An ultra-short-term photovoltaic power prediction method, system, device and medium

Abstract

The present application relates to the technical field of photovoltaic power generation, and particularly relates to an ultra-short-term photovoltaic power prediction method, system, device and medium. The present application fuses two types of heterogeneous observation data of all-sky imagers and geostationary satellites, respectively calculates cloud transmittance time series of different historical windows, utilizes the real-time and non-lagged characteristics of all-sky imagers to capture minute-level cloud changes, and utilizes the wide-range observation capability of geostationary satellites to obtain the macro trend of cloud movement, thereby effectively making up for the deficiencies of a single data source in timeliness and spatial coverage. On this basis, a pre-trained time series prediction model is used to estimate the cloud transmittance of a future target period, and is further converted into ground surface irradiance and photovoltaic power, which can significantly improve the accuracy and stability of ultra-short-term power prediction, and provide more reliable technical support for power grid dispatching and photovoltaic power station operation.

Description

Technical Field

[0001] This invention relates to the field of photovoltaic power generation technology, and in particular to an ultra-short-term photovoltaic power forecasting method, system, equipment and medium. Background Technology

[0002] Photovoltaic power forecasting refers to the technology of predicting the output power of photovoltaic power plants over a future period. Its main function is to help grid dispatching departments anticipate the fluctuations in photovoltaic power generation, rationally schedule the start-up and shutdown of conventional units, the charging and discharging of energy storage, and reserve capacity, ensuring the safe and stable operation of the power system, while also enhancing the market competitiveness of photovoltaic power. Because solar radiation is significantly affected by meteorological factors such as cloud movement and aerosol changes, ultra-short-term power forecasting faces challenges such as the difficulty in balancing data timeliness and spatial coverage, and the difficulty in predicting cloud processes, making it a current research hotspot and challenge.

[0003] Currently, photovoltaic power forecasting methods are mainly divided into three categories: physical methods based on numerical weather prediction, statistical methods based on historical data, and intelligent methods based on machine learning. Physical methods predict weather elements by solving atmospheric dynamic equations, but their accuracy in ultra-short-term forecasts is limited by computational resources and uncertainties in initial conditions. Statistical methods rely on historical power and meteorological data to build empirical models, but their generalization ability is weak under abrupt weather changes. Machine learning methods can uncover nonlinear relationships in data, but they require high-quality and high-quantity training samples. These methods generally suffer from common problems such as single data sources, mismatched spatiotemporal resolution, and low forecast accuracy, making it difficult to simultaneously meet the comprehensive requirements of timeliness, accuracy, and stability for ultra-short-term forecasts. Therefore, there is an urgent need to develop a new forecasting technology that integrates multi-source data. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide an ultra-short-term photovoltaic power forecasting method, system, device and medium to solve the above-mentioned technical problem.

[0005] The technical solution of this invention to solve the above-mentioned technical problems is as follows: A method for ultra-short-term photovoltaic power forecasting, comprising: acquiring all-sky imager observation data, ground meteorological station observation data, geostationary satellite observation data, reanalysis data, and surface albedo data for a target site; calculating the cloud transmittance time series within a first historical window prior to the current time based on the all-sky imager observation data; calculating the cloud transmittance time series within a second historical window prior to the current time based on the geostationary satellite observation data, the reanalysis data, and the surface albedo data; predicting the cloud transmittance time series for a future target period using a pre-trained time series prediction model based on the cloud transmittance time series within the first historical window and the cloud transmittance time series within the second historical window, and calculating the surface irradiance forecast sequence for the future target period based on the cloud transmittance time series for the future target period; and calculating the photovoltaic power forecast sequence for the future target period based on the surface irradiance forecast sequence for the future target period and the meteorological station observation data.

[0006] The beneficial effects of this invention are as follows: By fusing heterogeneous observation data from both all-sky imagers and geostationary satellites, this invention calculates cloud transmittance time series for different historical windows. It leverages the real-time, lag-free characteristics of all-sky imagers to capture minute-level cloud changes, while utilizing the wide-area observation capabilities of geostationary satellites to obtain macroscopic trends in cloud movement. This effectively compensates for the shortcomings of single data sources in terms of timeliness and spatial coverage. Based on this, a pre-trained time-series prediction model is used to estimate cloud transmittance for future target periods, and further converted into surface irradiance and photovoltaic power. This significantly improves the accuracy and stability of ultra-short-term power forecasts, providing more reliable technical support for grid dispatching and photovoltaic power plant operation and maintenance.

[0007] Based on the above technical solution, the present invention can be further improved as follows.

[0008] Furthermore, the step of calculating the cloud transmittance time series within the first historical window prior to the current moment based on the all-sky imager observation data includes: extracting the cloud cover at each time step within the first historical window using the red-blue channel ratio method based on the all-sky imager observation data; calculating the cloud transmittance at each time step within the first historical window based on the cloud cover at each time step within the first historical window and a pre-fitted localization coefficient; and obtaining the cloud transmittance time series within the first historical window based on the cloud transmittance at each time step within the first historical window.

