Power prediction method and device, electronic equipment and storage medium

Abstract

The invention discloses a power prediction method and device, electronic equipment and a storage medium, and relates to the technical field of power prediction. The method comprises the following steps: acquiring historical meteorological data of a wind power plant and historical operation data of each fan in the wind power plant; performing space mapping processing on the historical meteorological data and the historical operation data of each fan to obtain fan meteorological fusion data; based on the fan meteorological fusion data, predicting initial prediction power and prediction wind speed of each fan at each time point in a preset future duration; and based on the predicted meteorological data of the wind power plant in the preset future duration and the initial predicted power and the predicted wind speed of each fan, predicting the target predicted power of each fan at each time point in the preset future duration. According to the technical scheme of the embodiment of the invention, the power prediction accuracy of each fan in the wind power plant is improved.

Description

Technical Field

[0001] The present invention relates to the field of power prediction technology, and in particular to a power prediction method, apparatus, electronic device and storage medium. Background Technology

[0002] With the rapid development of the wind power industry, the proportion of wind power generation in the energy structure continues to increase. The refined operation and efficient dispatch of wind farms are crucial for the stable operation of the power system. Therefore, high-precision wind farm power forecasting is essential for ensuring grid security, optimizing power dispatch, and improving wind power absorption capacity.

[0003] However, most existing power prediction methods utilize historical meteorological data and historical power data of wind turbines in wind farms to construct statistical models for prediction using machine learning algorithms. However, this method has a single data dimension and is difficult to fully reflect the operating status of complex power systems, thus affecting the accuracy of prediction.

[0004] Therefore, there is an urgent need to propose a new method to solve the above problems. Summary of the Invention

[0005] This invention provides a power prediction method, apparatus, electronic device, and storage medium to improve the accuracy of power prediction for each wind turbine in a wind farm.

[0006] In a first aspect, embodiments of the present invention provide a power prediction method, the method comprising:

[0007] Acquire historical meteorological data of the wind farm and historical operating data of each wind turbine in the wind farm;

[0008] Spatial mapping processing is performed on the historical meteorological data and the historical operating data of each wind turbine to obtain wind turbine meteorological fusion data;

[0009] Based on the aforementioned wind turbine meteorological fusion data, the initial predicted power and predicted wind speed of each wind turbine at each time point within a preset future time period are predicted.

[0010] Based on the predicted meteorological data of the wind farm within the preset future time period, the initial predicted power and predicted wind speed of each wind turbine, the target predicted power of each wind turbine at each time point within the preset future time period is predicted.

[0011] The technical solution of this invention first acquires historical meteorological data of the wind farm and historical operating data of each wind turbine within the wind farm, providing a data foundation for subsequent wind turbine meteorological fusion data. Next, spatial mapping processing is performed on the historical meteorological data and the historical operating data of each wind turbine to obtain wind turbine meteorological fusion data, generating a highly unified spatiotemporal dimensional fusion dataset, thereby improving the accuracy and robustness of subsequent predictions. Then, based on the wind turbine meteorological fusion data, the initial predicted power and predicted wind speed of each wind turbine at each time point within a preset future timeframe are predicted, improving the accuracy of the initial prediction results and providing a high-quality data foundation for subsequent secondary predictions. This reduces the correction burden in the optimization process and effectively improves the accuracy of the target predicted power. Finally, based on the predicted meteorological data of the wind farm within a preset future timeframe, the initial predicted power and predicted wind speed of each wind turbine, the target predicted power of each wind turbine at each time point within the preset future timeframe is predicted. This not only reduces the error of prediction driven by a single data source but also effectively corrects the initial prediction deviation, significantly improving the power prediction accuracy. Therefore, the technical solution of the present invention solves the problem in the prior art that the reliance on single-dimensional data makes it difficult to fully reflect the operating status of complex power systems, thus leading to insufficient power prediction accuracy.

[0012] Secondly, embodiments of the present invention also provide a power prediction device, the device comprising:

[0013] The acquisition module is used to acquire historical meteorological data of the wind farm and historical operating data of each wind turbine in the wind farm;

[0014] The fusion module is used to perform spatial mapping processing on the historical meteorological data and the historical operating data of each wind turbine to obtain wind turbine meteorological fusion data.

[0015] The first prediction module is used to predict the initial predicted power and predicted wind speed of each wind turbine at each time point within a preset future time period based on the wind turbine meteorological fusion data.

[0016] The second prediction module is used to predict the target predicted power of each wind turbine at each time point within the preset future time period based on the predicted meteorological data of the wind farm within the preset future time period, the initial predicted power of each wind turbine, and the predicted wind speed.

[0017] Thirdly, embodiments of the present invention also provide an electronic device, the electronic device comprising:

[0018] At least one processor; and a memory communicatively connected to said at least one processor;

[0019] The memory stores a computer program that can be executed by the at least one processor to enable the at least one processor to implement any of the power prediction methods described in the first aspect.

[0020] Fourthly, embodiments of the present invention also provide a storage medium containing computer-executable instructions, which, when executed by a computer processor, implement any of the power prediction methods described in the first aspect.

[0021] It should be noted that the aforementioned computer instructions may be stored, in whole or in part, on a computer-readable storage medium. This computer-readable storage medium may be packaged together with the processor of the power prediction device, or it may be packaged separately from the processor of the power prediction device; this application does not impose any limitations on this.

[0022] The descriptions of the second, third, and fourth aspects in this application can be referenced to the detailed description of the first aspect; and the beneficial effects described in the second, third, and fourth aspects can be referenced to the analysis of the beneficial effects of the first aspect, which will not be repeated here.

[0023] In this application, the name of the aforementioned power prediction device does not limit the device or functional module itself. In actual implementation, these devices or functional modules may appear under other names. As long as the function of each device or functional module is similar to that of this application, it falls within the scope of the claims of this application and its equivalents.

[0024] These or other aspects of this application will become more readily apparent in the following description. Attached Figure Description

[0025] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0026] Figure 1 A flowchart of a power prediction method provided in an embodiment of the present invention;

[0027] Figure 2 A flowchart of another power prediction method provided in an embodiment of the present invention;

[0028] Figure 3 This is a schematic diagram of the structure of a power prediction device provided in an embodiment of the present invention;

[0029] Figure 4This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0030] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and not intended to limit it. Furthermore, it should be noted that, for ease of description, the accompanying drawings show only the parts relevant to the present invention, and not all of the structures.

[0031] In this article, the term "and / or" is merely a description of the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can represent three situations: A exists alone, A and B exist simultaneously, and B exists alone.

[0032] The terms "first" and "second," etc., used in the specification and drawings of this application are used to distinguish different objects or to distinguish different treatments of the same object, rather than to describe a specific order of objects.

