Photovoltaic ultra-short-term power prediction method, device, equipment and medium

Abstract

The present invention relates to the field of clean energy technology and discloses a photovoltaic ultra-short-term power prediction method, comprising: acquiring radar cloud image data of a target photovoltaic station; using real-time radar cloud image data as starting data, performing meteorological extrapolation based on cloud cluster change characteristics fused with the radar cloud image data to obtain extrapolated radar cloud image data; inputting the extrapolated radar cloud image data and historical radar cloud image data into a pre-trained extrapolation cloud image correction model to obtain corrected radar cloud image data; determining meteorological characteristics of the target photovoltaic station based on the corrected radar cloud image data; performing photovoltaic power prediction on the target photovoltaic station based on the meteorological characteristics to obtain a photovoltaic power prediction result; and using radar cloud image data to perform meteorological feature extraction and photovoltaic power prediction, thereby reducing the impact of low accuracy in distributed photovoltaic ultra-short-term power prediction caused by complex weather phenomena under low-precision or under-measured cloud cover.

Description

Technical Field

[0001] The present invention relates to the field of clean energy technology, and in particular to a photovoltaic ultra-short-term power prediction method, device, equipment and medium. Background Art

[0002] Distributed photovoltaic power generation is a new type of power generation and comprehensive energy utilization method with broad development prospects.

[0003] For the photovoltaic power generation power prediction of distributed photovoltaic sites, due to the lack of high-precision meteorological measurement devices near the photovoltaic equipment installation sites, high-precision meteorological information with high spatial and temporal precision cannot be found around most photovoltaic power stations, and meteorological data such as high-precision cloud information at the installation location cannot be obtained; since the power prediction business in related technologies is usually implemented using satellite meteorological data with a spatial resolution of 10 kilometers and a temporal resolution of 1 hour, the distributed photovoltaic sites in related technologies face the power prediction challenge of coarse-grained meteorological data.

[0004] At the same time, due to the impact of complex weather phenomena such as cloudy, foggy, hail, overcast, rainy, and snowy weather on ultra-short-term power forecasts, the fluctuation pattern of distributed photovoltaic power generation is difficult to capture accurately, resulting in inaccurate power forecasts for distributed photovoltaic sites. Summary of the Invention

[0005] In view of this, the present invention provides a photovoltaic ultra-short-term power prediction method, device, equipment and medium to solve the problem of inaccurate photovoltaic power generation prediction caused by coarse-grained meteorological data and complex weather phenomena.

[0006] In a first aspect, the present invention provides a photovoltaic ultra-short-term power prediction method, which includes: obtaining radar cloud map data of a target photovoltaic station; wherein the radar cloud map data includes real-time radar cloud map data and historical radar cloud map data within a preset historical time period; using the real-time radar cloud map data as the starting data, performing meteorological extrapolation based on the fused cloud cluster change characteristics of the radar cloud map data to obtain extrapolated radar cloud map data; wherein the fused cloud cluster change characteristics are obtained by fusing the cloud cluster change characteristics of time-adjacent radar cloud layer data with the corresponding time series characteristics; inputting the extrapolated radar cloud map data and the historical radar cloud map data into a pre-trained extrapolation cloud map correction model to obtain corrected radar cloud map data; wherein the extrapolation cloud map correction model uses the change characteristics of the extrapolated radar cloud map data as the guiding characteristics, and corrects the extrapolated radar cloud map data based on the cloud cluster characteristics of the historical radar cloud map data; determining the meteorological characteristics of the target photovoltaic station based on the corrected radar cloud map data; performing photovoltaic power prediction on the target photovoltaic station based on the meteorological characteristics to obtain a photovoltaic power prediction result.

[0007] As an exemplary embodiment, the real-time radar cloud image data is used as the starting reporting time, and meteorological extrapolation is performed based on the fused cloud cluster change characteristics of the radar cloud image data to obtain the extrapolated radar cloud image data, including: extracting the optical flow change characteristics of the radar cloud image data that are adjacent in time sequence as the cloud cluster change characteristics; obtaining the time corresponding to the optical flow change characteristics as the time feature; using the real-time radar cloud image data as the starting reporting data, and performing meteorological extrapolation based on the optical flow change characteristics to obtain the extrapolated radar cloud image data.

[0008] As an exemplary embodiment, the photovoltaic ultra-short-term power prediction method also includes: taking the real-time radar cloud image data as the starting data, performing meteorological extrapolation based on the optical flow change characteristics, and obtaining first sub-extrapolated radar cloud image data within a first preset time length; performing multiple rounds of cyclic iterative meteorological extrapolation based on the extrapolated radar cloud image data, and in each round of cyclic iterative meteorological extrapolation, performing meteorological extrapolation based on the fused cloud change characteristics of the historical extrapolated radar cloud image data obtained in the previous round of meteorological extrapolation, and obtaining second sub-extrapolated radar cloud image data within the first preset time length; obtaining the extrapolated radar cloud image data based on the first sub-extrapolated radar cloud image data and the second sub-extrapolated radar cloud image data.

[0009] As an exemplary embodiment, the method for training the extrapolated cloud map correction model includes: obtaining historical extrapolated radar cloud map data of the target photovoltaic station, first historical radar cloud map data corresponding to the time series of the historical extrapolated radar cloud map data, and second historical radar cloud map data whose time series is later than the historical extrapolated radar cloud map data; inputting the historical extrapolated radar cloud map data, the first historical radar cloud map data, and the second historical radar cloud map data into a pre-built preset model for model training, using the changing characteristics of the historical extrapolated radar cloud map data as the guiding characteristics and the cloud cluster characteristics of the first historical radar cloud map data as the benchmark, learning the mapping relationship when mapping the historical extrapolated radar cloud map data to the second historical radar cloud map data, and obtaining the extrapolated cloud map correction model.