[0009] Further, the step of calculating the cloud transmittance time series within the second historical window prior to the current time based on the geostationary satellite observation data, the reanalysis data, and the surface albedo data includes: calculating the all-weather surface solar radiation transmittance at each time step within the second historical window based on the geostationary satellite observation data, the reanalysis data, and the surface albedo data; calculating the surface solar radiance at each time step within the second historical window based on the all-weather surface solar radiation transmittance at each time step within the second historical window, as well as preset solar constants, Earth orbital eccentricity correction coefficients, and solar zenith angle cosines; calculating the clear-sky irradiance at each time step within the second historical window based on the geostationary satellite observation data and the reanalysis data; calculating the cloud transmittance at each time step within the second historical window based on the surface solar radiance and clear-sky irradiance; and obtaining the cloud transmittance time series within the second historical window based on the cloud transmittance at each time step within the second historical window.

[0010] Furthermore, the step of calculating the all-weather surface solar radiation transmittance at each time step within the second historical window based on the geostationary satellite observation data, the reanalysis data, and the surface albedo data includes: calculating the clear sky transmittance, water cloud transmittance, and ice cloud transmittance at each time step within the second historical window based on the geostationary satellite observation data, the reanalysis data, and the surface albedo data; and for each time step within the second historical window, performing a weighted calculation on the clear sky transmittance, water cloud transmittance, and ice cloud transmittance of the time step to obtain the all-weather surface solar radiation transmittance of the time step.

[0011] Furthermore, the pre-trained time-series prediction model includes a first prediction model and a second prediction model; the step of predicting the cloud transmittance time series for a future target period using the pre-trained time-series prediction model based on the cloud transmittance time series within the first historical window and the cloud transmittance time series within the second historical window includes: predicting the cloud transmittance time series for a first future period using the first prediction model based on the cloud transmittance time series within the first historical window; predicting the cloud transmittance time series for a second future period using the second prediction model based on the cloud transmittance time series within the second historical window; and obtaining the cloud transmittance time series for the future target period based on the cloud transmittance time series for the first future period and the cloud transmittance time series for the second future period.

[0012] Furthermore, both the first prediction model and the second prediction model are trained from the Cross ViViT model.

[0013] Furthermore, the step of calculating the photovoltaic power forecast sequence for the future target period based on the surface irradiance forecast sequence for the future target period and the meteorological station observation data includes: calculating the photovoltaic panel temperature at each time step within the future target period based on the surface irradiance forecast sequence for the future target period and the meteorological station observation data; and calculating the photovoltaic power forecast sequence for the future target period based on the surface irradiance forecast sequence for the future target period, the photovoltaic panel temperature at each time step within the future target period, and preset photovoltaic panel parameters.

[0014] To address the aforementioned technical problems, the present invention also provides an ultra-short-term photovoltaic power forecasting system, comprising: The data acquisition module is used to acquire all-sky imager observation data, ground meteorological station observation data, geostationary satellite observation data, reanalysis data, and surface albedo data for the target site; The first calculation module is used to calculate the cloud transmittance time series within the first historical window before the current moment based on the observation data of the all-sky imager. The second calculation module is used to calculate the cloud transmittance time series within the second historical window before the current moment based on the geostationary satellite observation data, the reanalysis data, and the surface albedo data. The model prediction module is used to predict the cloud transmittance time series for the future target period based on the cloud transmittance time series in the first historical window and the cloud transmittance time series in the second historical window using a pre-trained time series prediction model, and to calculate the surface irradiance forecast sequence for the future target period based on the cloud transmittance time series for the future target period. The power forecast module is used to calculate the photovoltaic power forecast sequence for the future target period based on the surface irradiance forecast sequence for the future target period and the meteorological station observation data.

[0015] To address the aforementioned technical problems, the present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the ultra-short-term photovoltaic power forecasting method as described above.

[0016] To address the aforementioned technical problems, the present invention also provides a non-transitory computer-readable storage medium storing computer instructions for causing a computer to execute the ultra-short-term photovoltaic power forecasting method described above. Attached Figure Description

[0017] Figure 1 This is a flowchart of an ultra-short-term photovoltaic power forecasting method according to the present invention; Figure 2This is a schematic diagram of an ultra-short-term photovoltaic power forecasting system according to the present invention; Figure 3 This is a schematic diagram of an electronic device according to the present invention. Detailed Implementation

[0018] The principles and features of the present invention are described below. The examples given are only for explaining the present invention and are not intended to limit the scope of the present invention.

[0019] As mentioned earlier, the output power of photovoltaic (PV) power generation depends on solar radiation, which is significantly affected by weather conditions, exhibiting natural fluctuations and intermittency. Among these factors, cloud movement is the most frequent and has the greatest impact; when PV is connected to the grid on a large scale, these fluctuations can trigger severe grid oscillations. Therefore, high-precision ultra-short-term PV power forecasts are crucial for the safe operation of the power grid.

[0020] However, single data sources all have inherent limitations in ultra-short-term forecasting: all-sky imagers can observe cloud information in real time, but their observation range is small and their lead time is short; geostationary satellites have a large observation range, but due to data latency, they are usually lagging by about 45 minutes, making them difficult to use directly for short-term forecasting. Numerical weather prediction, while having clear physical meaning, has low spatiotemporal resolution, limited ability to capture rapidly moving cloud changes, and model warm-up issues make it difficult to achieve ultra-short-term forecasting. Traditional machine learning methods lack physical constraints and have poor portability between different photovoltaic power plants.