[0033] Furthermore, the terms "comprising" and "having," and any variations thereof, used in the description of this application are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the steps or units listed, but may optionally include other steps or units not listed, or may optionally include other steps or units inherent to such process, method, product, or apparatus.

[0034] Before discussing the exemplary embodiments in more detail, it should be noted that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although the flowcharts describe the operations (or steps) as sequential processes, many of these operations can be performed in parallel, concurrently, or simultaneously. Furthermore, the order of the operations can be rearranged. The process can be terminated when its operation is completed, but may also have additional steps not included in the figures. The process can correspond to a method, function, procedure, subroutine, subroutine, etc. Moreover, without conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0035] It should be noted that in the embodiments of this application, the words "exemplary" or "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design scheme described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design schemes. Specifically, the use of the words "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0036] In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0037] Figure 1 This is a flowchart illustrating a power prediction method provided in an embodiment of the present invention. This embodiment is applicable to situations requiring power prediction of each wind turbine in a wind farm. The method can be executed by a power prediction device, which can be implemented in software and / or hardware. For example, the device can be an electronic device. (See reference...) Figure 1 The power prediction method in this embodiment specifically includes the following steps:

[0038] Step 110: Obtain historical meteorological data of the wind farm and historical operating data of each wind turbine in the wind farm.

[0039] Specifically, a wind farm refers to a large-scale power generation site consisting of a batch of wind turbine generators (wind turbines), supporting power collection systems (such as collection lines and step-up substations), and auxiliary facilities (such as maintenance roads and monitoring centers), used to centrally develop wind energy resources and convert them into electrical energy. A wind turbine is the core equipment in a wind farm that directly converts wind energy into electrical energy. Historical meteorological data for a wind farm refers to the collection of meteorological information for the area where the wind farm is located from a certain point in the past to the present (e.g., the past 1 hour, 3 hours, or 1 day). For example, historical meteorological data may include wind speed, wind direction, temperature, humidity, and air pressure. In this embodiment, the format of historical meteorological data can be gridded meteorological data. Historical operating data for each wind turbine refers to the operating status information recorded by its own sensors and control system during the operation of each wind turbine in the wind farm from a certain point in the past to the present (e.g., the past 1 hour, 3 hours, or 1 day). For example, historical operating data may include nacelle position, turbine speed, blade angle, gearbox temperature, and output power.

[0040] In practice, based on a preset period (such as 1 hour) and the latitude and longitude range of the wind farm, historical meteorological data from a third-party meteorological service platform can be retrieved to obtain the initial historical meteorological data of the wind farm. Then, the data can be preprocessed (such as converting it to a compatible format, removing outliers, and supplementing missing values) to obtain the historical meteorological data of the wind farm.

[0041] Similarly, based on the aforementioned preset period, the operating data of each wind turbine can be obtained from the database of the data acquisition and monitoring system for each wind turbine in the wind farm. To avoid data confusion, a unique identifier (such as turbine number, installation location coordinates, unit model, etc.) can be added to each piece of operating data to obtain the initial historical operating data of each wind turbine. This data can then be preprocessed to obtain the historical operating data of each wind turbine.

[0042] In this embodiment, the above steps provide a data foundation for obtaining wind turbine meteorological fusion data in the future.

[0043] Step 120: Perform spatial mapping processing on historical meteorological data and historical operating data of each wind turbine to obtain wind turbine meteorological fusion data.

[0044] Specifically, wind turbine meteorological fusion data refers to a structured dataset formed by integrating the historical operating data of each wind turbine in a wind farm with the historical meteorological data of the corresponding spatial location.

[0045] In the specific implementation, firstly, a regular two-dimensional spatial grid is generated based on the latitude and longitude boundaries of the wind farm [e.g., 116°10′E - 116°20′E, 39°20′N - 39°30′N] and a preset grid resolution (e.g., 500m × 500m, 1km × 1km). It should be noted that a unified coordinate system (e.g., WGS84 coordinate system) must be used when establishing the grid to ensure consistency with subsequent meteorological data and wind turbine location coordinates; at the same time, a unique identifier is assigned to each grid cell [e.g., "row number - column number", for example: grid (2,3)], and the latitude and longitude of its center point are recorded (e.g., the center point of grid (2,3) is 39°22′30″N, 116°13′01″E), providing a basis for subsequent distance calculation and data binding.

[0046] Subsequently, an appropriate interpolation method is selected based on the terrain characteristics of the wind farm. For example, if the wind farm has flat terrain, the inverse distance weighted interpolation method is used; if the wind farm has complex terrain, the Kriging interpolation method is used. Then, the selected interpolation method is combined with the location information of historical meteorological data to interpolate the historical meteorological data to each grid cell, resulting in a two-dimensional meteorological spatial feature matrix with "grid identifier - meteorological data - acquisition time" as the core dimension.

[0047] Next, based on the installation location of each wind turbine (such as latitude and longitude coordinates), the distance between each wind turbine and the center of all grid cells in the meteorological two-dimensional spatial feature matrix is ​​calculated using a distance calculation formula (such as the Haversian formula). Then, following the principle of "nearest matching", the nearest grid cell is matched for each wind turbine, and its historical operating data (such as rotational speed, output power, and blade angle) is fused with the meteorological data of the corresponding grid cell (such as the interpolated wind speed or wind speed and temperature of that cell). Specifically, the time granularity (such as a 15-minute interval) and timestamp format (such as YYYY-MM-DDHH:MM:SS) of the two can be unified first to ensure that the wind turbine operating data and meteorological data at the same time can be accurately aligned. Subsequently, the aligned two types of data are spliced ​​according to the timestamp, and finally all splicing results are summarized to obtain the wind turbine meteorological fusion data.

[0048] In this embodiment, the above steps generate a fusion dataset with highly unified spatiotemporal dimensions, thereby improving the accuracy and robustness of subsequent predictions.

[0049] Step 130: Based on the wind turbine meteorological fusion data, predict the initial predicted power and predicted wind speed of each wind turbine at each time point within a preset future time period.

[0050] Specifically, the preset future duration refers to a predicted time range pre-set based on actual conditions or needs, such as the next hour, two hours, one day, or one week. Each time point refers to a specific moment after the preset future duration is divided into fixed time intervals; it is the time dimension carrier of the prediction results. For example, if the preset future duration is one hour and the time interval is 10 minutes, then each time point is the next 0 hours (the current moment), the next 10 minutes, 20 minutes… the next hour. Each time point corresponds to a set of independent prediction results (such as the initial predicted power and predicted wind speed of a wind turbine in the next 10 minutes). The initial predicted power refers to the power value that the wind turbine may output at a certain future time point, calculated based on the wind turbine's meteorological fusion data. The predicted wind speed refers to the expected wind speed value at the location of the wind turbine at a certain future time point, calculated based on the wind turbine's meteorological fusion data.

