Method, device, equipment and medium for integrating satellite and ground data photovoltaic output prediction

Abstract

The present application relates to the technical field of power system load forecasting, and particularly relates to a method, device, equipment and medium for predicting photovoltaic output by integrating satellite and ground data, the method comprising: preprocessing data, the data including remote sensing inversion radiation data, ground observation radiation data and satellite remote sensing data; performing spatial grid division based on the spatially continuous remote sensing inversion radiation data, the grid division being based on geographical position and meteorological characteristics, and regions with different meteorological characteristics being divided into different grid units, each grid unit containing at least one ground observation station; establishing a solar radiation prediction model for each grid unit; calculating the solar radiation prediction value of each grid according to the trained solar radiation prediction model, in combination with real-time satellite cloud images and ground observation radiation data; and overcoming the limitation that traditional ground-based methods cannot accurately predict the fluctuation of radiation caused by changes in cloud cover, thereby significantly improving the prediction accuracy of solar radiation and photovoltaic output.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power system load forecasting, and in particular to a photovoltaic power output prediction method, device, equipment and medium integrating satellite and ground data. BACKGROUND

[0002] Due to the uncertainty of solar radiation, solar photovoltaic power generation often exhibits random, volatile and intermittent characteristics. In order to ensure the stability of the power grid system with high photovoltaic power generation penetration, reliable and accurate photovoltaic power output prediction is crucial, which can provide key support for decision-making of power system optimization scheduling, safe operation and reasonable pricing.

[0003] Photovoltaic power output prediction generally includes two core links of meteorological parameter prediction and photovoltaic power generation simulation, and the prediction of ground surface solar radiation is the most critical. Solar radiation is the most important factor affecting solar power generation and is the main source of power generation prediction uncertainty. The commonly used methods for solar radiation prediction include statistical methods, physical methods and hybrid methods. The main input of the above methods is the historical solar radiation and its attribute information observed on the ground, which can be collectively referred to as ground-based methods. This kind of method usually cannot output the random fluctuations of solar radiation caused by changes in cloud cover, thereby leading to a decrease in prediction performance. Due to the rapid movement of clouds, clouds far from the point of interest can also affect the local weather within tens of minutes to hours, which makes local observations insufficient to predict future changes in solar radiation, and it is necessary to expand the spatiotemporal observation range.

[0004] The use of cloud images observed by high-frequency meteorological satellites can infer the movement state of the cloud layer, thereby improving the accuracy of solar radiation and photovoltaic power output prediction. The defect of this method is that the calculation of satellite images is spatiotemporally intensive, and in the process of regional application, it faces high computational cost, resulting in low service efficiency.

[0005] The information disclosed in this BACKGROUND section is only intended to enhance the understanding of the general background of the application, and should not be considered as recognition or suggestion that this information forms the prior art known by those skilled in the art. SUMMARY

[0006] The present application provides a photovoltaic power output prediction method, device, equipment and medium integrating satellite and ground data, thereby effectively solving the problems in the background art.

[0007] In order to achieve the above purpose, the technical scheme adopted by the present application is as follows: a photovoltaic power output prediction method integrating satellite and ground data, comprising the following steps:

[0008] S10: preprocessing data, the data including remote sensing inversion radiation data, ground observation radiation data and satellite remote sensing data;

[0009] S20: performing spatial grid division based on the spatially continuous remote sensing inversion radiation data, the grid division being based on geographical position and meteorological characteristics, and dividing regions with different meteorological characteristics into different grid units, each grid unit containing at least one ground observation station;

[0010] S30: establishing a solar radiation prediction model for each grid unit, and training the model in combination with the ground observation radiation data and the satellite remote sensing data;

[0011] S40: calculating a solar radiation prediction value for each grid according to the trained solar radiation prediction model in combination with real-time satellite cloud images and the ground observation radiation data;

[0012] S50: calculating power output of a photovoltaic system in a target time period based on the solar radiation prediction value and by using a photovoltaic system simulation tool.