[0021] Example 1 Based on this, such as Figure 1 As shown, this embodiment provides a method for ultra-short-term photovoltaic power forecasting, including: S101. Acquire all-sky imager observation data, ground meteorological station observation data, geostationary satellite observation data, reanalysis data, and surface albedo data for the target site.

[0022] S102. Based on the observation data from the all-sky imager, calculate the cloud transmittance time series within the first historical window prior to the current moment.

[0023] S103. Based on the geostationary satellite observation data, the reanalysis data, and the surface albedo data, calculate the cloud transmittance time series within the second historical window prior to the current moment.

[0024] S104. Based on the cloud transmittance time series in the first historical window and the cloud transmittance time series in the second historical window, use a pre-trained time series prediction model to predict the cloud transmittance time series for the future target period, and calculate the surface irradiance forecast sequence for the future target period based on the cloud transmittance time series for the future target period.

[0025] S105. Calculate the photovoltaic power forecast sequence for the future target period based on the surface irradiance forecast sequence for the future target period and the meteorological station observation data.

[0026] This method integrates heterogeneous observation data from both all-sky imagers and geostationary satellites to calculate cloud transmittance time series for different historical windows. It leverages the real-time, lag-free nature of all-sky imagers to capture minute-level cloud changes, while utilizing the wide-area observation capabilities of geostationary satellites to obtain macroscopic trends in cloud movement. This effectively compensates for the limitations of single data sources in terms of timeliness and spatial coverage. Based on this, a pre-trained time-series prediction model is used to estimate cloud transmittance for future target periods, which is then converted into surface irradiance and photovoltaic power. This significantly improves the accuracy and stability of ultra-short-term power forecasts, providing more reliable technical support for grid dispatching and photovoltaic power plant operation and maintenance.

[0027] Specifically, all-sky imager observation data refers to the image sequences and derived parameters collected by all-sky imagers installed at photovoltaic power plants. All-sky imagers can continuously capture visible light or infrared images covering the entire sky, typically possessing a near 180° field of view. Equipped with fisheye lenses or specialized optical systems, they can cover the entire sky from the horizon to the zenith in a single frame. They are commonly used for continuous monitoring in meteorological observations (cloud cover, cloud type, sky radiation), astronomy (night sky brightness, star background), and space physics and atmospheric science research (such as upper atmospheric luminescence, auroral structure, gravity waves, etc.). Time-series analysis of all-sky images helps study the evolution of sky conditions and related physical processes. Image processing algorithms (such as the red-blue channel ratio method) can further extract cloud cover, cloud movement speed, and cloud movement direction.

[0028] Ground-based meteorological station observation data refers to near-surface meteorological elements collected in real time by on-site meteorological stations at photovoltaic power plants, such as air temperature, relative humidity, surface air pressure, wind speed, and visibility. Geostationary satellite observation data refers to gridded quantitative products provided by geostationary orbit meteorological satellites (such as Himawari-8 / 9), mainly including cloud optical thickness and aerosol optical thickness.

[0029] Himawari-8 / 9 are the Japan Meteorological Agency's next-generation geostationary weather satellites, located at 140.7°E, approximately 36,000 km above the equator. Equipped with an AHI sensor, they have 16 spectral bands (0.47-13.3 μm) and can monitor meteorological parameters such as clouds, aerosols, and radiation in the East Asia-West Pacific region with a temporal resolution of 10 minutes and a spatial resolution of 0.5-2 km. They are widely used in weather forecasting, climate research, and photovoltaic power generation prediction.

[0030] In this embodiment, ERA5 reanalysis data provided by ECMWF are used, specifically including surface air pressure, total column water vapor content, and total column ozone content. Surface albedo data refers to the reflectivity of the Earth's surface to solar shortwave radiation, and the surface albedo retrieved from the MODIS satellite (MOD43C3) is used.

[0031] After all data undergoes quality control and spatial interpolation, it proceeds to the next steps. Quality control primarily includes outlier removal and time alignment.

[0032] Optionally, in an embodiment, the step of calculating the cloud transmittance time series within a first historical window prior to the current moment based on the all-sky imager observation data includes: extracting the cloud cover at each time step within the first historical window using the red-blue channel ratio method based on the all-sky imager observation data; calculating the cloud transmittance at each time step within the first historical window based on the cloud cover at each time step within the first historical window and a pre-fitted localization coefficient; and obtaining the cloud transmittance time series within the first historical window based on the cloud transmittance at each time step within the first historical window.

[0033] In this embodiment, the first historical window is one hour prior to the current time. Specifically, cloud cover is extracted using all-sky imager images and the red-blue channel ratio method. First, calculate the red / blue ratio for each pixel. : ; in, This represents the pixel value of the red channel. This represents the pixel value of the blue channel.

[0034] Set a threshold, if If the value is less than a threshold, it is considered a cloud, which can be specifically represented as: .

[0035] Cloud cover It is calculated using the following formula: ; in, For cloud pixels, This represents the total number of valid pixels.