[0051] In practice, the fused meteorological data of the wind turbines is input into a pre-trained primary prediction model to obtain the initial predicted power and predicted wind speed of each wind turbine at various time points within a preset future timeframe. The primary prediction model refers to a model obtained by training a deep learning model (such as Transformer) using historical fused meteorological data of different wind turbines as input features and their corresponding future actual power and wind speed as training labels. To improve model quality and prediction accuracy, a multi-task learning mechanism can be employed to simultaneously optimize both wind speed and power prediction tasks.

[0052] In this embodiment, the above steps improve the accuracy of the initial prediction results, providing a high-quality data foundation for subsequent secondary predictions, thereby reducing the correction burden of the optimization process and effectively improving the accuracy of the target prediction power.

[0053] Step 140: Based on the predicted meteorological data of the wind farm within a preset future time period, the initial predicted power of each wind turbine and the predicted wind speed, predict the target predicted power of each wind turbine at each time point within the preset future time period.

[0054] Specifically, forecast meteorological data refers to the expected meteorological data for a predetermined future time period obtained through professional meteorological forecasting systems (such as numerical weather prediction models or third-party meteorological service platforms). Target forecast power refers to the expected final power generation value of wind turbines at each time point within a predetermined future time period, obtained by correcting and optimizing the initial forecast power based on forecast meteorological data.

[0055] In practice, the predicted meteorological data of the wind farm, along with the initial predicted power and predicted wind speed of each wind turbine, can be input into a pre-trained advanced prediction model to obtain the target predicted power of each wind turbine at each time point within a preset future timeframe. The advanced prediction model refers to the model obtained by training a deep learning model (such as Transformer) using the output of the primary prediction model and the predicted meteorological data for future times as input features, and the actual power at the corresponding future times as training labels.

[0056] In this embodiment, the above steps not only reduce the error of single data-driven prediction, but also effectively correct the initial prediction deviation and significantly improve the power prediction accuracy.

[0057] The power prediction method provided in this invention first acquires historical meteorological data of the wind farm and historical operating data of each wind turbine within the wind farm, providing a data foundation for subsequent wind turbine meteorological fusion data. Next, spatial mapping processing is performed on the historical meteorological data and the historical operating data of each wind turbine to obtain wind turbine meteorological fusion data, generating a highly unified spatiotemporal dimensional fusion dataset, thereby improving the accuracy and robustness of subsequent predictions. Then, based on the wind turbine meteorological fusion data, the initial predicted power and predicted wind speed of each wind turbine at each time point within a preset future timeframe are predicted, improving the accuracy of the initial prediction results and providing a high-quality data foundation for subsequent secondary predictions. This reduces the correction burden in the optimization process and effectively improves the accuracy of the target predicted power. Finally, based on the predicted meteorological data of the wind farm within a preset future timeframe, the initial predicted power and predicted wind speed of each wind turbine, the target predicted power of each wind turbine at each time point within the preset future timeframe is predicted. This not only reduces the error of prediction driven by a single data source but also effectively corrects the initial prediction deviation, significantly improving the power prediction accuracy. Therefore, the technical solution of the present invention solves the problem in the prior art that the reliance on single-dimensional data makes it difficult to fully reflect the operating status of complex power systems, thus leading to insufficient power prediction accuracy.

[0058] Figure 2 A flowchart of another power prediction method provided by an embodiment of the present invention is shown. This embodiment is a specific embodiment based on the above embodiment. In this embodiment, the method may further include:

[0059] Step 210: Obtain historical meteorological data of the wind farm and historical operating data of each wind turbine in the wind farm.

[0060] Furthermore, after step 210, the method further includes: performing outlier processing on the historical meteorological data and the historical operating data of each wind turbine to obtain intermediate historical meteorological data and intermediate historical operating data of each wind turbine; performing time alignment processing on the intermediate historical meteorological data and the intermediate historical operating data of each wind turbine to obtain updated historical meteorological data and updated historical operating data of each wind turbine.

[0061] Specifically, intermediate historical meteorological data refers to the preliminary purified data obtained after outlier processing of historical meteorological data. Intermediate historical operating data for each wind turbine refers to the preliminary purified data obtained after outlier processing of the historical operating data for each wind turbine. Updated historical meteorological data refers to the standardized meteorological data obtained after time alignment processing of intermediate historical meteorological data. Updated historical operating data for each wind turbine refers to the standardized operating data obtained after time alignment processing of the intermediate historical operating data for each wind turbine.

[0062] In practical implementation, for historical meteorological data, anomalies can be identified using a combination of rule-based judgment and statistical analysis. Examples include: range anomalies (data exceeding physically reasonable ranges); logical anomalies [inconsistent data between indicators (e.g., wind direction recorded as 0° but recorded as backwind power generation)]; and statistical anomalies—identifying values ​​deviating from the main distribution using methods such as box plots and the 3σ principle (e.g., wind speed exceeding the average for the same period by more than 10 times). After identification, the data is processed according to type. For example, missing values ​​are supplemented using linear interpolation or filling with the average of adjacent time periods; invalid, outlier, or contradictory values ​​are directly deleted. After processing, intermediate historical meteorological data is obtained.

[0063] For the historical operating data of each wind turbine, abnormal data can be identified first, such as: abnormal status [operating data during shutdown or fault maintenance periods (e.g., power output despite status "fault")]; abnormal indicators [data exceeding the rated parameter range of the equipment (e.g., power exceeding the rated value, or speed exceeding the design range)]; and abnormal transmission (incorrect data format or duplicate data). Then, abnormal data is processed, for example: data from fault / shutdown periods is directly deleted; after deleting transmission errors / duplicate data, only one valid record is retained; if the data is within the allowable fluctuation range (e.g., power ±5%), it is retained without correction. After processing, intermediate historical operating data for each wind turbine is obtained.

[0064] Subsequently, the intermediate historical meteorological data and the intermediate historical operating data of each wind turbine are time-aligned. Specifically, a unified collection frequency is first set according to actual needs (e.g., 15 minutes / data point). If the data frequencies are inconsistent (e.g., meteorological data 5 minutes / data point, wind turbine operating data 1 minute / data point), aggregation is performed according to the target granularity. During aggregation, a suitable method must be selected based on the characteristics of the indicators; for example, the maximum value of the interval can be used for wind speed, and the average value of the interval can be used for temperature and power. Then, the timestamps of all data are uniformly converted to a standard format (e.g., "YYYY-MM-DDHH:MM:SS"), and records with mismatched timestamps or exceeding the analysis time range are removed. For unilateral data loss caused by frequency unification or missing records, interpolation methods (e.g., linear interpolation, filling with the average of preceding and following time periods) are used to supplement the missing data, ensuring that the meteorological data and operating data completely correspond for any valid timestamp. Finally, updated historical meteorological data and updated historical operating data of each wind turbine are obtained.