[0010] As an exemplary embodiment, the target photovoltaic station includes multiple sub-distributed photovoltaic stations, and the photovoltaic power prediction of the target photovoltaic station based on the meteorological characteristics to obtain the photovoltaic power prediction result includes: predicting the photovoltaic power of the target photovoltaic station based on the meteorological characteristics to obtain the power prediction result to be corrected; obtaining the target combination reflectivity greater than the preset combination reflectivity in the corrected radar cloud map data; obtaining the photovoltaic power prediction result that matches the target combination reflectivity in time and space as the power prediction result to be corrected; obtaining the correlation coefficient of the sub-distributed photovoltaic station corresponding to the power prediction result to be corrected when performing geographic location matching as the influence coefficient; if there is a historical power prediction result, obtaining the historical predicted power and actual power at the end time of the historical power prediction result; correcting the power prediction result to be corrected based on the size of the historical predicted power relative to the actual power and the influence coefficient to obtain the photovoltaic power prediction result.

[0011] As an exemplary embodiment, the power prediction result to be corrected is corrected based on the deviation between the historical predicted power and the actual power and the influence coefficient to obtain a photovoltaic power prediction result, including: if the historical predicted power is greater than the actual power, the power prediction result to be corrected is weakened based on the influence coefficient to obtain the photovoltaic power prediction result; if the historical predicted power is less than the actual power, the power prediction result to be corrected is enhanced based on the influence coefficient to obtain the photovoltaic power prediction result.

[0012] As an exemplary embodiment, the photovoltaic power prediction method further includes: if there is no historical power prediction result, weakening and correcting the power prediction result to be corrected based on the influence coefficient to obtain the photovoltaic power prediction result.

[0013] In a second aspect, the present invention provides a photovoltaic ultra-short-term power prediction device, which includes: an acquisition module for acquiring radar cloud map data of a target photovoltaic station; wherein the radar cloud map data includes real-time radar cloud map data and historical radar cloud map data within a preset historical period; a meteorological extrapolation module for using the real-time radar cloud map data as the starting data, performing meteorological extrapolation based on the fused cloud cluster change characteristics of the radar cloud map data, and obtaining extrapolated radar cloud map data; wherein the fused cloud cluster change characteristics are obtained by fusing the cloud cluster change characteristics of the radar cloud layer data adjacent in time series with the corresponding time series characteristics; and a correction module for obtaining the extrapolated radar cloud map data. A positive module is used to input the extrapolated radar cloud image data and the historical radar cloud image data into a pre-trained extrapolated cloud image correction model to obtain corrected radar cloud image data; wherein, the extrapolated cloud image correction model uses the change characteristics of the extrapolated radar cloud image data as a guiding feature, and corrects the extrapolated radar cloud image data based on the cloud cluster characteristics of the historical radar cloud image data; a meteorological feature determination module is used to determine the meteorological characteristics of the target photovoltaic station based on the corrected radar cloud image data; a photovoltaic power prediction module is used to predict the photovoltaic power of the target photovoltaic station based on the meteorological characteristics to obtain a photovoltaic power prediction result.

[0014] In a third aspect, the present invention provides a computer device comprising: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the photovoltaic ultra-short-term power prediction method of the first aspect or any corresponding embodiment thereof by executing the computer instructions.

[0015] In a fourth aspect, the present invention provides a computer-readable storage medium having computer instructions stored thereon, the computer instructions being used to enable a computer to execute the photovoltaic ultra-short-term power prediction method of the first aspect or any corresponding embodiment thereof.

[0016] The invention provides a photovoltaic ultra-short-term power prediction method, which includes: obtaining radar cloud map data of a target photovoltaic station; wherein the radar cloud map data includes real-time radar cloud map data and historical radar cloud map data within a preset historical time period; using the real-time radar cloud map data as the starting data, performing meteorological extrapolation based on the fused cloud cluster change characteristics of the radar cloud map data to obtain extrapolated radar cloud map data; wherein the fused cloud cluster change characteristics are obtained by fusing the cloud cluster change characteristics of radar cloud layer data adjacent in time series with the corresponding time series characteristics; inputting the extrapolated radar cloud map data and the historical radar cloud map data into a pre-trained extrapolation cloud map correction model to obtain corrected radar cloud map data; wherein the extrapolation cloud map correction model uses the change characteristics of the extrapolated radar cloud map data as the guiding characteristics, and corrects the extrapolated radar cloud map data based on the cloud cluster characteristics of the historical radar cloud map data; based on the corrected radar The cloud image data determines the meteorological characteristics of the target photovoltaic station; based on the meteorological characteristics, the photovoltaic power of the target photovoltaic station is predicted to obtain a photovoltaic power prediction result; the above method, first, the radar data used has the advantage of higher accuracy; secondly, the meteorological extrapolation is realized by integrating cloud change characteristics, which can consider the influence of the time interval between the real-time radar cloud image data and the historical radar cloud image data when calculating the cloud change characteristics to obtain the extrapolated radar cloud image data; and, the extrapolated radar cloud image correction model is used to extrapolate the extrapolated radar cloud image data, and the extrapolated radar cloud image data is corrected to adapt to the cloud characteristics with the guidance of the extrapolated radar cloud image data and the cloud characteristics provided by the historical radar cloud image data, so as to obtain more accurate corrected cloud image data; finally, the corrected cloud image data is used to extract meteorological characteristics and predict photovoltaic power, which can reduce the impact of low accuracy of distributed photovoltaic ultra-short-term power prediction caused by complex weather phenomena under low precision or missing cloud cover. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the specific embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0018] Figure 1 is a schematic flow chart of a photovoltaic ultra-short-term power prediction method according to an embodiment of the present invention;

[0019] Figure 2 is a structural block diagram of a photovoltaic ultra-short-term power prediction device according to an embodiment of the present invention;

[0020] Figure 3Schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0021] To make the purpose, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative efforts shall fall within the scope of protection of the present invention.

[0022] According to an embodiment of the present invention, an embodiment of a photovoltaic ultra-short-term power prediction method is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0023] For the photovoltaic power generation power prediction of distributed photovoltaic sites, due to the lack of high-precision meteorological measurement devices near the photovoltaic equipment installation sites, high-precision meteorological information with high spatial and temporal precision cannot be found around most photovoltaic power stations, and meteorological data such as high-precision cloud information at the installation location cannot be obtained; since the power prediction business in related technologies is usually implemented using satellite meteorological data with a spatial resolution of 10 kilometers and a temporal resolution of 1 hour, the distributed photovoltaic sites in related technologies face the power prediction challenge of coarse-grained meteorological data.