[0036] The attenuation of solar radiation by clouds is measured by cloud transmittance. Characterization: ; Where k is the fitting coefficient, which can be obtained by regression analysis using historical all-sky imager cloud cover and ground radiation to fit a k suitable for the local area.

[0037] Calculate the radiation from the top of the atmosphere: ; in, The solar constant is (1367±7W / m) 2 ), This is the eccentricity correction factor for Earth's orbit (current value approximately 0.0167). It is the cosine of the solar zenith angle, which can be determined given the location.

[0038] By correcting atmospheric transparency using weather station data, clear-sky irradiance is obtained: ; in, , , , , The transmittance values ​​are, in order, ozone absorption transmittance, water vapor absorption transmittance, homogeneous gas mixture transmittance, Rayleigh scattering transmittance, and aerosol transmittance. The water vapor absorption transmittance is calculated from the air temperature and relative humidity observed at the meteorological station.

[0039] The shortwave radiation from clouds and sky is calculated using the following formula: .

[0040] Therefore, the irradiance at each time step within the first historical window can be obtained. Irradiance is the radiant energy that is perpendicularly incident on a unit area surface per unit time. In the photovoltaic field, it specifically refers to solar irradiance, which is the most critical meteorological factor determining the power generation of a photovoltaic power plant.

[0041] Optionally, in an embodiment, calculating the cloud transmittance time series within a second historical window prior to the current time based on the geostationary satellite observation data, the reanalysis data, and the surface albedo data includes: calculating the all-weather surface solar radiation transmittance at each time step within the second historical window based on the geostationary satellite observation data, the reanalysis data, and the surface albedo data; calculating the surface solar radiance at each time step within the second historical window based on the all-weather surface solar radiation transmittance at each time step within the second historical window, as well as preset solar constants, Earth orbital eccentricity correction coefficients, and solar zenith angle cosines; calculating the clear-sky irradiance at each time step within the second historical window based on the geostationary satellite observation data and the reanalysis data; calculating the cloud transmittance at each time step within the second historical window based on the surface solar radiance and clear-sky irradiance; and obtaining the cloud transmittance time series within the second historical window based on the cloud transmittance at each time step within the second historical window.

[0042] In this embodiment, the second historical window is the interval of 1 to 4 hours prior to the current time, totaling 3 hours. Specifically, the surface solar radiance R is expressed as follows: ; in, The solar constant is (1367±7W / m) 2 ), This is the eccentricity correction factor for Earth's orbit (current value approximately 0.0167). It is the cosine of the solar zenith angle, which can be determined given the location.

[0043] Optionally, in an embodiment, calculating the all-weather surface solar radiation transmittance at each time step within the second historical window based on the geostationary satellite observation data, the reanalysis data, and the surface albedo data includes: calculating the clear sky transmittance, water cloud transmittance, and ice cloud transmittance at each time step within the second historical window based on the geostationary satellite observation data, the reanalysis data, and the surface albedo data; and for each time step within the second historical window, performing a weighted calculation on the clear sky transmittance, water cloud transmittance, and ice cloud transmittance of the time step to obtain the all-weather surface solar radiation transmittance of the time step.

[0044] T represents the transmittance of solar radiation on the Earth's surface under all weather conditions, including clear sky transmittance on sunny days. Water cloud transmittance when there are clouds and ice cloud transmittance , can be represented as: ; in, and These represent the proportions of water clouds and ice clouds, respectively. and They can be represented as: ; ; Where A represents the surface albedo, provided by MODIS albedo products. and These are the reflectance values ​​for water clouds and ice clouds, respectively, and the default values ​​are used.

[0045] Solar radiation transmittance on Earth's surface under clear weather conditions It is mainly related to atmospheric composition and can be given by the following formula. ; ; ; Among them, the transmittance caused by aerosol attenuation Transmittance caused by absorption Transmittance caused by scattering The relationship between them can be represented as: ; in, This represents the aerosol single-scattering albedo.

[0046] The downward transmittance is caused by Rayleigh single scattering. The downward transmittance caused by aerosol single scattering is given by the following expressions: ; .

[0047] Here, since Rayleigh scattering is symmetrical between the upward and downward hemispheres, a coefficient of 0.5 is used to represent the downward scattering component. However, aerosol scattering is asymmetrical, so the coefficient for the downward component is... , It is the solar zenith angle and the asymmetry factor (g or A function is given by the following expression: ; ; .

[0048] In cloudy conditions, transmittance mainly depends on the clouds themselves, specifically the transmittance of ice clouds or water clouds. or () can be given in the following ways.

[0049] In cloudy conditions, the shortwave band is divided into two segments: band a (0.2-0.7μm) and band b (0.7-5.0μm). The transmittance of ice clouds or water clouds will be parameterized separately in different bands.

[0050] Transmittance of ice clouds in band a Can be parameterized as cloud optical thickness in the visible light band Ozone content Cosine of the solar zenith angle The function of water cloud transmittance Can be parameterized as cloud optical thickness in the visible light band Cosine of the solar zenith angle The function. Specifically as follows: like ,but: ; like ,but: .

[0051] like and ,but: ; like ,but: .

[0052] Transmittance of ice clouds and water clouds in band b ( and All of these can be parameterized as the optical thickness of clouds in the visible light band. Water vapor content Cosine of the solar zenith angle The function. Specifically as follows: like ,but: ; like ,but: .