[0013] Further, in step S10, the data is preprocessed, including:

[0014] S11: performing quality control inspection on the remote sensing inversion radiation data, comparing with ground observation true data, eliminating data records with an absolute difference greater than 20%, and filling the eliminated data with inversion multi-year average values;

[0015] S12: performing quality detection on the ground observation radiation data, eliminating noise and missing values by using filtering and interpolation and the like preprocessing techniques, detecting and eliminating abnormal values by using statistical methods and physical models, and finally calibrating and verifying the data to ensure its accuracy and consistency;

[0016] S13: performing geometric and radiation correction on the satellite remote sensing data to eliminate errors caused by instruments and observation angles, applying quality evaluation indexes to screen out contaminated or distorted data, and filling in the contaminated areas based on adjacent observation images;

[0017] S14: matching the processed remote sensing inversion radiation data, ground observation radiation data and satellite remote sensing data through time and space information to form a paired sequence of satellite data and ground data.

[0018] Further, in step S20, the spatial grid division is performed based on the spatially continuous remote sensing inversion radiation data, and the step includes:

[0019] S21: constructing an infinite Gaussian mixture model, initializing mean and covariance matrix of the infinite Gaussian mixture model, performing parameter estimation by Gibbs sampling, calculating posterior probability of each data point belonging to each Gaussian distribution, and updating parameter values to convergence;

[0020] S22: determining the optimal number of clusters in the data using an adaptive method through the output results of the infinite Gaussian mixture model;

[0021] S23: inputting the preprocessed data into the infinite Gaussian mixture model fitted with the optimal number of clusters to predict the partition to which each pixel belongs, forming a spatially continuous partition prediction result;

[0022] S24: performing edge detection on the partition prediction result, extracting boundary lines and performing smoothing processing to determine the accurate boundary and range of each partition, and converting it into spatial coordinates for storage.

[0023] Further, in step S22, the adaptive method is used to determine the optimal number of clusters in the data, and the adaptive method is realized by Bayesian information criterion, and the model of the Bayesian information criterion includes:

[0024]

[0025] In the formula, BIC is the Bayesian information criterion, n is the sample size, k is the number of model parameters, is the maximum likelihood estimation of the model.

[0026] Further, in step S30, a solar radiation prediction model is established for each grid cell, and the step includes:

[0027] S31: matching the ground observation radiation data and the satellite remote sensing data according to the geographic coordinates and time stamp of the partition in units of each spatial grid cell partition to form paired data;

[0028] S32: constructing a solar radiation prediction model integrating the ground observation radiation data and the satellite remote sensing data based on the paired data;

[0029] S33: extracting training samples including input features and target variables from historical observation data, dividing the data set into training set, validation set and test set in the ratio of 3:1:1, and using the training set to train the solar radiation prediction model to optimize the model parameters;

[0030] S34: using the validation set and the test set to evaluate the solar radiation prediction model, calculating the performance indicators of the solar radiation prediction model to ensure that the performance of the solar radiation prediction model on the validation set and the test set is optimal.

[0031] Further, in step S32, the solar radiation prediction model integrating the ground observation radiation data and the satellite remote sensing data is constructed, including constructing the solar radiation prediction model as:

[0032]

[0033] wherein, represents the solar radiation to be predicted; t represents the time at which the latest historical observation is located; m represents the maximum prediction length; f represents the deep learning prediction model; D represents the model input; x i is the solar radiation observation data at the i-th time point; is the visible light satellite image block of size h x w at the i-th time point; T is the input historical sequence length.

[0034] Further, in step S40, the solar radiation prediction value of each grid is calculated, and the step includes:

[0035] S41: determining the geographic coordinates and time period of the point to be predicted, wherein the geographic coordinates include longitude and latitude, and the time period includes year, month, day, hour, and minute information;

[0036] S42: mapping the point to be predicted into the grid unit that has been divided according to the geographic coordinates, and matching to the solar radiation prediction model corresponding to the grid unit;

[0037] S43: extracting the historical observation data of the ground observation station closest to the geographic position of the point to be predicted in the grid unit corresponding to the point to be predicted;

[0038] S44: extracting a visible light cloud image sequence of T hours before the prediction time period from the satellite image, pairing with the historical observation data, and forming a data pair meeting the input format requirements of the prediction model, with the point to be predicted as the center;

[0039] S45: inputting the paired data pair into the solar radiation prediction model to generate the solar radiation prediction result of the point to be predicted.