[0024] To solve this problem, in this embodiment, a photovoltaic ultra-short-term power prediction method is provided. Figure 1 FIG. 1 is a flow chart of a photovoltaic ultra-short-term power prediction method according to an embodiment of the present invention. Figure 1 As shown, the process includes the following steps:

[0025] Step S101, obtaining radar cloud map data of a target photovoltaic station; wherein the radar cloud map data includes real-time radar cloud map data and historical radar cloud map data within a preset historical time period.

[0026] As mentioned above, the photovoltaic power generation power prediction method in the related art faces the power prediction challenge of coarse-grained meteorological data; compared with satellite meteorological data with coarse granularity and lower accuracy, radar data usually has a spatial resolution of 250 meters and a temporal resolution of 6 minutes, which is more advantageous in terms of accuracy, and at the same time there will be no sea and land objects that affect reflectivity (such as deserts, snow, ice, etc.); therefore, in the present invention, radar cloud map data is used to realize ultra-short-term power prediction of photovoltaic stations.

[0027] Exemplarily, the radar cloud map data includes real-time radar cloud map data and historical radar cloud map data within a preset historical period; the radar cloud map data of the target photovoltaic station can be acquired through Doppler weather radar.

[0028] Specifically, the new generation of Doppler weather radar single volume scan data is stored in a binary file compressed in bz2 format and bin format. This is a weather radar with a scanning radius of 200km, a spatial resolution of 250M, and a time resolution of 6 minutes. The file includes Julian time, volume scan mode, reflectivity factor, radial velocity, spectral width, etc. There are three volume scan modes, corresponding to three different numbers of elevation angles; the file name includes radar parameters, station code, recording time and other information; the data is recorded in units of radial line data at each azimuth angle of each layer of elevation, and this unit is called a radial data; in the present invention, since the combined reflectivity is used to determine cloud thickness and height and predict severe convective weather such as heavy rain and storms, the combined reflectivity of the Doppler weather radar is used to form the radar raw data.

[0029] Exemplarily, after obtaining the radar raw data, the radar raw data is subjected to data standardization processing; specifically, first, the missing values ​​in the radar raw data are filled in, and further, in order to reduce the amount of subsequent forecast calculations and align with the time resolution of the station data, the data is time interpolated, and the average of 12 minutes and 18 minutes of the radar data of each hour is calculated as 15-minute radar data, and the average of 42 and 48 minutes is calculated as 45-minute radar data, and finally 00, 15, 30, and 45-minute radar data are obtained; then the polar coordinate system is converted into an isometric plane coordinate system, which can be regarded as an interpolation process, and bilinear interpolation is used to finally generate combined reflectivity data of 1600*1600 equal latitude and longitude grids, and standard grid radar data with a spatial resolution of 250m and a time resolution of 15 minutes.

[0030] After obtaining the standard radar grid data, the unit of the combined reflectivity contained in the standard radar grid data is dBZ, which represents the physical value of the radar echo intensity. Generally, an echo intensity greater than -10 dBZ indicates the presence of initial clouds, an echo intensity greater than 10 dBZ indicates the presence of cloudy clouds, an echo intensity greater than 25 dBZ indicates the presence of precipitation clouds, and an echo intensity greater than 35 dBZ indicates the presence of severe convective clouds. The radar data thresholds in Table 1 correspond to the following weather types:

[0031] Table 1. Correspondence between reflectivity factor, cloud (weather) type and characteristics

[0032] Reflectivity factor (dBZ) Cloud (weather) type feature -10 ~ 10 Thin clouds (ice crystals, high-altitude cirrus) No precipitation, weak radar echo 10 ~ 25 Cumulus / Stratus without precipitation May contain supercooled water droplets, but no effective precipitation 25 ~ 35 Light precipitation clouds (light rain / snow) The cloud contains dense water droplets or snowflakes, and precipitation may begin. ≥35 Heavy rainfall clouds (thunderstorms, hail) Strong convection, tight echo structure, and significant vertical development

[0033] According to the initial cloud appearance, the corresponding radar combination reflectivity factor is -10dbz. Now it is necessary to return the radar data less than or equal to -10dbz to zero, and normalize the data greater than -10dbz to between 0 and 1, where 0 represents 0 cloud cover, 1 represents 1 cloud cover, and the values ​​in between represent different degrees of cloud cover. Then, the normalized radar data is plotted into a grayscale radar cloud map with a grayscale value of 0-255 to complete the generation of radar cloud types and cloud maps.

[0034] Step S102, using the real-time radar cloud image data as the starting data, meteorological extrapolation is performed based on the fused cloud cluster change characteristics of the radar cloud image data to obtain extrapolated radar cloud image data; wherein the fused cloud cluster change characteristics are obtained by fusing the cloud cluster change characteristics of the radar cloud layer data adjacent in time series with the corresponding time series characteristics.

[0035] In this embodiment, real-time radar cloud image data is used as the starting data, and meteorological extrapolation is performed based on the fused cloud cluster change characteristics of the radar cloud image data to obtain extrapolated radar cloud image data.

[0036] Exemplarily, the optical flow method or a pre-trained deep learning extrapolation model can be used to use real-time radar cloud map data as the starting data, and meteorological data can be extrapolated based on historical radar cloud map data close to the real-time radar cloud map data to obtain extrapolated radar cloud map data.

[0037] For example, when implementing meteorological extrapolation of radar cloud image data based on cloud change characteristics, meteorological extrapolation is performed by calculating the cloud change characteristics of historical radar cloud image data N moments before the real-time radar cloud image, and then combining them with the real-time radar cloud image data.

[0038] For example, in order to consider the influence of the time interval between real-time radar cloud map data and historical radar cloud map data when calculating cloud change characteristics, after calculating the cloud change characteristics of the historical radar cloud map data N moments before the real-time radar cloud map, the cloud change characteristics of each cloud map are fused based on the time interval between the real-time radar cloud map and the historical radar cloud map data to obtain fused cloud change characteristics. The fused cloud change characteristics are further combined with the optical flow method or a pre-trained deep learning extrapolation model to realize meteorological extrapolation and obtain extrapolated radar cloud map data.