[0053] like and ,but: ; like ,but: .

[0054] Optionally, in an embodiment, the pre-trained time-series prediction model includes a first prediction model and a second prediction model; the step of predicting the cloud transmittance time series for a future target period using the pre-trained time-series prediction model based on the cloud transmittance time series within the first historical window and the cloud transmittance time series within the second historical window includes: predicting the cloud transmittance time series for a first future period using the first prediction model based on the cloud transmittance time series within the first historical window; predicting the cloud transmittance time series for a second future period using the second prediction model based on the cloud transmittance time series within the second historical window; and obtaining the cloud transmittance time series for the future target period based on the cloud transmittance time series for the first future period and the cloud transmittance time series for the second future period.

[0055] Optionally, in an embodiment, both the first prediction model and the second prediction model are trained using the Cross ViViT (Cross Video-Vision Transformer) model.

[0056] The first and second prediction models are for 0-1 hour and 1-4 hour forecasts, respectively. The first model is input with the cloud transmittance sequence of the past 1 hour, and the second model is input with the cloud transmittance sequence of the past 3 hours.

[0057] Both the first and second prediction models are built on the Cross ViViT model and have the same model structure. For model training, one year's worth of historical data can be used, with mean squared error as the loss function, and the Adam optimizer employed. For example, N×N×t time-series cloud transmittance data blocks for daytime conditions can be obtained from satellite remote sensing cloud images over a historical period as samples, with each sample covering N×N grid points centered on the forecast station, to train the model.

[0058] By inputting the cloud transmittance sequence before the forecast time into the model, the cloud transmittance sequence for multiple time steps in the future can be obtained through image block embedding, spatial context encoding, temporal feature encoding, cross-modal cross-fusion, and temporal decoding prediction.

[0059] Finally, multiplying the cloud transmittance by the clear-sky solar radiation at the corresponding time (calculated using the solar altitude angle formula) yields the predicted ground solar irradiance value (unit: W / m²). 2 ).

[0060] Cross ViViT is a novel architecture combining numerical time-series data, text, and images, primarily used for time-series forecasting tasks. The Transformer is a deep learning model architecture for sequence-to-series tasks, introducing a self-attention mechanism that allows the model to consider all positions in the input sequence simultaneously, thus better handling long-term series or spatial data. The input part of the Cross ViViT model contains two Transformer modules: a Vision Transformer for processing images and extracting spatial information, and a Temporal Transformer for processing time-series data and extracting temporal information. The two inputs are fused by the Cross Transformer and then passed through a Temporal Transformer and a Multilayer Perceptron (MLP) module to predict future time-series changes. This algorithm uses the Cross Video-Vision Transformer model, combined with cloud parameters from Himawari-8 (H8) and atmospheric parameters from multi-source data, to obtain cloud transmittance at the forecast start time, thereby improving the accuracy of short-term (1-4 hours) radiation forecasts.

[0061] Optionally, in an embodiment, calculating the photovoltaic power forecast sequence for the future target period based on the surface irradiance forecast sequence for the future target period and the meteorological station observation data includes: calculating the photovoltaic panel temperature at each time step within the future target period based on the surface irradiance forecast sequence for the future target period and the meteorological station observation data; and calculating the photovoltaic power forecast sequence for the future target period based on the surface irradiance forecast sequence for the future target period, the photovoltaic panel temperature at each time step within the future target period, and preset photovoltaic panel parameters.

[0062] Photovoltaic panel temperature refers to the actual temperature of a solar cell module during operation, which significantly affects photovoltaic power generation. Generally, under the same irradiance conditions, an increase in temperature leads to a decrease in the output voltage of the photovoltaic module, thereby reducing the output power. The typical temperature coefficient of polycrystalline or monocrystalline silicon modules is approximately −0.3% to −0.5% / ℃, meaning that for every 1℃ increase in temperature, the power decreases by approximately 0.3% to 0.5%. In actual outdoor operation, the temperature of photovoltaic panels is usually about 20℃ higher than the ambient temperature: for example, when the ambient temperature is 25℃, the surface temperature of the module may reach 45~55℃, and under strong sunlight and weak wind conditions in summer, it can even reach 60~70℃.

[0063] The power output is affected by factors such as irradiance, module temperature, tilt angle, and shading. The calculation formula is as follows: P = η × A × G; Where η is the conversion efficiency, A is the component area, and G is the irradiance.

[0064] Conversion efficiency is closely related to photovoltaic (PV) panel temperature. Specifically, an increase in PV panel temperature leads to a decrease in power output. For every 1°C increase in PST (Power Stem Temperature) compared to standard test conditions (25°C), the photoelectric conversion efficiency decreases by approximately 0.3%-0.6%. In most PV power plants, the actual PST during peak midday power generation in summer is 15-25°C higher than standard test conditions, corresponding to an instantaneous power loss of approximately 4.5%-15%. This efficiency deviation directly impacts power generation estimation. Therefore, the more accurate the PV panel temperature reading, the more accurate the power output estimation.