[0040] Further, in step S50, a photovoltaic system simulation tool is used to calculate the power output of the photovoltaic system in the target time period, and the step includes:

[0041] S51: selecting a photovoltaic power generation simulator; configuring the running environment of the photovoltaic power generation simulator, including dependent libraries and initial parameters; and setting the basic configuration of the photovoltaic power generation simulator by calling an initialization function;

[0042] S52: configuring various parameters of the photovoltaic system, including component parameters, installation parameters, and geographic location;

[0043] S53: inputting meteorological data of the photovoltaic power station location, wherein the meteorological data at least includes the predicted solar radiation and the near-real-time observed temperature;

[0044] S54: using the configured photovoltaic system parameters and the input meteorological data, calculating the direct current output of the photovoltaic module in the target time period through the photovoltaic power generation simulator;

[0045] S55: configuring the parameters of the photovoltaic power station inverter model, the parameters of the photovoltaic power station inverter model at least including maximum direct current input power, maximum alternating current output power, and efficiency;

[0046] S56: inputting the calculated direct current power generation of the photovoltaic module into the inverter model, performing direct current to alternating current conversion calculation through the inverter model, and obtaining the alternating current output of the photovoltaic system in the target time period;

[0047] S57: outputting the calculated alternating current output and performing necessary verification and evaluation.

[0048] The application also includes a device for predicting photovoltaic output integrating satellite and ground data, using the method as described above, comprising:

[0049] A data preprocessing unit for preprocessing data, the data including remote sensing inversion radiation data, ground observation radiation data, and satellite remote sensing data;

[0050] A spatial grid division unit for dividing spatial grids based on the spatially continuous remote sensing inversion radiation data, the grid division being based on geographical location and meteorological characteristics, and regions with different meteorological characteristics being divided into different grid units, each grid unit containing at least one ground observation station;

[0051] A partition prediction modeling unit for establishing a solar radiation prediction model for each grid unit, and training in combination with the ground observation radiation data and satellite remote sensing data;

[0052] A solar radiation prediction unit for calculating the solar radiation prediction value of each grid according to the trained solar radiation prediction model in combination with real-time satellite cloud images and the ground observation radiation data;

[0053] A photovoltaic output calculation unit for calculating the power output of the photovoltaic system in the target time period based on the solar radiation prediction value and using a photovoltaic system simulation tool.

[0054] The application also includes a computer device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, the processor implementing the computer program to realize the method as described above.

[0055] The application also includes a storage medium having a computer program stored thereon, the computer program being executable by a processor to realize the method as described above.

[0056] The beneficial effects of the present application are:

[0057] By analyzing the spatio-temporal heterogeneity characteristics of meteorological factors affecting photovoltaic output, the region is divided into grids, and only one solar radiation prediction model needs to be established for each grid, then combined with the specific location and parameters of the photovoltaic station to predict the photovoltaic output, which greatly reduces the calculation consumption of the solar radiation prediction link; by simultaneously taking the satellite and ground observation sequence as the input, and using the convolutional neural network to extract the spatio-temporal characteristics, and using the LSTM to conduct time series reasoning, the spatio-temporal dynamic changes of the cloud layer are effectively inferred and quantified, and the influence of the cloud layer on the photovoltaic output is improved, thereby improving the photovoltaic output prediction accuracy under cloudy weather conditions. BRIEF DESCRIPTION OF DRAWINGS

[0058] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or prior art description will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments described in the present application, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.