[0039] Step S103: input the extrapolated radar cloud map data and the historical radar cloud map data into a pre-trained extrapolated cloud map correction model to obtain corrected radar cloud map data; wherein, the extrapolated cloud map correction model uses the change characteristics of the extrapolated radar cloud map data as a guiding feature and corrects the extrapolated radar cloud map data based on the cloud cluster characteristics of the historical radar cloud map data.

[0040] For the extrapolated radar cloud image data obtained by meteorological extrapolation based on the fused cloud cluster change characteristics of the radar cloud image data, the data is relatively clear at the beginning of the extrapolation; but as the extrapolation time increases, the extrapolated radar cloud image data gradually becomes unclear. This shortcoming is particularly obvious when the optical flow method is used. Specifically, for ultra-short-term photovoltaic power prediction, the time scale is usually 4 hours. For the extrapolated radar data within 0 hours to 1 hour, its clarity can meet the application requirements; but for the extrapolated radar cloud image data within 1 hour to 4 hours, it gradually becomes distorted as the forecast time increases. For the extrapolated radar cloud image data obtained by the optical flow method, after 2.5 hours, many cloud clusters are limited by the long prediction time of the optical flow method, and the birth and death of cloud clusters are difficult to predict, resulting in inaccurate extrapolated radar cloud image data, thereby affecting the accuracy of photovoltaic power prediction by affecting the accuracy of meteorological characteristics.

[0041] Based on this, in the present invention, after obtaining the extrapolated radar cloud image data, it is further corrected based on a pre-trained extrapolated cloud image correction model to obtain the corrected radar cloud image data.

[0042] The extrapolated cloud image correction model uses the change characteristics of the extrapolated radar cloud image data as a guiding feature, and corrects the extrapolated radar cloud image data based on the cloud cluster characteristics of the historical radar cloud image data.

[0043] The above method of inputting the extrapolated radar cloud image data and the historical radar cloud image data into a pre-trained extrapolated cloud image correction model to obtain the corrected radar cloud image data, wherein the extrapolated cloud image correction model uses the changing characteristics of the extrapolated radar cloud image data as the guiding characteristics, and can consider the guidance of the extrapolated radar cloud image data and the cloud cluster characteristics provided by the historical radar cloud image data to correct the extrapolated radar cloud image data to adapt to the cloud cluster characteristics, thereby obtaining the corrected radar cloud image data.

[0044] Step S104: determining the meteorological characteristics of the target photovoltaic station based on the corrected radar cloud image data.

[0045] In this embodiment, the spatial scale corresponding to the modified radar cloud image data includes the target photovoltaic station. Therefore, spatial matching is first performed based on the modified radar cloud image data to obtain meteorological data at the corresponding position of the target photovoltaic station.

[0046] For example, the proximity matching method can be used to complete the spatial matching of the target photovoltaic station and the corrected radar cloud map data, obtain the location information of the target photovoltaic station in the corrected radar cloud map data, and further determine the meteorological characteristics of the target photovoltaic station based on the comparison table of the combined reflectivity corresponding to the location information and the precipitation type.

[0047] For example, Table 2 is a comparison table of an exemplary combination of reflectivity and precipitation type according to the present invention, see Table 2:

[0048] Reflectivity factor (dBZ) Precipitation type <10 sunny 10~15 partly cloudy 15~30 light rain (snow) 30~40 moderate rain (snow) 40~46 Heavy rain (snow) 46~50 Heavy rain (snow) 50~56 Heavy rain (snow) >56 Heavy rain (snow)

[0049] Step S105 , performing photovoltaic power prediction on the target photovoltaic station based on the meteorological characteristics to obtain a photovoltaic power prediction result.

[0050] After obtaining the meteorological characteristics, the photovoltaic power of the target photovoltaic station is further predicted based on the meteorological characteristics to obtain the photovoltaic power prediction result.

[0051] For example, meteorological characteristics may be input into a pre-trained photovoltaic power prediction model to obtain a photovoltaic power prediction result.

[0052] Exemplarily, the target photovoltaic station includes a distributed photovoltaic station; since the distributed photovoltaic station is a discrete distribution structure, spatial matching of the grid and the discrete photovoltaic station is required; the proximity matching method can be used to complete spatial matching based on the geographical characteristics of each distributed photovoltaic station to obtain the power prediction results of each distributed photovoltaic station.

[0053] The present invention provides a photovoltaic ultra-short-term power prediction method, which comprises: obtaining radar cloud map data of a target photovoltaic station; wherein the radar cloud map data comprises real-time radar cloud map data and historical radar cloud map data within a preset historical time period; taking the real-time radar cloud map data as the starting data, performing meteorological extrapolation based on the fused cloud cluster change characteristics of the radar cloud map data to obtain extrapolated radar cloud map data; wherein the fused cloud cluster change characteristics are obtained by fusing the cloud cluster change characteristics of radar cloud layer data adjacent in time sequence with the corresponding time series characteristics; inputting the extrapolated radar cloud map data and the historical radar cloud map data into a pre-trained extrapolation cloud map correction model to obtain corrected radar cloud map data; wherein the extrapolation cloud map correction model uses the change characteristics of the extrapolated radar cloud map data as the guiding characteristics, and corrects the extrapolated radar cloud map data based on the cloud cluster characteristics of the historical radar cloud map data; and correcting the extrapolated radar cloud map data based on the corrected radar cloud map data. The cloud image data determines the meteorological characteristics of the target photovoltaic station; based on the meteorological characteristics, the photovoltaic power of the target photovoltaic station is predicted to obtain a photovoltaic power prediction result; the above method, first, the radar data used has the advantage of higher accuracy; secondly, the meteorological extrapolation is realized by integrating cloud change characteristics, which can consider the influence of the time interval between the real-time radar cloud image data and the historical radar cloud image data when calculating the cloud change characteristics to obtain the extrapolated radar cloud image data; and, the extrapolated radar cloud image correction model is used to extrapolate the extrapolated radar cloud image data, and the extrapolated radar cloud image data is corrected to adapt to the cloud characteristics with the guidance of the extrapolated radar cloud image data and the cloud characteristics provided by the historical radar cloud image data, so as to obtain more accurate corrected cloud image data; finally, the corrected cloud image data is used to extract meteorological characteristics and predict photovoltaic power, which can reduce the impact of low accuracy of distributed photovoltaic ultra-short-term power prediction caused by complex weather phenomena under low precision or missing cloud cover.