[0065] In this embodiment, a photovoltaic panel temperature calculation based on an energy balance model is used to replace the traditional empirical photovoltaic panel temperature algorithm. A steady-state thermal balance equation is established between incident solar radiation absorption, convective heat transfer, and long-wave radiation heat dissipation. This equation incorporates the predicted solar radiation values. With ambient temperature Wind speed Combined, calculate the photovoltaic panel temperature using the following formula. : .

[0066] Overall, this method improves the robustness of photovoltaic panel temperature calculation through an energy balance model. Furthermore, the increased accuracy of the input irradiance leads to improved accuracy in photovoltaic panel temperature calculation. These two factors combined significantly enhance the accuracy of the irradiance-to-power conversion. Finally, based on the predicted irradiance and the corrected photoelectric conversion efficiency, a 0-4 hour ultra-short-term photovoltaic power forecast is output.

[0067] Example 2 like Figure 2 As shown, this embodiment provides an ultra-short-term photovoltaic power forecasting system 200, including: The data acquisition module 201 is used to acquire all-sky imager observation data, ground meteorological station observation data, geostationary satellite observation data, reanalysis data and surface albedo data for the target site; The first calculation module 202 is used to calculate the cloud transmittance time series within the first historical window before the current moment based on the observation data of the all-sky imager. The second calculation module 203 is used to calculate the cloud transmittance time series within the second historical window before the current moment based on the geostationary satellite observation data, the reanalysis data and the surface albedo data; The model prediction module 204 is used to predict the cloud transmittance time series of the future target period based on the cloud transmittance time series in the first historical window and the cloud transmittance time series in the second historical window using a pre-trained time series prediction model, and to calculate the surface irradiance forecast sequence of the future target period based on the cloud transmittance time series of the future target period. The power forecast module 205 is used to calculate the photovoltaic power forecast sequence for the future target period based on the surface irradiance forecast sequence for the future target period and the meteorological station observation data.

[0068] Optionally, in an embodiment, the first computing module 202 includes: The cloud cover calculation unit is used to extract the cloud cover at each time step within the first historical window based on the observation data of the all-sky imager using the red-blue channel ratio method. The first cloud transmittance calculation unit is used to calculate the cloud transmittance of each time step within the first historical window based on the cloud amount at each time step within the first historical window and the pre-fitted localization coefficient. The first sequence generation unit is used to obtain the cloud transmittance time series within the first historical window based on the cloud transmittance at each time step within the first historical window.

[0069] Optionally, in an embodiment, the second computing module 203 includes: The surface solar radiation transmittance calculation unit is used to calculate the all-weather surface solar radiation transmittance at each time step within the second historical window based on the geostationary satellite observation data, the reanalysis data, and the surface albedo data. The surface solar radiance calculation unit is used to calculate the surface solar radiance at each time step within the second historical window based on the all-weather surface solar radiance transmittance at each time step within the second historical window, as well as the preset solar constant, Earth orbit eccentricity correction coefficient, and solar zenith angle cosine. The clear sky irradiance calculation unit is used to calculate the clear sky irradiance at each time step within the second historical window based on the geostationary satellite observation data and the reanalysis data. The second cloud transmittance calculation unit is used to calculate the cloud transmittance at each time step within the second historical window based on the surface solar irradiance and clear sky irradiance at each time step within the second historical window. The second sequence generation unit is used to obtain the cloud transmittance time series within the second historical window based on the cloud transmittance at each time step within the second historical window.

[0070] Optionally, in an embodiment, the surface solar radiation transmittance calculation unit includes: The transmittance calculation subunit is used to calculate the clear sky transmittance, water cloud transmittance, and ice cloud transmittance at each time step within the second historical window based on the geostationary satellite observation data, the reanalysis data, and the surface albedo data. The weighted calculation subunit is used to perform weighted calculations on the clear sky transmittance, water cloud transmittance, and ice cloud transmittance of each time step within the second historical window to obtain the all-weather surface solar radiation transmittance of the time step.

[0071] Optionally, in an embodiment, the pre-trained time series prediction model includes a first prediction model and a second prediction model; the model prediction module 204 includes: The first prediction unit is used to predict the cloud transmittance time series for a first future period based on the cloud transmittance time series within the first historical window and using the first prediction model. The second prediction unit is used to predict the cloud transmittance time series for a second future period based on the cloud transmittance time series within the second historical window and using the second prediction model. The result generation unit is used to obtain the cloud transmittance time series of the future target period based on the cloud transmittance time series of the first future period and the cloud transmittance time series of the second future period.

[0072] Optionally, in an embodiment, both the first prediction model and the second prediction model are trained from the Cross ViViT model.

[0073] Optionally, in an embodiment, the power prediction module 205 includes: The photovoltaic panel temperature calculation unit is used to calculate the photovoltaic panel temperature at each time step within the future target period based on the surface irradiance forecast sequence for the future target period and the meteorological station observation data. The power calculation unit is used to calculate the photovoltaic power forecast sequence for the future target period based on the surface irradiance forecast sequence for the future target period, the photovoltaic panel temperature at each time step within the future target period, and the preset photovoltaic panel parameters.