[0059] Figure 1 The flowchart of the photovoltaic output prediction method integrating satellite and ground data;

[0060] Figure 2 The technical roadmap of the photovoltaic output prediction method integrating satellite and ground data;

[0061] Figure 3 The architecture diagram of the integrated satellite and ground prediction model;

[0062] Figure 4 The grid division result of Jiangsu Zhenjiang obtained by the Infinite Gaussian Mixture (IGMM) model;

[0063] Figure 5 The prediction example diagram at 12:00 on June 20, 2023;

[0064] Figure 6 The accuracy comparison diagram of the prediction results under different preposition lengths;

[0065] Figure 7 The structural diagram of the photovoltaic output prediction device integrating satellite and ground data;

[0066] Figure 8 The structural diagram of the computer device. DETAILED DESCRIPTION

[0067] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0068] Example 1:

[0069] like Figures 1 to 3 As shown: A method for predicting photovoltaic output by integrating satellite and ground data includes the following steps:

[0070] S10: Preprocess the data, which includes remote sensing inversion radiometric data, ground observation radiometric data, and satellite remote sensing data;

[0071] S20: Spatial gridding is performed based on spatially continuous remote sensing inversion radiation data. The purpose of gridding is to group locations with similar spatiotemporal variation characteristics of solar radiation and to separate areas with large differences in variation characteristics. This guides the ground data collection process for calibrating photovoltaic output prediction results. Gridding is based on geographical location and meteorological characteristics, dividing regions with different meteorological characteristics into different grid units. Each grid unit contains at least one ground observation station. Since meteorological conditions vary in different regions, directly using national or large-area prediction models may not be accurate enough. Dividing the region into multiple different grid units ensures that the meteorological characteristics within each unit are as consistent as possible, which can improve the prediction accuracy of each unit. Ensuring that each grid unit contains at least one ground observation station ensures that the model can obtain sufficient calibration data.

[0072] S30: Establish a solar radiation prediction model for each grid cell and train it using ground-based observation radiation data and satellite remote sensing data. The advantages of this step are: firstly, since the spatiotemporal variation of solar radiation in each partition has similar characteristics, it can ensure that the solar radiation prediction results of the model can effectively support the output prediction of all photovoltaic stations in the partition; secondly, it is only necessary to establish a prediction model for each partition rather than each station, which avoids the high computational cost of dense prediction and improves the computational efficiency of regional applications.

[0073] S40: Based on the trained solar radiation prediction model, combined with real-time satellite cloud imagery and ground-based radiation observation data, calculate the predicted solar radiation value for each grid cell; using real-time data ensures the timeliness of the prediction results and adapts to dynamic changes in meteorological conditions. This step provides specific radiation input data for subsequent photovoltaic power output prediction.

[0074] S50: Based on the predicted solar radiation, use photovoltaic system simulation tools to calculate the power output of the photovoltaic system within the target time period.

[0075] By analyzing the spatio-temporal heterogeneity characteristics of meteorological factors affecting photovoltaic output, the region is divided into grids, at least one solar radiation prediction model is established for each grid, and then combined with the specific location and parameters of the photovoltaic station, the photovoltaic output is predicted, which greatly reduces the calculation consumption of the solar radiation prediction link;

[0076] By combining remote sensing inversion radiation data, ground observation radiation data and satellite remote sensing data, the spatio-temporal variation characteristics of solar radiation can be more comprehensively captured, especially the influence of cloud cover and its rapid movement on solar radiation. This method overcomes the limitations of traditional ground-based methods that cannot accurately predict the fluctuation of radiation caused by changes in cloud cover, thereby significantly improving the prediction accuracy of solar radiation and photovoltaic output.

[0077] By combining satellite remote sensing data and ground observation radiation data, integrating radiation information from different sources, the limitations of a single data source are effectively supplemented, especially under complex weather conditions such as cloud cover, which helps to improve the accuracy of photovoltaic output prediction.

[0078] Spatial grid division is based on geographical location and meteorological characteristics, which gathers regions with similar meteorological characteristics in the same grid and ensures that at least one ground observation station is used for data calibration in each grid, which can ensure that the prediction model can accurately reflect the meteorological changes in each grid, enhancing the reliability of photovoltaic output prediction in the region.

[0079] By establishing a solar radiation prediction model for each grid unit instead of modeling each observation station separately, the calculation amount is greatly reduced, especially in large-scale regional applications, avoiding the high cost of repeated calculations while ensuring the improvement of overall calculation efficiency.