[0054] As an exemplary embodiment, the real-time radar cloud image data is used as the starting reporting time, and meteorological extrapolation is performed based on the fused cloud cluster change characteristics of the radar cloud image data to obtain the extrapolated radar cloud image data, including: extracting the optical flow change characteristics of the radar cloud image data that are adjacent in time sequence as the cloud cluster change characteristics; obtaining the time corresponding to the optical flow change characteristics as the time feature; using the real-time radar cloud image data as the starting reporting data, and performing meteorological extrapolation based on the optical flow change characteristics to obtain the extrapolated radar cloud image data.

[0055] In this embodiment, the optical flow method is used to implement meteorological extrapolation based on the cloud cluster change characteristics fused with the radar cloud image data.

[0056] Specifically, the real-time radar cloud map data is used as the starting time, and a total of N-1 historical radar cloud map data and 1 real-time radar cloud map data at the starting time and the N nearest moments are used. The optical flow change characteristics of the real-time radar cloud map data and the historical radar cloud map data at the N nearest moments are calculated by the optical flow method (cloud map pixels will move over time, and the optical flow change characteristics are the speed and direction of the moving pixels). Different fusion weights are used to fuse to obtain a comprehensive optical flow change characteristic, and then the real-time radar cloud map data is combined with the comprehensive optical flow change characteristic to obtain a new radar cloud map as the radar cloud map forecast for the future moment; wherein N is a positive integer greater than 1, the sum of the fusion weights is 1, and is inversely proportional to the time characteristic and the time interval determined by the starting time.

[0057] In one embodiment, N may be 16.

[0058] In one embodiment, N may be 4; that is, a total of 3 historical radar cloud map data and 1 real-time radar cloud map data at the reporting time and the 4 moments before the reporting time are used, and the optical flow change characteristics of the real-time radar cloud map data and the historical radar cloud map data at the 4 moments before the reporting time are calculated by the optical flow method (cloud map pixels will move over time, and the optical flow change characteristics are the speed and direction of the moving pixels). Different fusion weights are used to fuse to obtain a comprehensive optical flow change characteristic, and then the real-time radar cloud map data is combined with the comprehensive optical flow change characteristic to obtain a new radar cloud map as the radar cloud map forecast for the future moment.

[0059] In one embodiment, a loop iteration method is used to perform extrapolation based on the obtained new sub-extrapolated radar cloud map data to obtain extrapolated radar cloud map data that meets the prediction time of photovoltaic ultra-short-term power prediction.

[0060] Specifically, as an exemplary embodiment, the photovoltaic ultra-short-term power prediction method also includes: taking the real-time radar cloud map data as the starting data, performing meteorological extrapolation based on the optical flow change characteristics, and obtaining first sub-extrapolated radar cloud map data within a first preset time length; performing multiple rounds of cyclic iterative meteorological extrapolation based on the extrapolated radar cloud map data, and in each round of cyclic iterative meteorological extrapolation, performing meteorological extrapolation based on the fused cloud change characteristics of the historical extrapolated radar cloud map data obtained in the previous round of meteorological extrapolation, and obtaining second sub-extrapolated radar cloud map data within the first preset time length; obtaining the extrapolated radar cloud map data based on the first sub-extrapolated radar cloud map data and the second sub-extrapolated radar cloud map data.

[0061] Among them, the preset duration can be 1 hour.

[0062] As an exemplary embodiment, the method for training the extrapolated cloud map correction model includes: obtaining historical extrapolated radar cloud map data of the target photovoltaic station, first historical radar cloud map data corresponding to the time series of the historical extrapolated radar cloud map data, and second historical radar cloud map data whose time series is later than the historical extrapolated radar cloud map data; inputting the historical extrapolated radar cloud map data, the first historical radar cloud map data, and the second historical radar cloud map data into a pre-built preset model for model training, using the changing characteristics of the historical extrapolated radar cloud map data as the guiding characteristics and the cloud cluster characteristics of the first historical radar cloud map data as the benchmark, learning the mapping relationship when mapping the historical extrapolated radar cloud map data to the second historical radar cloud map data, and obtaining the extrapolated cloud map correction model.

[0063] As described above, the extrapolated radar cloud image data using the optical flow method is relatively clear in the first hour to the first hour, and becomes increasingly unclear as the forecast time increases from the first hour to the fourth hour. In particular, after the second and a half hour, many cloud clusters are limited by the long forecast time of the optical flow method, resulting in difficulty in predicting the birth and extinction of cloud clusters. The adversarial network can use its own generator and discriminator game principle to learn the detailed characteristics of the birth and extinction of cloud clusters and the changes in cloud clusters. At the same time, the preliminary ultra-short-term radar cloud image is instructive, reducing the difficulty of the adversarial network prediction. Therefore, in this embodiment, a preset model such as a Wasserstein Generative Adversarial Net (WGAN) or a Deep Convolutional Generative Adversarial Network (DCGAN) can be constructed, and the historical extrapolated radar cloud image data, the first historical radar cloud image data, and the second historical radar cloud image data can be input into the pre-constructed preset model for model training, thereby obtaining an extrapolated cloud image correction model.

[0064] Exemplarily, in the training process of the preset model, the cloud cluster features of 16 historical extrapolated radar cloud map data and the 16 first historical radar cloud map data before the reporting time are used, and the 16 first historical radar cloud map data after the reporting time are used as labels, and every 15 minutes is the reporting time to train an ultra-short-term radar cloud map extrapolation model; after obtaining the extrapolation cloud map correction model, the cloud cluster features provided by the 16 third historical radar cloud map time series data before the reporting time and the guidance features provided by the 16 extrapolated radar cloud map data are used to generate ultra-short-term radar cloud maps every 15 minutes in the next 4 hours, thereby obtaining the corrected radar cloud map data.