[0065] Optionally, to further improve the spatial correlation and matching accuracy of the data, after obtaining the updated historical meteorological data and the updated historical operating data of each wind turbine, a preset spatial alignment method (such as the distance-weighted matching method based on the wind turbine and meteorological collection points, the grid point attribution mapping method, etc.) can be used to perform precise spatial dimension correlation processing on the two types of data, thereby obtaining high-quality standardized data with consistent spatial attributes that are suitable for subsequent fusion needs.

[0066] In this embodiment, the above steps can effectively remove invalid interference from the data, ensure data reliability, unify the data benchmark, and achieve accurate correlation between meteorological data and wind turbine operation data, thereby reducing the workload of format adaptation during subsequent fusion processing.

[0067] Step 211: Perform spatial mapping processing on historical meteorological data and historical operating data of each wind turbine to obtain wind turbine meteorological fusion data.

[0068] Further, step 211 may specifically include: constructing a two-dimensional spatial grid covering the wind farm based on the location information of the wind farm and a preset grid resolution; interpolating historical meteorological data into each cell of the two-dimensional spatial grid to obtain a meteorological two-dimensional spatial feature matrix; and mapping the historical operating data of each wind turbine to the meteorological two-dimensional spatial feature matrix based on the location information of each wind turbine to obtain wind turbine meteorological fusion data.

[0069] Specifically, the location information of a wind farm refers to the data that defines the geographical scope of the wind farm. For example, the location information of a wind farm can include its latitude and longitude boundaries, and the coordinates of its geographical center point. The preset grid resolution refers to the grid cell size of a two-dimensional spatial grid pre-set according to actual conditions or needs, such as 50m × 50m or 100m × 100m. The two-dimensional spatial grid refers to a set of regular rectangular cells formed by dividing the geographical scope defined by the wind farm location information according to the preset grid resolution. This grid must completely cover the wind farm area, and each grid cell must have a unique identifier (such as row and column numbers, and center point coordinates), with no spatial overlap or omissions between cells. The meteorological two-dimensional spatial feature matrix refers to a structured data matrix formed by filling each cell of the two-dimensional spatial grid with discrete historical meteorological data (such as wind tower data and third-party grid data) through spatial interpolation algorithms. Each cell stores the meteorological data for the corresponding location. The location information of each wind turbine refers to the geographical coordinate data (such as latitude and longitude coordinates) of each wind turbine within the wind farm.

[0070] In practice, the preset grid resolution can be converted into the corresponding latitude and longitude difference based on the Earth curvature correction coefficient of the latitude where the wind farm is located. For example, if the preset grid resolution is 1 kilometer, then the latitude difference = preset grid resolution / 111, and the longitude difference = preset grid resolution / (111 × Earth curvature correction coefficient).

[0071] Then, based on the latitude and longitude boundaries of the wind farm and the difference between them, the number of columns (longitude direction) and rows (latitude direction) of the grid are calculated by rounding up. For example, the number of columns = ⌈(original right longitude boundary of the wind farm - original left longitude boundary of the wind farm) / longitude difference⌉, and the number of rows = ⌈(original upper latitude boundary of the wind farm - original lower latitude boundary of the wind farm) / latitude difference⌉. Next, the actual boundary of the grid is deduced from the calculated number of rows and columns and the resolution to ensure that the grid completely covers the original wind farm boundary. If the deduced actual boundary exceeds the original wind farm boundary, no adjustment is needed; if it does not completely cover the boundary, 1-2 rows or columns need to be added and the boundary needs to be recalibrated. Afterward, a unique identifier (such as row and column number) is assigned to each unit in the calibrated grid in a preset order [such as from left to right (increasing longitude) or from bottom to top (increasing latitude)].

[0072] Next, the historical meteorological data after outlier processing and time alignment are grouped according to timestamps to ensure that each group of data corresponds to a meteorological record at a specific time node. Based on the actual situation, an appropriate spatial interpolation algorithm (such as Kriging interpolation or inverse distance weighted interpolation) is selected. Then, for the discrete meteorological data at each time node, the meteorological index value of each grid cell is calculated using an interpolation algorithm on a two-dimensional spatial grid as the carrier, thereby obtaining a meteorological two-dimensional spatial feature matrix.

[0073] Finally, for each wind turbine, its grid cell is determined based on its location information (such as latitude and longitude coordinates). Then, based on the timestamp, the wind turbine operation data at the same time is fused with the meteorological data of the corresponding grid cell (such as horizontal stitching) to obtain wind turbine meteorological fusion data.

[0074] In this embodiment, the above steps not only improve the scene adaptability of the data and realize the accurate binding of meteorological data and wind turbine operation data and strengthen the data correlation, but also improve the quality of the final wind turbine meteorological fusion data, thereby reducing the difficulty of model learning and improving training efficiency.

[0075] Furthermore, based on the location information of each wind turbine, the historical operating data of each wind turbine is mapped to a meteorological two-dimensional spatial feature matrix to obtain wind turbine meteorological fusion data. This includes: calculating the distance between each wind turbine and the center of each grid point in the meteorological two-dimensional spatial feature matrix based on the location information of each wind turbine, to obtain the first distance matrix corresponding to each wind turbine; taking the grid point corresponding to the minimum distance value in the first distance matrix corresponding to each wind turbine as the target associated grid point of each wind turbine; and fusing the historical operating data of each wind turbine with the historical meteorological data of the corresponding target associated grid point according to the timestamps of the historical meteorological data of the corresponding target associated grid point to obtain wind turbine meteorological fusion data.

[0076] Specifically, the grid center refers to the geometric center point of each grid cell (grid point) in the meteorological two-dimensional spatial feature matrix. The center point longitude = (minimum grid point longitude + maximum grid point longitude) / 2, and the center point latitude = (minimum grid point latitude + maximum grid point latitude) / 2. The first distance matrix refers to the set of distances between a single wind turbine and the centers of all grid points in the meteorological two-dimensional spatial feature matrix, calculated using spatial distance algorithms (such as spherical distance algorithm, planar approximate distance algorithm, etc.). The target associated grid point refers to the grid point matched for each wind turbine that is spatially closest to the wind turbine. The minimum distance value in the first distance matrix is ​​the minimum value of all distance values ​​in the first distance matrix for a single wind turbine.