[0080] As a preferred embodiment of the above, in step S10, the data is preprocessed, including:

[0081] S11: Quality control inspection of remote sensing inversion radiation data, comparison with ground observation true data, elimination of data records with absolute difference greater than 20%, and filling of eliminated data with inversion multi-year average value;

[0082] S12: Quality detection of ground observation radiation data, elimination of noise and missing values by filtering and interpolation, detection and elimination of outliers by statistical methods and physical models; finally, calibration and verification of the data to ensure its accuracy and consistency;

[0083] S13: Geometric and radiometric correction of satellite remote sensing data to eliminate errors caused by instruments and observation angles, application of quality evaluation indicators to screen contaminated or distorted data, and filling of contaminated areas based on adjacent observation images;

[0084] S14: Match the processed remote sensing inversion radiation data, ground observation radiation data and satellite remote sensing data through time and space information to form a paired sequence of satellite data and ground data.

[0085] Through step-by-step quality control inspection and preprocessing, the accuracy and consistency of the data are ensured, thereby improving the reliability of the subsequent prediction model; by eliminating outliers, filling missing data and correcting image data, data errors are reduced, and the influence of inaccurate data on the prediction result is avoided; through the pairing of data in time and space, the consistency between different data sources is ensured, which is crucial for model training and prediction accuracy; the high-quality data after processing enables the model to better fit the actual situation during training, improving the prediction accuracy and reliability; through effective data preprocessing and quality control, the interference of incorrect data in subsequent model training is reduced, thereby improving the calculation efficiency.

[0086] In this embodiment, in step S20, spatial grid division is performed based on spatially continuous remote sensing inversion radiation data, and the step includes:

[0087] S21: Construct an infinite Gaussian mixture (IGMM) model, which is a model based on Bayesian non-parametric method, and its main advantage is that it does not need to pre-set the number of clusters, initialize the mean and covariance matrix of the infinite Gaussian mixture (IGMM) model, estimate the parameters through Gibbs sampling, calculate the posterior probability of each data point belonging to each Gaussian distribution, and update the parameter value to convergence; this step is the training process of the infinite Gaussian mixture (IGMM) model, which ensures that the model can correctly capture the distribution of the data; update the parameter value to convergence to ensure that the parameters of the model get stable values through repeated iterations, thereby improving the accuracy of the model;

[0088] S22: Use an adaptive method to determine the optimal number of clusters in the data through the output results of the infinite Gaussian mixture (IGMM) model;

[0089] S23: Input the preprocessed data into the infinite Gaussian mixture (IGMM) model fitted with the optimal number of clusters to predict the partition to which each pixel belongs, forming a spatially continuous partition prediction result; ensures that the data used is quality-controlled, thereby improving the accuracy of the prediction;

[0090] S24: Perform edge detection on the partition prediction result, extract the boundary line and perform smoothing processing to determine the accurate boundary and range of each partition, and convert it into spatial coordinates for storage.

[0091] Through the unsupervised learning characteristics of the infinite Gaussian mixture (IGMM) model, the optimal partition of the data can be automatically determined without manually setting the number of clusters, thereby improving the processing efficiency; through edge detection and smoothing processing, the boundary of each partition can be more accurately determined, thereby improving the spatial resolution of the model; accurate spatial grid division and refined partitions can make subsequent solar radiation prediction more accurate, thereby improving the reliability and practicality of photovoltaic power output prediction.

[0092] As a preferred embodiment of the above, in step S22, an adaptive method is used to determine the optimal number of clusters in the data, and the adaptive method is implemented by the Bayesian information criterion (BIC). The BIC model includes:

[0093]

[0094] In the formula, BIC is the Bayesian information criterion, n is the number of samples, k is the number of model parameters, is the maximum likelihood estimation of the model. BIC is calculated for different cluster numbers k, and the k with the minimum BIC is selected as the optimal cluster number.