[0065] In the present invention, for distributed photovoltaic stations, specific types of weather data are further repaired in combination with their weather types; specifically, as an exemplary embodiment, the target photovoltaic station includes multiple sub-distributed photovoltaic stations, and the photovoltaic power prediction for the target photovoltaic station based on the meteorological characteristics is performed to obtain the photovoltaic power prediction result, including: performing photovoltaic power prediction for the target photovoltaic station based on the meteorological characteristics to obtain a power prediction result to be corrected; in the corrected radar cloud map data, obtaining a target combination reflectivity greater than a preset combination reflectivity; obtaining the photovoltaic power prediction result that matches the target combination reflectivity in time and space as the power prediction result to be corrected; obtaining the correlation coefficient of the sub-distributed photovoltaic station corresponding to the power prediction result to be corrected when performing geographic location matching as an influence coefficient; if there is a historical power prediction result, obtaining the historical predicted power and actual power at the end time of the historical power prediction result; correcting the power prediction result to be corrected based on the size of the historical predicted power relative to the actual power and the influence coefficient to obtain a photovoltaic power prediction result.

[0066] In this embodiment, after obtaining the meteorological characteristics, a general photovoltaic power prediction is first performed on a preset area based on the meteorological characteristics to obtain a photovoltaic power prediction result; wherein the preset area is a larger area containing the target photovoltaic station.

[0067] Furthermore, the actual combined reflectivity of each distributed photovoltaic station is compared with a preset reflectivity threshold to obtain the precipitation type corresponding to each distributed photovoltaic station.

[0068] For example, Table 3 is a comparison table of combined reflectivity and precipitation type of an exemplary distributed photovoltaic station according to the present invention:

[0069] Table 3. Comparison table of combined reflectivity and precipitation types of distributed photovoltaic stations

[0070] Serial number Reflectivity factor (dBZ) Precipitation type 1 <10 sunny 2 10~15 partly cloudy 3 15~30 light rain (snow) 4 30~40 moderate rain (snow) 5 40~46 Heavy rain (snow) 6 46~50 Heavy rain (snow) 7 50~56 Heavy rain (snow) 8 >56 Heavy rain (snow)

[0071] In the present invention, the ultra-short-term power forecast data of the distributed photovoltaic station corresponding to the weather phenomenon of thicker clouds above thin clouds is selected for correction. Based on this, the target combination reflectivity greater than the preset combination reflectivity is obtained in the corrected radar cloud map data; wherein the preset combination reflectivity can be 10dBZ.

[0072] Furthermore, the photovoltaic power prediction result that is spatially and temporally matched with the target combined reflectivity is obtained as the power prediction result to be corrected; and the correlation coefficient of the sub-distributed photovoltaic station corresponding to the power prediction result to be corrected when performing geographic location matching is obtained as the influence coefficient.

[0073] Since the ultra-short-term predicted power of distributed photovoltaic stations tends to be too high or too low compared to the actual power, in the present invention, the calculated influence coefficient is used to weaken or improve the ultra-short-term predicted power of distributed photovoltaic stations.

[0074] In addition, the ultra-short-term power prediction time of distributed photovoltaic stations is usually 4 hours. During the prediction, the prediction offset increases with the prediction time. According to historical distributed photovoltaic ultra-short-term data, if the fourth hour is high, the previous hours will be slightly higher, and if the fourth hour is low, the previous hours will be slightly lower.

[0075] Therefore, in the present invention, it is first determined whether there are historical power prediction results for each photovoltaic station; if there are historical power prediction results, the historical predicted power and actual power at the end time of the historical power prediction results are obtained; based on the size of the historical predicted power relative to the actual power and the influence coefficient, the power prediction result to be corrected is corrected to obtain the photovoltaic power prediction result.

[0076] Exemplarily, if the historical predicted power is greater than the actual power, a weakening strategy is adopted to use the influence coefficient to weaken the ultra-short-term predicted power of the distributed photovoltaic station in proportion to the influence coefficient; that is, if the historical predicted power is greater than the actual power, the power prediction result to be corrected is weakened and corrected based on the influence coefficient to obtain the photovoltaic power prediction result.

[0077] Exemplarily, if the historical predicted power is less than the actual power, an improvement strategy is adopted to use the influence coefficient to improve the ultra-short-term predicted power of the distributed photovoltaic station in proportion to the influence coefficient; that is, if the historical predicted power is less than the actual power, the power prediction result to be corrected is improved and corrected based on the influence coefficient to obtain the photovoltaic power prediction result.

[0078] Furthermore, for distributed photovoltaic stations that cannot fully identify the high and low prediction trends, the present invention adopts a unified weakening strategy for correction; based on this, as an exemplary embodiment, the photovoltaic power prediction method also includes: if there is no historical power prediction result, the power prediction result to be corrected is weakened and corrected based on the influence coefficient to obtain the photovoltaic power prediction result.

[0079] This embodiment provides a photovoltaic ultra-short-term power prediction device, such as Figure 2 Shown, including:

[0080] An acquisition module 501 is configured to acquire radar cloud map data of a target photovoltaic station; wherein the radar cloud map data includes real-time radar cloud map data and historical radar cloud map data within a preset historical period;

[0081] The meteorological extrapolation module 502 is configured to use the real-time radar cloud image data as the starting data and perform meteorological extrapolation based on the cloud cluster change characteristics fused with the radar cloud image data to obtain extrapolated radar cloud image data; wherein the fused cloud cluster change characteristics are obtained by fusing the cloud cluster change characteristics based on the time-series adjacent radar cloud layer data with the corresponding time series characteristics;

[0082] A correction module 503 is configured to input the extrapolated radar cloud image data and the historical radar cloud image data into a pre-trained extrapolated cloud image correction model to obtain corrected radar cloud image data; wherein the extrapolated cloud image correction model uses the change characteristics of the extrapolated radar cloud image data as a guide feature and corrects the extrapolated radar cloud image data based on the cloud cluster characteristics of the historical radar cloud image data;

[0083] A meteorological characteristic determination module 504 is configured to determine the meteorological characteristics of the target photovoltaic station based on the corrected radar cloud image data;

[0084] The photovoltaic power prediction module 505 is configured to perform photovoltaic power prediction on the target photovoltaic station based on the meteorological characteristics to obtain a photovoltaic power prediction result.