[0077] In the specific implementation, for each wind turbine, the distance to the center of all grid cells is calculated sequentially using spatial distance algorithms (such as Vincent's formula, the great circle distance formula, and the spherical cosine theorem), and these distance values ​​are stored in a fixed order according to the unique identifier of the grid cells, resulting in the first distance matrix corresponding to that wind turbine. Then, based on the distance sorting, the spatially closest unique grid cell for each wind turbine is determined, i.e., the target associated grid cell. If multiple grid cells have the same distance to the wind turbine and all are the minimum, one can be randomly selected, or the determination can be made according to the principle of meteorological data integrity (such as prioritizing grid cells with no missing data). Next, wind turbine operation data with the same timestamp are bound (e.g., field concatenation) to the meteorological data of their associated grid cells (i.e., the corresponding target associated grid cells), resulting in wind turbine meteorological fusion data. The wind turbine meteorological fusion data can be understood as superimposing wind turbine operation data onto the corresponding spatial position of the meteorological two-dimensional spatial feature matrix, achieving a structured fusion of two types of data: wind turbine operation data superimposed on the meteorological two-dimensional spatial feature matrix.

[0078] In this embodiment, the above steps achieve precise spatial matching, allowing meteorological data to match the actual operating conditions of the wind turbine, improving data adaptability, ensuring the correspondence between the environment and the operating sequence, strengthening data correlation, and thus improving the quality of the fused data.

[0079] Step 212: Obtain historical meteorological data from each meteorological observation station within the wind farm.

[0080] Specifically, a meteorological observation station refers to a fixed observation facility (such as a wind measurement tower, integrated meteorological monitoring equipment, or a small weather station) deployed within a wind farm according to a pre-planned schedule for continuous monitoring of local meteorological conditions. Historical meteorological data refers to the raw meteorological observation data continuously recorded by each meteorological observation station within the wind farm from a certain point in the past to the present (such as the past 1 hour, 3 hours, or 1 day), at a pre-set collection frequency (such as 1 minute / data or 5 minutes / data).

[0081] In practice, based on a preset period, historical meteorological data of each meteorological station in the wind farm can be obtained from the database storing observation data of each meteorological station, thus providing a data foundation for obtaining updated wind turbine meteorological fusion data.

[0082] Step 213: Merge the historical meteorological data of each meteorological observation station with the wind turbine meteorological fusion data to obtain the updated wind turbine meteorological fusion data.

[0083] In practical implementation, the same method as described above, which involves spatially mapping historical meteorological data and historical operating data of each wind turbine to obtain wind turbine meteorological fusion data, can be used. This involves fusing historical meteorological data from each meteorological observation station with existing wind turbine meteorological fusion data to obtain updated wind turbine meteorological fusion data. Specifically, the historical meteorological data can be preprocessed (e.g., outlier handling, time alignment), and then, based on the location information of each meteorological observation station, the preprocessed historical meteorological data can be mapped onto the wind turbine meteorological fusion data to obtain updated wind turbine meteorological fusion data. The updated wind turbine meteorological fusion data can be understood as the historical meteorological data superimposed onto the corresponding spatial location of the wind turbine meteorological fusion data.

[0084] In this embodiment, the quality of the fused data is optimized through the above steps, thereby improving the accuracy of subsequent predictions.

[0085] Further, step 213 may specifically include: calculating the distance between each wind turbine and each meteorological observation station to obtain a second distance matrix corresponding to each wind turbine; identifying meteorological observation stations in the second distance matrix corresponding to each wind turbine whose distance is less than a preset distance threshold as associated observation stations of each wind turbine; and fusing the historical site meteorological data of the associated observation stations of each wind turbine with the wind turbine meteorological fusion data based on the timestamps of the wind turbine meteorological fusion data and the historical site meteorological data of the associated observation stations of each wind turbine to obtain updated wind turbine meteorological fusion data.

[0086] Specifically, the second distance matrix refers to a matrix formed by calculating the spatial distance between each wind turbine and all meteorological observation stations within the wind farm using a spatial distance algorithm, and then storing the distance values ​​in a fixed order (e.g., increasing the meteorological observation station numbers). The preset distance threshold refers to a critical value, such as 500 meters or 1 kilometer, pre-set according to requirements to determine whether the spatial distance between a meteorological observation station and a wind turbine is sufficiently close. The associated observation stations refer to the set of observation stations selected from all meteorological observation stations whose distance to a particular wind turbine is less than the preset distance threshold.

[0087] In the specific implementation, for each wind turbine, the distance between it and all meteorological observation stations within the wind farm is calculated sequentially using a spatial distance algorithm. These distance values ​​are then stored in a fixed order according to the unique identifiers of the meteorological observation stations, resulting in a second distance matrix corresponding to that wind turbine. Next, the distance values ​​in the second distance matrix are filtered based on a preset distance threshold, identifying meteorological observation stations with distances less than the threshold as associated observation stations for that wind turbine. Then, records with the same timestamp as the meteorological data of the associated observation stations are extracted from the wind turbine meteorological fusion data. If a wind turbine corresponds to multiple associated observation stations, the mean value of the meteorological data at the same timestamp for these observation stations (e.g., mean wind speed, mean temperature) is first calculated, and then this mean value is fused with the wind turbine meteorological fusion data at the corresponding timestamp (e.g., field concatenation, direct replacement of existing meteorological data, weighted fusion replacement of existing meteorological data). If it corresponds to only a single associated observation station, the meteorological data of that observation station is directly fused with the wind turbine meteorological fusion data, ultimately obtaining the updated wind turbine meteorological fusion data.

[0088] To ensure the temporal consistency between the wind turbine meteorological fusion data and the historical station meteorological data, and to avoid fusion deviations caused by differences in time granularity and format, thereby improving the accuracy and reliability of the updated wind turbine meteorological fusion data, the two types of data can be standardized and time aligned before data fusion processing.

[0089] In this embodiment, the above steps not only enhance the robustness and scenario adaptability of the data, further optimize the quality of wind turbine meteorological fusion data and strengthen data credibility, but also provide a higher-value data source for subsequent model input, thereby effectively improving prediction accuracy.

[0090] Step 214: Based on the wind turbine meteorological fusion data, predict the initial predicted power and predicted wind speed of each wind turbine at each time point within a preset future time period.

[0091] Step 215: Perform spatial mapping processing on the predicted meteorological data of the wind farm, the initial predicted power of each wind turbine, and the predicted wind speed within the preset future time period to obtain the target fusion data.

[0092] Specifically, target fusion data refers to structured forecast data formed by spatial mapping, which consists of three dimensions: spatial location, predicted weather, predicted power, and wind speed.