[0095] In the model selection process, BIC selects the best model by balancing the accuracy of data fitting and the complexity of the model; specifically, BIC adds a penalty term to the maximum likelihood estimation, and the penalty term is related to the number of model parameters. Therefore, when selecting a model, BIC can avoid selecting a model that is too complex, i.e., overfitting.

[0096] BIC provides a scientific method for automatically selecting the optimal number of clusters. By calculating the BIC values of different cluster numbers, the model with the best balance can be objectively selected, avoiding the bias that may be caused by manually setting the number of clusters. By selecting the optimal number of clusters, the spatial area can be more accurately divided, so that the subsequent solar radiation prediction and photovoltaic power output calculation can reflect the true spatial distribution characteristics. The optimal number of clusters improves the prediction accuracy of the model, further enhancing the reliability and practicality of photovoltaic power output prediction.

[0097] In step S30, a solar radiation prediction model is established for each grid cell, and the step includes:

[0098] S31: In the unit of each spatial grid cell partition, the ground observation radiation data and satellite remote sensing data are matched according to the geographical coordinates and time stamp of the partition to form paired data; this step is the basis for the entire model training. First, the ground observation radiation data and satellite observation data are matched in the unit of each spatial grid cell. This step ensures that the input data for model training contains both actual ground observation values and visible light images in satellite data, providing a complete data perspective;

[0099] S32: Based on the paired data, a solar radiation prediction model integrating ground observation radiation data and satellite remote sensing data is constructed; by integrating multiple sources of data, the characteristics and spatio-temporal information of different data are fully utilized, further providing a model architecture for subsequent training;

[0100] S33: Training samples including input features and target variables are extracted from historical observation data, the training samples are divided into training set, validation set and test set according to the ratio of 3:1:1, and the training set is used to train the solar radiation prediction model to optimize the model parameters; during the training process, the root mean square error is selected as the loss function, and the Adam optimizer is selected for gradient calculation and parameter optimization; this step can enable the model to learn the characteristics of the data and optimize the parameters, thereby improving the prediction accuracy, and is the core step to improve the performance of the model.

[0101] S34: The solar radiation prediction model is evaluated using the validation set and the test set, and the performance indicators (such as mean square error, R 2 , accuracy, etc.) of the solar radiation prediction model are calculated to ensure that the performance of the solar radiation prediction model on the validation set and the test set is optimal. The optimal model is saved as a.h5 file for subsequent application and deployment. Through the double verification of the validation set and the test set, problems such as model overfitting can be avoided, and the optimal model is finally used for actual application.

[0102] By matching ground observation radiation data with satellite data, the data from different sources is fully utilized, and the fusion of multiple sources of data improves the diversity and comprehensiveness of the model input, which can better capture the spatio-temporal variation characteristics of solar radiation, thereby improving the prediction accuracy; the deep learning model can effectively capture complex nonlinear relationships, especially the extraction of spatio-temporal features, and the model can learn from historical data of satellite images and ground observations and predict future trends of solar radiation changes; by refining the prediction process to each spatial grid cell, the model can adapt to the specific radiation characteristics of different geographical regions, and this targeted processing improves the accuracy of large-area prediction, which helps to improve the prediction effect in different regions and climate conditions; through the evaluation of the validation set and the test set, the performance of the model on different data sets is stable, and overfitting is avoided. This training, verification, and testing structure can ensure that the model not only performs well on historical data, but also maintains high prediction accuracy in future actual applications.

[0103] As a preferred embodiment of the above, in step S32, the solar radiation prediction model integrating ground observation radiation data and satellite remote sensing data is constructed, including constructing the solar radiation prediction model as:

[0104]

[0105] In the formula, represents the solar radiation to be predicted; t represents the time at which the latest historical observation is located; m represents the maximum prediction length; f represents the deep learning prediction model; D represents the model input; x i is the solar radiation observation data at the i-th time point; is the visible light satellite image block of the i-th time point with a size of h x w; T is the input historical sequence length.