[0085] It should be noted here that the examples and application scenarios implemented by the above modules and corresponding steps are the same, but are not limited to the contents disclosed in the above embodiments.

[0086] It should be noted that the above modules as part of the device can be implemented through software or hardware, wherein the hardware environment includes a network environment.

[0087] An embodiment of the present invention also provides a computer device, including a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other through the communication bus, the memory is used to store computer programs; the processor is used to execute the method in any of the above embodiments by running the computer program stored in the memory.

[0088] Figure 3 is a structural block diagram of an optional computer device according to an embodiment of the present application, such as Figure 3 As shown, it includes a processor 10, a communication interface 20, a memory 30 and a communication bus 40, wherein the processor 10, the communication interface 20 and the memory 30 communicate with each other through the communication bus 40, wherein,

[0089] Memory 30, for storing computer programs;

[0090] The processor 10 is configured to implement the photovoltaic ultra-short-term power prediction method according to any of the above embodiments when executing the computer program stored in the memory 30 .

[0091] Optionally, in this embodiment, the communication bus may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus. The communication bus may be divided into an address bus, a data bus, a control bus, and the like. For ease of representation, Figure 3 Only one thick line is used in the diagram, but this does not mean that there is only one bus or one type of bus.

[0092] The communication interface is used for communication between the above-mentioned computer device and other devices.

[0093] The memory may include RAM, or may include non-volatile memory, such as at least one disk memory. Alternatively, the memory may also be at least one storage device located away from the aforementioned processor.

[0094] The above-mentioned processor can be a general-purpose processor, which can include but is not limited to: CPU (Central Processing Unit), NP (Network Processor), etc.; it can also be DSP (Digital Signal Processing), ASIC (Application Specific Integrated Circuit), FPGA (Field-Programmable Gate Array) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0095] Optionally, the specific examples in this embodiment may refer to the examples described in the above embodiments, and this embodiment will not be described in detail here.

[0096] It can be understood by those skilled in the art that Figure 3 The structure shown is for illustration only. The device for implementing any one of the methods in the above embodiments may be a terminal device, which may be a smart phone (such as an Android phone, an IOS phone, etc.), a tablet computer, a PDA, a mobile Internet device (MID), a PAD, or other terminal devices. Figure 3 It does not limit the structure of the above electronic device. For example, the terminal device may also include Figure 3 More or fewer components (such as network interfaces, display devices, etc.) shown in, or with Figure 3 Different configurations shown.

[0097] A person skilled in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing the hardware related to the terminal device through a program, and the program can be stored in a computer-readable storage medium, which can include: a flash drive, ROM, RAM, a magnetic disk or an optical disk, etc.

[0098] As an exemplary embodiment, the present application further provides a computer-readable storage medium, in which a computer program is stored, wherein the computer program is configured to execute any one of the method steps of the present embodiment when run.

[0099] Optionally, in this embodiment, the above-mentioned storage medium can be used to execute the program code of the method steps of the embodiment of the present application.

[0100] Optionally, in this embodiment, the above-mentioned storage medium may be located on at least one network device among the multiple network devices in the network shown in the above-mentioned embodiment.

[0101] Optionally, in this embodiment, the storage medium is configured to store data for executing the method in the above embodiment.

[0102] Optionally, the specific examples in this embodiment may refer to the examples described in the above embodiments, which will not be described in detail in this embodiment.

[0103] Optionally, in this embodiment, the storage medium may include, but is not limited to, various media that can store program codes, such as a USB flash drive, a ROM, a RAM, a mobile hard disk, a magnetic disk, or an optical disk.

[0104] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0105] If the integrated units in the above embodiments are implemented in the form of software functional units and sold or used as independent products, they can be stored in the above-mentioned computer-readable storage medium. Based on this understanding, the technical solution of this application, or the part that contributes to the existing technology, or all or part of the technical solution can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes a number of instructions for causing one or more computer devices (such as personal computers, servers, or network devices) to execute all or part of the steps of the method in the above embodiments.

[0106] In the several embodiments provided in this application, it should be understood that the disclosed client can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is merely a logical functional division. In actual implementation, there may be other division methods, such as combining or integrating multiple units or components into another system, or ignoring or not implementing some features. Furthermore, the mutual coupling or direct coupling or communication connection shown or discussed may be through some interface, indirect coupling or communication connection of units or modules, and may be electrical or other forms.

[0107] Units described as separate components may or may not be physically separate, and components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected based on actual needs to achieve the purpose of the solution provided in this embodiment.

[0108] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0109] In the above embodiments of the present application, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, please refer to the relevant description of other embodiments.

[0110] The above is only a preferred embodiment of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A photovoltaic ultra-short-term power prediction method, characterized in that: The photovoltaic ultra-short-term power prediction method includes: Acquire radar cloud map data of the target photovoltaic station; wherein the radar cloud map data includes real-time radar cloud map data and historical radar cloud map data within a preset historical period; Using the real-time radar cloud image data as the starting data, meteorological extrapolation is performed based on the fused cloud cluster change characteristics of the radar cloud image data to obtain extrapolated radar cloud image data; wherein the fused cloud cluster change characteristics are obtained by fusing the cloud cluster change characteristics based on the temporally adjacent radar cloud layer data with the corresponding temporal characteristics; Inputting the extrapolated radar cloud image data and the historical radar cloud image data into a pre-trained extrapolated cloud image correction model to obtain corrected radar cloud image data; wherein the extrapolated cloud image correction model uses the change characteristics of the extrapolated radar cloud image data as a guiding feature and corrects the extrapolated radar cloud image data based on the cloud cluster characteristics of the historical radar cloud image data; Determining meteorological characteristics of the target photovoltaic station based on the corrected radar cloud image data; The photovoltaic power prediction result is obtained by performing photovoltaic power prediction on the target photovoltaic station based on the meteorological characteristics, wherein the target photovoltaic station includes a plurality of sub-distributed photovoltaic stations, and the photovoltaic power prediction result is obtained by performing photovoltaic power prediction on the target photovoltaic station based on the meteorological characteristics. Performing photovoltaic power forecasting on the target photovoltaic station based on the meteorological characteristics to obtain a power forecast result to be corrected; In the corrected radar cloud image data, obtaining a target combined reflectivity greater than a preset combined reflectivity; Acquire the photovoltaic power prediction result that is spatially and temporally matched with the target combined reflectivity as the power prediction result to be corrected; Obtaining a correlation coefficient of the sub-distributed photovoltaic station corresponding to the power prediction result to be corrected when performing geographical location matching as an influence coefficient; If there is a historical power prediction result, obtain the historical predicted power and actual power at the end time of the historical power prediction result; The power prediction result to be corrected is corrected based on the magnitude of the historical predicted power relative to the actual power and the influence coefficient to obtain a photovoltaic power prediction result.