[0093] In practice, the same method as described above, which involves "spatial mapping of historical meteorological data and historical operating data of each wind turbine to obtain wind turbine meteorological fusion data", can be used to spatially map the predicted meteorological data of the wind farm, the initial predicted power of each wind turbine, and the predicted wind speed within a preset future time period to obtain the target fusion data.

[0094] In this embodiment, the above steps improve the accuracy and reliability of the target fusion data, providing high-quality core data support for subsequent power prediction optimization.

[0095] Step 216: Input the target fusion data into the pre-trained power prediction model to obtain the target predicted power of each wind turbine at each time point within the preset future time period.

[0096] Specifically, the power prediction model refers to a model obtained by training a deep learning model (such as Transformer, temporal convolutional network, generative adversarial network, etc.) with the historical target fusion data of different wind farms as input features and the actual output power of each wind turbine in a preset future period after this historical period as the label.

[0097] In practice, after obtaining the target fusion data, it can be input into a pre-trained power prediction model to obtain the target predicted power of each wind turbine at various time points within a preset future timeframe. Furthermore, to continuously improve model quality and prediction reliability, a full-process model optimization and validation mechanism can be established: cross-validation (such as k-fold cross-validation) reduces the risk of overfitting, and rolling training (updating training data with fixed time windows) enhances the model's generalization ability; simultaneously, historical data is used to backtest and validate the trained model, and model parameters and weights are automatically adjusted based on feedback indicators such as prediction error and accuracy. The training dataset is also updated regularly (supplementing the latest historical target fusion data and the corresponding actual output power of the wind turbines) to ensure the model always maintains good timeliness and prediction accuracy.

[0098] Furthermore, after obtaining the target predicted power results, post-processing (such as smoothing and anomaly detection) can be performed to optimize the stability and reliability of the prediction results. Subsequently, the processed prediction results are output in standardized formats (such as CSV, Parquet, JSON), and key information (such as the predicted power trend of each wind turbine, the total predicted power of the entire wind farm, and the predicted distribution at different time points) is displayed through visualization tools (such as line charts, heat maps, etc.) to intuitively present the prediction results and support subsequent decision-making (such as the formulation of grid dispatch schemes and the optimization of wind farm operation and maintenance plans).

[0099] Optionally, to further improve prediction accuracy, prediction strategies can be dynamically adjusted based on seasonal changes and regional climate characteristics. For example, for typical climate scenarios such as strong winds in spring and low temperatures in winter, a dedicated seasonal model can be trained—only historical data of the corresponding season (including fused data of the target during the same period and the actual output power of the wind turbine) is used for model training, making the model more suitable for the meteorological patterns and wind turbine operating characteristics of that season.

[0100] In this embodiment, the above steps improve the accuracy and stability of the final output target predicted power of each wind turbine, providing reliable power prediction support for subsequent grid dispatching, wind farm operation and maintenance and other scenarios.

[0101] The power prediction method provided in this invention first acquires historical meteorological data of the wind farm and historical operating data of each wind turbine within the wind farm, providing a data foundation for subsequent wind turbine meteorological fusion data. Next, spatial mapping processing is performed on the historical meteorological data and the historical operating data of each wind turbine to obtain wind turbine meteorological fusion data, generating a highly unified spatiotemporal dimensional fusion dataset, thereby improving the accuracy and robustness of subsequent predictions. Then, historical meteorological data from each meteorological observation station within the wind farm is acquired, providing a data foundation for subsequent updated wind turbine meteorological fusion data. The historical meteorological data from each meteorological observation station is fused with the wind turbine meteorological fusion data to obtain updated wind turbine meteorological fusion data, optimizing the quality of the fusion data and thus improving the accuracy of subsequent predictions. Finally, based on the wind turbine meteorological fusion data, the initial predicted power and predicted wind speed of each wind turbine at various time points within a preset future timeframe are predicted, improving the accuracy of the initial prediction results and providing a high-quality data foundation for subsequent secondary predictions, thereby reducing the correction burden of the optimization process and effectively improving the accuracy of the target predicted power. Then, spatial mapping processing is performed on the predicted meteorological data of the wind farm, the initial predicted power of each wind turbine, and the predicted wind speed within a preset future time period to obtain target fusion data. This improves the accuracy and reliability of the target fusion data, providing high-quality core data support for subsequent power prediction optimization. Finally, the target fusion data is input into a pre-trained power prediction model to obtain the target predicted power of each wind turbine at each time point within the preset future time period. This improves the accuracy and stability of the final output target predicted power of each wind turbine, providing reliable power prediction support for subsequent grid dispatching, wind farm operation and maintenance, and other scenarios. Therefore, the technical solution of this invention solves the problem of insufficient power prediction accuracy in existing technologies due to reliance on single-dimensional data, which makes it difficult to comprehensively reflect the operating status of complex power systems.

[0102] Figure 3 This is a schematic diagram of a power prediction device provided in an embodiment of the present invention. This device belongs to the same inventive concept as the power prediction methods in the above embodiments. Details not described in detail in the embodiments of the power prediction device can be found in the embodiments of the power prediction methods described above. Figure 3 As shown, the device includes:

[0103] like Figure 3 As shown, the device includes:

[0104] The acquisition module 310 is used to acquire historical meteorological data of the wind farm and historical operating data of each wind turbine in the wind farm;

[0105] The fusion module 320 is used to perform spatial mapping processing on the historical meteorological data and the historical operating data of each wind turbine to obtain wind turbine meteorological fusion data.

[0106] The first prediction module 330 is used to predict the initial predicted power and predicted wind speed of each wind turbine at each time point within a preset future time period based on the wind turbine meteorological fusion data.

[0107] The second prediction module 340 is used to predict the target predicted power of each wind turbine at each time point within the preset future time period based on the predicted meteorological data of the wind farm within the preset future time period, the initial predicted power of each wind turbine, and the predicted wind speed.

[0108] Based on the above embodiments, the fusion module 320 is specifically used for:

[0109] Based on the location information of the wind farm and the preset grid resolution, a two-dimensional spatial grid covering the wind farm is constructed; the historical meteorological data is interpolated into each cell of the two-dimensional spatial grid to obtain a meteorological two-dimensional spatial feature matrix; based on the location information of each wind turbine, the historical operating data of each wind turbine is mapped to the meteorological two-dimensional spatial feature matrix to obtain the wind turbine meteorological fusion data.

[0110] Based on the above embodiments, the fusion module 320 maps the historical operating data of each wind turbine to the meteorological two-dimensional spatial feature matrix based on the location information of each wind turbine, to obtain the wind turbine meteorological fusion data, including:

[0111] Based on the location information of each wind turbine, the distance between each wind turbine and the center of each grid point in the meteorological two-dimensional spatial feature matrix is ​​calculated to obtain the first distance matrix corresponding to each wind turbine; the grid point corresponding to the minimum distance value in the first distance matrix corresponding to each wind turbine is taken as the target associated grid point of each wind turbine; according to the timestamps of the historical operating data of each wind turbine and the historical meteorological data of the corresponding target associated grid point, the historical operating data of each wind turbine and the historical meteorological data of the corresponding target associated grid point are fused to obtain the wind turbine meteorological fusion data.