[0074] In some embodiments, the ultra-short-term photovoltaic power forecasting system 200 of the present invention can be implemented in a combination of hardware and software. As an example, the ultra-short-term photovoltaic power forecasting system 200 of the present invention can be a processor in the form of a hardware decoding processor, which is programmed to execute the ultra-short-term photovoltaic power forecasting method of the present invention. For example, the processor in the form of a hardware decoding processor can be one or more application-specific integrated circuits (ASICs), DSPs, programmable logic devices (PLDs), complex programmable logic devices (CPLDs), field-programmable gate arrays (FPGAs), or other electronic components.

[0075] The modules described in the embodiments of this invention can be implemented in software or hardware. The names of the modules are not, in some cases, limiting the scope of the module itself.

[0076] Example 3 like Figure 3 As shown, this embodiment provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements an ultra-short-term photovoltaic power forecasting method as described in Embodiment 1.

[0077] In other words, an electronic device according to an embodiment of the present invention may include, but is not limited to, a processor and a memory; the memory is used to store a computer program; the processor is used to execute an ultra-short-term photovoltaic power forecasting method shown in any embodiment of the present invention by calling the computer program.

[0078] In one alternative embodiment, an electronic device is provided. Figure 3The illustrated electronic device 300 includes a processor 301 and a memory 303. The processor 301 and the memory 303 are connected, for example, via a bus 302. Optionally, the electronic device 300 may further include a transceiver 304, which can be used for data interaction between the electronic device and other electronic devices, such as sending and / or receiving data. It should be noted that in practical applications, the transceiver 304 is not limited to one type, and the structure of the electronic device 300 does not constitute a limitation on the embodiments of the present invention.

[0079] Processor 301 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It can implement or execute the various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this invention. Processor 301 may also be a combination that implements computational functions, such as including one or more microprocessor combinations, a combination of a DSP and a microprocessor, etc.

[0080] Bus 302 may include a path for transmitting information between the aforementioned components. Bus 302 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. Bus 302 can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 3 The bus 302 is represented by only one thick line, but this does not mean that there is only one bus or one type of bus.

[0081] The memory 303 may be a ROM (Read Only Memory) or other type of static storage device capable of storing static information and instructions, RAM (Random Access Memory) or other type of dynamic storage device capable of storing information and instructions, or an EEPROM (Electrically Erasable Programmable Read Only Memory), CD-ROM (Compact Disc Read Only Memory) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto.

[0082] The memory 303 is used to store application code (computer program) for executing the present invention, and its execution is controlled by the processor 301. The processor 301 is used to execute the application code stored in the memory 303 to implement the content shown in the foregoing method embodiments.

[0083] Among them, electronic devices can also be terminal devices, which can be any device that can install applications, including at least one of smartphones, tablets, laptops, desktop computers, smart speakers, smartwatches, smart TVs, and smart in-vehicle devices.

[0084] It should be noted that, Figure 3 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of use of the embodiments of the present invention.

[0085] Example 4 This embodiment provides a non-transitory computer-readable storage medium that stores computer instructions for causing a computer to execute an ultra-short-term photovoltaic power forecasting method as described in Embodiment 1.

[0086] Alternatively, the computer-readable storage medium may be a read-only memory (ROM), a random access memory (RAM), a compact disc read-only memory (CD-ROM), magnetic tape, a floppy disk, and an optical data storage device, etc.

[0087] In an exemplary embodiment, a computer program product or computer program is also provided, which includes computer instructions stored in a computer-readable storage medium. A processor of an electronic device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the electronic device to perform the aforementioned ultra-short-term photovoltaic power forecasting method.

[0088] Computer program code for performing the operations of this invention can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as C or similar 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 can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0089] It should be understood that the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of methods and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0090] The computer-readable storage medium provided in this invention can be, but is not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: 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 (EEPROM 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 invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0091] The aforementioned computer-readable storage medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to perform the method shown in the above embodiments.

[0092] The above description is merely a preferred embodiment of the present invention and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of disclosure in this invention is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-disclosed concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features with similar functions disclosed in this invention.

[0093] It should be noted that the terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects and represent a limitation on a specific order or sequence. Where appropriate, the order of use for similar objects can be interchanged so that the embodiments of this application described herein can be implemented in an order other than that shown or described.

[0094] Those skilled in the art will recognize that this invention can be implemented as a system, method, or computer program product. Therefore, this invention can be specifically implemented in the following forms: it can be entirely hardware, entirely software (including firmware, resident software, microcode, etc.), or a combination of hardware and software, generally referred to herein as a "circuit," "module," or "system." Furthermore, in some embodiments, this invention can also be implemented as a computer program product contained in one or more computer-readable media, which includes computer-readable program code.

[0095] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A method for ultra-short-term photovoltaic power forecasting, characterized in that, include: Acquire all-sky imager observation data, ground meteorological station observation data, geostationary satellite observation data, reanalysis data, and surface albedo data for the target site; Based on the observation data from the all-sky imager, calculate the cloud transmittance time series within the first historical window prior to the current moment; Based on the geostationary satellite observation data, the reanalysis data, and the surface albedo data, calculate the cloud transmittance time series within the second historical window prior to the current moment; Based on the cloud transmittance time series within the first historical window and the cloud transmittance time series within the second historical window, a pre-trained time series prediction model is used to predict the cloud transmittance time series for the future target period, and the surface irradiance forecast sequence for the future target period is calculated based on the cloud transmittance time series for the future target period. Based on the surface irradiance forecast sequence for the future target period and the meteorological station observation data, calculate the photovoltaic power forecast sequence for the future target period.