2. The photovoltaic ultra-short-term power prediction method according to claim 1, characterized in that: The method of taking the real-time radar cloud image data as the starting time and performing meteorological extrapolation based on the cloud cluster change characteristics fused with the radar cloud image data to obtain the extrapolated radar cloud image data includes: extracting optical flow change features of the radar cloud image data adjacent in time sequence as the cloud cluster change features; Obtaining the moment corresponding to the optical flow change feature as a time feature; The real-time radar cloud image data is used as the starting data, and meteorological extrapolation is performed based on the optical flow change characteristics to obtain the extrapolated radar cloud image data.

3. The photovoltaic ultra-short-term power prediction method according to claim 2, characterized in that: The photovoltaic ultra-short-term power prediction method further includes: Using the real-time radar cloud image data as the starting data, meteorological extrapolation is performed based on the optical flow change characteristics to obtain first sub-extrapolated radar cloud image data within a first preset time period; Performing multiple rounds of cyclic iterative meteorological extrapolation based on the extrapolated radar cloud image data, and in each round of cyclic iterative meteorological extrapolation, performing meteorological extrapolation based on the fused cloud cluster change characteristics of the historical extrapolated radar cloud image data obtained in the previous round of meteorological extrapolation, to obtain second sub-extrapolated radar cloud image data within a first preset time period; The extrapolated radar cloud image data is obtained based on the first sub-extrapolated radar cloud image data and the second sub-extrapolated radar cloud image data.

4. The photovoltaic ultra-short-term power prediction method according to claim 1, wherein: The method for training the extrapolated cloud image correction model includes: Acquire historical extrapolated radar cloud image data of a target photovoltaic station, first historical radar cloud image data corresponding to a time sequence of the historical extrapolated radar cloud image data, and second historical radar cloud image data whose time sequence is later than the historical extrapolated radar cloud image data; The historical extrapolated radar cloud map data, the first historical radar cloud map data and the second historical radar cloud map data are input into a pre-built preset model for model training. The changing characteristics of the historical extrapolated radar cloud map data are used as guiding features, and the cloud cluster characteristics of the first historical radar cloud map data are used as a benchmark. The mapping relationship when mapping the historical extrapolated radar cloud map data to the second historical radar cloud map data is learned to obtain the extrapolated cloud map correction model.

5. The photovoltaic ultra-short-term power prediction method according to claim 1, wherein: The step of correcting the power prediction result to be corrected based on the deviation between the historical predicted power and the actual power and the influence coefficient to obtain a photovoltaic power prediction result includes: If the historical predicted power is greater than the actual power, the power prediction result to be corrected is weakened and corrected based on the influence coefficient to obtain the photovoltaic power prediction result; If the historical predicted power is less than the actual power, the power prediction result to be corrected is improved based on the influence coefficient to obtain the photovoltaic power prediction result.

6. The photovoltaic ultra-short-term power prediction method according to claim 1, wherein: Also includes: If there is no historical power prediction result, the power prediction result to be corrected is weakened and corrected based on the influence coefficient to obtain the photovoltaic power prediction result.

7. A photovoltaic ultra-short-term power prediction device, characterized in that: The photovoltaic ultra-short-term power prediction device comprises: An acquisition module is used to acquire radar cloud map data of a target photovoltaic station; wherein the radar cloud map data includes real-time radar cloud map data and historical radar cloud map data within a preset historical period; A meteorological extrapolation module is configured to use the real-time radar cloud image data as the starting data and perform meteorological extrapolation based on the fused cloud cluster change characteristics of the radar cloud image data to obtain extrapolated radar cloud image data; wherein the fused cloud cluster change characteristics are obtained by fusing the cloud cluster change characteristics of the radar cloud layer data adjacent in time sequence with the corresponding time series characteristics; a correction module, configured to input the extrapolated radar cloud image data and the historical radar cloud image data into a pre-trained extrapolated cloud image correction model to obtain corrected radar cloud image data; wherein the extrapolated cloud image correction model uses the change characteristics of the extrapolated radar cloud image data as a guiding feature and corrects the extrapolated radar cloud image data based on the cloud cluster characteristics of the historical radar cloud image data; A meteorological characteristic determination module, configured to determine the meteorological characteristics of the target photovoltaic station based on the corrected radar cloud image data; A photovoltaic power prediction module is used to predict the photovoltaic power of the target photovoltaic station based on the meteorological characteristics to obtain a photovoltaic power prediction result; The target photovoltaic station includes a plurality of sub-distributed photovoltaic stations, and the photovoltaic power prediction module is further used for: Performing photovoltaic power forecasting on the target photovoltaic station based on the meteorological characteristics to obtain a power forecast result to be corrected; In the corrected radar cloud image data, obtaining a target combined reflectivity greater than a preset combined reflectivity; Acquire the photovoltaic power prediction result that is spatially and temporally matched with the target combined reflectivity as the power prediction result to be corrected; Obtaining a correlation coefficient of the sub-distributed photovoltaic station corresponding to the power prediction result to be corrected when performing geographical location matching as an influence coefficient; If there is a historical power prediction result, obtain the historical predicted power and actual power at the end time of the historical power prediction result; The power prediction result to be corrected is corrected based on the magnitude of the historical predicted power relative to the actual power and the influence coefficient to obtain a photovoltaic power prediction result.

8. A computer device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the photovoltaic ultra-short-term power prediction method according to any one of claims 1 to 6 by executing the computer instructions.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the photovoltaic ultra-short-term power prediction method according to any one of claims 1 to 6.

Images