[0112] Based on the above embodiments, the device further includes:

[0113] The update module is used to obtain historical meteorological data from each meteorological observation station in the wind farm after obtaining the wind turbine meteorological fusion data; and to fuse the historical meteorological data from each meteorological observation station with the wind turbine meteorological fusion data to obtain the updated wind turbine meteorological fusion data.

[0114] Based on the above embodiments, the update module fuses the historical meteorological data of each meteorological observation station with the wind turbine meteorological fusion data to obtain updated wind turbine meteorological fusion data, including:

[0115] Calculate the distance between each wind turbine and each meteorological observation station to obtain the second distance matrix corresponding to each wind turbine; identify the meteorological observation stations in the second distance matrix corresponding to each wind turbine whose distance is less than a preset distance threshold as the associated observation stations of each wind turbine; based on the timestamps of the wind turbine meteorological fusion data and the historical station meteorological data of the associated observation stations of each wind turbine, fuse the historical station meteorological data of the associated observation stations of each wind turbine with the wind turbine meteorological fusion data to obtain the updated wind turbine meteorological fusion data.

[0116] Based on the above embodiments, the second prediction module 340 is specifically used for:

[0117] Spatial mapping processing is performed on the predicted meteorological data of the wind farm within the preset future time period, the initial predicted power of each wind turbine, and the predicted wind speed to obtain target fusion data; the target fusion data is input into a pre-trained power prediction model to obtain the target predicted power of each wind turbine at each time point within the preset future time period.

[0118] Based on the above embodiments, the device further includes:

[0119] The preprocessing module is used to perform outlier processing on the historical meteorological data and the historical operating data of each wind turbine in the wind farm after acquiring the historical meteorological data and the historical operating data of each wind turbine, respectively, to obtain intermediate historical meteorological data and intermediate historical operating data of each wind turbine; and to perform time alignment processing on the intermediate historical meteorological data and the intermediate historical operating data of each wind turbine to obtain updated historical meteorological data and updated historical operating data of each wind turbine.

[0120] The power prediction device provided in the embodiments of the present invention can execute the power prediction method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of executing the method.

[0121] It is worth noting that in the embodiments of the power prediction device described above, the various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional unit are only for easy differentiation and are not used to limit the scope of protection of the present invention.

[0122] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Figure 4A block diagram of an exemplary electronic device 4 suitable for implementing embodiments of the present invention is shown. Figure 4 The electronic device 4 shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of the present invention.

[0123] like Figure 4 As shown, electronic device 4 is represented in the form of a general-purpose computing electronic device. The components of electronic device 4 may include, but are not limited to: one or more processors or processing units 16, system memory 28, and bus 18 connecting different system components (including system memory 28 and processing unit 16).

[0124] Bus 18 represents one or more of several bus architectures, including a memory bus or memory controller, a peripheral bus, a graphics acceleration port, a processor, or a local bus using any of the various bus architectures. For example, these architectures include, but are not limited to, the Industry Standard Architecture (ISA) bus, the Micro Channel Architecture (MAC) bus, the Enhanced ISA bus, the Video Electronics Standards Association (VESA) local bus, and the Peripheral Component Interconnect (PCI) bus.

[0125] Electronic device 4 typically includes a variety of computer system readable media. These media can be any available media that can be accessed by electronic device 4, including volatile and non-volatile media, removable and non-removable media.

[0126] System memory 28 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 30 and / or cache memory 32. Electronic device 4 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, storage system 34 may be used to read and write non-removable, non-volatile magnetic media (… Figure 4 Not shown; usually referred to as a "hard drive"). Although Figure 4 Not shown, a disk drive for reading and writing to a removable non-volatile disk (e.g., a "floppy disk") and an optical disk drive for reading and writing to a removable non-volatile optical disk (e.g., a CD-ROM, DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to bus 18 via one or more data media interfaces. System memory 28 may include at least one program product having a set (e.g., at least one) of program modules configured to perform the functions of the embodiments of the present invention.

[0127] A program / utility 40 having a set (at least one) of program modules 42 may be stored, for example, in system memory 28. Such program modules 42 include, but are not limited to, an operating system, one or more application programs, other program modules, and program data. Each or some combination of these examples may include an implementation of a network environment. Program modules 42 typically perform the functions and / or methods described in the embodiments of the present invention.

[0128] Electronic device 4 can also communicate with one or more external devices 14 (e.g., keyboard, pointing device, display 24, etc.), and with one or more devices that enable a user to interact with electronic device 4, and / or with any device that enables electronic device 4 to communicate with one or more other computing devices (e.g., network card, modem, etc.). This communication can be performed through input / output (I / O) interface 22. Furthermore, electronic device 4 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) via network adapter 20. Figure 4 As shown, network adapter 20 communicates with other modules of electronic device 4 via bus 18. It should be understood that, although... Figure 4 Not shown, it can be combined with electronic device 4 to use other hardware and / or software modules, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0129] Processing unit 16 executes various functional applications and page displays by running programs stored in system memory 28, such as implementing the power prediction method provided in this embodiment of the invention, which includes:

[0130] Acquire historical meteorological data of the wind farm and historical operating data of each wind turbine in the wind farm;

[0131] Spatial mapping processing is performed on the historical meteorological data and the historical operating data of each wind turbine to obtain wind turbine meteorological fusion data;

[0132] Based on the aforementioned wind turbine meteorological fusion data, the initial predicted power and predicted wind speed of each wind turbine at each time point within a preset future time period are predicted.

[0133] Based on the predicted meteorological data of the wind farm within the preset future time period, the initial predicted power and predicted wind speed of each wind turbine, the target predicted power of each wind turbine at each time point within the preset future time period is predicted.

[0134] Of course, those skilled in the art will understand that the processor can also implement the technical solutions of the power prediction method provided in any embodiment of the present invention.

[0135] This invention provides a computer-readable storage medium storing a computer program thereon. When executed by a processor, the program implements, for example, the power prediction method provided in this invention, which includes:

[0136] Acquire historical meteorological data of the wind farm and historical operating data of each wind turbine in the wind farm;

[0137] Spatial mapping processing is performed on the historical meteorological data and the historical operating data of each wind turbine to obtain wind turbine meteorological fusion data;

[0138] Based on the aforementioned wind turbine meteorological fusion data, the initial predicted power and predicted wind speed of each wind turbine at each time point within a preset future time period are predicted.