2. The ultra-short-term photovoltaic power forecasting method according to claim 1, characterized in that, The step of calculating the cloud transmittance time series within the first historical window prior to the current moment based on the all-sky imager observation data includes: Based on the observation data from the all-sky imager, the cloud cover at each time step within the first historical window is extracted using the red-blue channel ratio method. Based on the cloud cover at each time step within the first historical window and the pre-fitted localization coefficient, calculate the cloud transmittance at each time step within the first historical window. Based on the cloud transmittance at each time step within the first historical window, the cloud transmittance time series within the first historical window is obtained.

3. The ultra-short-term photovoltaic power forecasting method according to claim 1, characterized in that, The step of calculating the cloud transmittance time series within the second historical window prior to the current moment based on the geostationary satellite observation data, the reanalysis data, and the surface albedo data includes: Based on the geostationary satellite observation data, the reanalysis data, and the surface albedo data, calculate the all-weather surface solar radiation transmittance at each time step within the second historical window; Based on the all-weather surface solar radiation transmittance at each time step within the second historical window, as well as the preset solar constant, Earth orbital eccentricity correction coefficient, and solar zenith angle cosine, the surface solar radiance at each time step within the second historical window is calculated. Based on the geostationary satellite observation data and the reanalysis data, calculate the clear sky irradiance at each time step within the second historical window; Based on the surface solar irradiance and clear sky irradiance at each time step within the second historical window, calculate the cloud transmittance at each time step within the second historical window. Based on the cloud transmittance at each time step within the second historical window, the cloud transmittance time series within the second historical window is obtained.

4. The ultra-short-term photovoltaic power forecasting method according to claim 3, characterized in that, The step of calculating the all-weather surface solar radiation transmittance at each time step within the second historical window based on the geostationary satellite observation data, the reanalysis data, and the surface albedo data includes: Based on the geostationary satellite observation data, the reanalysis data, and the surface albedo data, calculate the clear sky transmittance, water cloud transmittance, and ice cloud transmittance at each time step within the second historical window; For each time step within the second historical window, the clear sky transmittance, water cloud transmittance, and ice cloud transmittance of the time step are weighted and calculated to obtain the all-weather surface solar radiation transmittance of the time step.

5. The ultra-short-term photovoltaic power forecasting method according to claim 1, characterized in that, The pre-trained time-series prediction model includes a first prediction model and a second prediction model; the step of predicting the cloud transmittance time series for a future target period using the pre-trained time-series prediction model based on the cloud transmittance time series within the first historical window and the cloud transmittance time series within the second historical window includes: Based on the cloud transmittance time series within the first historical window, the cloud transmittance time series for the first future period is predicted using the first prediction model. Based on the cloud transmittance time series within the second historical window, the cloud transmittance time series for the second future period is predicted using the second prediction model. Based on the cloud transmittance time series of the first future time period and the cloud transmittance time series of the second future time period, the cloud transmittance time series of the future target time period is obtained.

6. The ultra-short-term photovoltaic power forecasting method according to claim 5, characterized in that, Both the first prediction model and the second prediction model are obtained by training the Cross ViViT model.

7. The ultra-short-term photovoltaic power forecasting method according to claim 1, characterized in that, The step of calculating the photovoltaic power forecast sequence for the future target period based on the surface irradiance forecast sequence for the future target period and the meteorological station observation data includes: Based on the surface irradiance forecast sequence for the future target period and the meteorological station observation data, calculate the photovoltaic panel temperature at each time step within the future target period; Based on the surface irradiance forecast sequence for the future target period, the photovoltaic panel temperature at each time step within the future target period, and the preset photovoltaic panel parameters, the photovoltaic power forecast sequence for the future target period is calculated.

8. An ultra-short-term photovoltaic power forecasting system, characterized in that, include: The data acquisition module is used to acquire all-sky imager observation data, ground meteorological station observation data, geostationary satellite observation data, reanalysis data, and surface albedo data for the target site; The first calculation module is used to calculate the cloud transmittance time series within the first historical window before the current moment based on the observation data of the all-sky imager. The second calculation module is used to calculate the cloud transmittance time series within the second historical window before the current moment based on the geostationary satellite observation data, the reanalysis data, and the surface albedo data. The model prediction module is used to predict the cloud transmittance time series for the future target period based on the cloud transmittance time series in the first historical window and the cloud transmittance time series in the second historical window using a pre-trained time series prediction model, and to calculate the surface irradiance forecast sequence for the future target period based on the cloud transmittance time series for the future target period. The power forecast module is used to calculate the photovoltaic power forecast sequence for the future target period based on the surface irradiance forecast sequence for the future target period and the meteorological station observation data.

9. An electronic device, characterized in that, The method includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement an ultra-short-term photovoltaic power forecasting method as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium storing computer instructions, characterized in that, The computer instructions are used to cause the computer to execute the ultra-short-term photovoltaic power forecasting method according to any one of claims 1 to 7.

Images