitored each time within the preset time period; when the number of wind speed data is greater than the first preset number, outputting a warning prompt for recommending shutting down the wind turbines in the wind farm.

17. The method according to claim 15, characterized in that The step of providing an early warning based on the initial wind speed at the wind turbine hub acquired each time during the preset monitoring period includes: When the initial wind speed at the wind turbine hub acquired each time during the preset time period is greater than the second preset wind speed, outputting a warning prompt for recommending shutting down the wind turbines in the wind farm; or, Counting the number of wind speed data that are greater than the second preset wind speed among the initial wind turbine hub wind speeds monitored each time within the preset time period; when the number of wind speed data is greater than the second preset number, outputting a warning prompt for recommending shutting down the wind turbines in the wind farm.

18. An electronic device, characterized in that: The invention comprises a processor, a memory and a program stored in the memory and executable on the processor, wherein when the program is executed by the processor, the steps of the wind power output prediction method according to any one of claims 1 to 17 are implemented.

19. A computer non-transitory readable storage medium, characterized in that: When the instructions in the storage medium are executed by a processor of an electronic device, the electronic device is enabled to execute the wind power output prediction method according to any one of claims 1 to 17.

20. A wind power control system, characterized in that: The device comprises a plurality of data acquisition devices, a control device, and the electronic device according to claim 18, wherein the data acquisition device is arranged in a wind farm, the data acquisition device is communicatively connected to the control device, and the control device is communicatively connected to the electronic device; The data acquisition device is configured to collect raw meteorological sub-data in the wind farm and transmit the raw meteorological sub-data to the control device; The control device is configured to generate an initial meteorological data set based on each of the original meteorological sub-data, and transmit the initial meteorological data set to the electronic device at a preset time interval, so that the electronic device performs wind power output prediction; Each of the initial meteorological data sets corresponds to a receiving time node at which the data is received by the electronic device.

Images