[0139] Based on the predicted meteorological data of the wind farm within the preset future time period, the initial predicted power and predicted wind speed of each wind turbine, the target predicted power of each wind turbine at each time point within the preset future time period is predicted.

[0140] The computer storage medium of this invention can be any combination of one or more computer-readable media. A computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. A computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this document, a computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0141] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, capable of sending, propagating, or transmitting programs for use by or in connection with an instruction execution system, apparatus, or device.

[0142] Program code contained on a computer-readable medium may be transmitted using any suitable medium, including but not limited to: wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0143] Computer program code for performing the operations of this invention can be written in one or more programming languages ​​or a combination thereof. Programming languages ​​include object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages—such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0144] Those skilled in the art will understand that the modules or steps of the present invention described above can be implemented using general-purpose computing devices. They can be centralized on a single computing device or distributed across a network of multiple computing devices. Optionally, they can be implemented using computer-executable program code, thereby allowing them to be stored in a storage device for execution by a computing device, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Thus, the present invention is not limited to any particular combination of hardware and software.

[0145] Furthermore, the acquisition, storage, use, and processing of data in the technical solution of this invention all comply with relevant laws and regulations.

[0146] Note that the above description is merely a preferred embodiment of the present invention and the technical principles employed. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and various obvious changes, readjustments, and substitutions can be made without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in detail through the above embodiments, the present invention is not limited to the above embodiments, and may include many other equivalent embodiments without departing from the concept of the present invention, the scope of which is determined by the scope of the appended claims.

Claims

1. A power prediction method characterized by, The method comprises: acquiring historical meteorological data of a wind farm and historical operation data of each wind turbine in the wind farm; performing spatial mapping processing on the historical meteorological data and the historical operation data of each wind turbine to obtain wind turbine meteorological fusion data; based on the wind turbine meteorological fusion data, predicting initial predicted power and predicted wind speed of each wind turbine at each time point within a preset future time period; based on predicted meteorological data of the wind farm within the preset future time period, the initial predicted power and the predicted wind speed of each wind turbine, predicting target predicted power of each wind turbine at each time point within the preset future time period.

2. The power prediction method of claim 1, wherein, The spatial mapping processing on the historical meteorological data and the historical operation data of each wind turbine to obtain wind turbine meteorological fusion data comprises: based on the location information of the wind farm and a preset grid resolution, constructing a two-dimensional spatial grid covering the wind farm; interpolating the historical meteorological data into each unit of the two-dimensional spatial grid to obtain a meteorological two-dimensional spatial feature matrix; based on the location information of each wind turbine, mapping the historical operation data of each wind turbine to the meteorological two-dimensional spatial feature matrix to obtain the wind turbine meteorological fusion data.

3. The power prediction method of claim 2, wherein, The mapping of the historical operation data of each wind turbine to the meteorological two-dimensional spatial feature matrix based on the location information of each wind turbine to obtain the wind turbine meteorological fusion data comprises: based on the location information of each wind turbine, calculating the distance between each wind turbine and the center of each grid point in the meteorological two-dimensional spatial feature matrix to obtain a first distance matrix corresponding to each wind turbine; taking the grid point corresponding to the minimum distance value in the first distance matrix corresponding to each wind turbine as the target associated grid point of each wind turbine; based on the time stamps of the historical operation data of each wind turbine and the historical meteorological data of the corresponding target associated grid point, fusing the historical operation data of each wind turbine and the historical meteorological data of the corresponding target associated grid point to obtain the wind turbine meteorological fusion data.

4. The power prediction method of claim 1, wherein, After obtaining the wind turbine meteorological fusion data, further comprising: acquiring historical site meteorological data of each meteorological observation station in the wind farm; fusing the historical site meteorological data of each meteorological observation station and the wind turbine meteorological fusion data to obtain updated wind turbine meteorological fusion data.

5. The power prediction method of claim 4, wherein, The fusion of the historical site meteorological data of each meteorological observation station and the wind turbine meteorological fusion data to obtain updated wind turbine meteorological fusion data comprises: calculating the distance between each wind turbine and each meteorological observation station to obtain a second distance matrix corresponding to each wind turbine; determining the meteorological observation stations with a distance less than a preset distance threshold in the second distance matrix corresponding to each wind turbine as the associated observation stations of each wind turbine; based on the time stamps of the wind turbine meteorological fusion data and the historical site meteorological data of the associated observation stations of each wind turbine, fusing the historical site meteorological data of the associated observation stations of each wind turbine and the wind turbine meteorological fusion data to obtain updated wind turbine meteorological fusion data.

6. The power prediction method of claim 1, wherein, The prediction of the target predicted power of each wind turbine at each time point within the preset future time period based on the predicted meteorological data of the wind farm within the preset future time period, the initial predicted power and the predicted wind speed of each wind turbine comprises: The predicted meteorological data of the wind farm in the preset future time length, the initial predicted power and the predicted wind speed of each wind turbine are subjected to spatial mapping processing to obtain target fusion data; The target fusion data is input into a pre-trained power prediction model to obtain the target predicted power of each wind turbine at each time point in the preset future time length.

7. The power prediction method of claim 1, wherein, After obtaining the historical meteorological data of the wind farm and the historical operation data of each wind turbine in the wind farm, the method further includes: The historical meteorological data and the historical operation data of each wind turbine are subjected to abnormal value processing to obtain intermediate historical meteorological data and intermediate historical operation data of each wind turbine; The intermediate historical meteorological data and the intermediate historical operation data of each wind turbine are subjected to time alignment processing to obtain updated historical meteorological data and updated historical operation data of each wind turbine.

8. A power prediction apparatus characterized by comprising: The device includes: An acquisition module configured to acquire historical meteorological data of a wind farm and historical operation data of each wind turbine in the wind farm; A fusion module configured to perform spatial mapping processing on the historical meteorological data and the historical operation data of each wind turbine to obtain wind turbine meteorological fusion data; A first prediction module configured to predict initial predicted power and predicted wind speed of each wind turbine at each time point in a preset future time length based on the wind turbine meteorological fusion data; A second prediction module configured to predict target predicted power of each wind turbine at each time point in the preset future time length based on predicted meteorological data of the wind farm in the preset future time length, the initial predicted power and the predicted wind speed of each wind turbine.

9. An electronic device, comprising: The electronic device includes: at least one processor; and a memory connected with the at least one processor in communication; wherein the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor to enable the at least one processor to execute the power prediction method of any one of claims 1-7.

10. A storage medium containing computer-executable instructions, wherein: The computer executable instructions, when executed by a computer processor, are used to execute the power prediction method of any one of claims 1-7.

Images