[0106] By constructing a solar radiation prediction model integrating ground observation radiation data and satellite remote sensing data, the advantages of the two types of data are fully utilized, and the actual observation on the ground and the wide spatial coverage of satellite remote sensing are combined, so that the future change of solar radiation can be more accurately predicted; by introducing the visible light image block of satellite remote sensing, the model can capture information such as cloud layer change and weather fluctuation, thereby improving the photovoltaic power output prediction ability under complex weather conditions.

[0107] The structural architecture of the deep learning model is as shown in Figure 3 The 3D convolution is used to extract the spatio-temporal features of the satellite image sequence, the 1D convolution is used to extract the fluctuation features of the ground observation sequence, then all the features are aligned in the time dimension and input to the LSTM structure for time series reasoning, and finally the MLP is used to establish the nonlinear relationship between the reasoning information and the subsequent solar radiation.

[0108] In this embodiment, in step S40, the solar radiation prediction value of each grid is calculated, and the step includes:

[0109] S41: Determine the geographic coordinates and time period of the point to be predicted, the geographic coordinates including longitude and latitude, and the time period including year, month, day, hour, and minute information;

[0110] S42: Map the point to be predicted to the grid unit that has been divided according to the geographic coordinates, and match to the solar radiation prediction model corresponding to the grid unit; this step ensures the pertinence and accuracy of the prediction model;

[0111] S43: Extract the historical observation data of the ground observation station nearest to the geographic position of the point to be predicted in the grid unit corresponding to the point to be predicted; by using the principle of spatial proximity, the data of the nearest neighbor ground observation station is selected as the reference, which is helpful to capture the meteorological characteristics of the local area;

[0112] S44: Take the point to be predicted as the center, extract the visible light cloud image sequence of T hours before the prediction time period from the satellite image, and pair with the historical observation data to form a data pair meeting the input format requirements of the prediction model;

[0113] S45: Input the paired data pair into the solar radiation prediction model to generate the solar radiation prediction result of the point to be predicted.

[0114] By precisely matching the prediction model of the grid where the to-be-predicted point is located, and combining the historical data of the nearest ground observation station and the real-time satellite cloud image, the spatio-temporal variation characteristics of solar radiation can be more accurately captured, thereby improving the prediction accuracy.

[0115] In step S50, the power output of the photovoltaic system in the target time period is calculated using a photovoltaic system simulation tool, and the step includes:

[0116] S51: Select a suitable photovoltaic power generation simulator, such as PVlib or other professional software; configure the running environment of the photovoltaic power generation simulator, including dependent libraries and initial parameters; set the basic configuration of the photovoltaic power generation simulator by calling the initialization function;

[0117] S52: Configure the parameters of the photovoltaic system, including component parameters, installation parameters, and geographic location;

[0118] The parameters can be divided into three groups: ① Component parameters: set the electrical parameters of the photovoltaic module, such as open-circuit voltage (Voc), short-circuit current (Isc), maximum power point voltage (Vmp), maximum power point current (Imp), etc.; ② Installation parameters: set the installation angle and azimuth angle of the photovoltaic module; ③ Geographic location: set the geographic location (longitude, latitude, and height) of the photovoltaic power station and the prediction time period (year-month-day-hour-minute);

[0119] S53: Input the meteorological data of the photovoltaic power station location, which at least includes the predicted solar radiation and the near-real-time observed temperature; use the configured photovoltaic system parameters to simulate the direct current output of the photovoltaic module;

[0120] S54: Use the configured photovoltaic system parameters and input meteorological data to calculate the direct current output of the photovoltaic module in the target time period through the photovoltaic power generation simulator;

[0121] S55: Configure the parameters of the photovoltaic power station inverter model, which at least include the maximum direct current input power, the maximum alternating current output power, and the efficiency;

[0122] S56: Input the calculated direct current power generation of the photovoltaic module into the inverter model, and perform direct current to alternating current conversion calculation through the inverter model to obtain the alternating current output of the photovoltaic system in the target time period;

[0123] S57: Output the calculated alternating current output, and perform necessary verification and evaluation to confirm the accuracy and reliability of the simulation results, and provide scientific basis for decision-making of power system optimization scheduling, safe operation, and reasonable pricing.

[0124] By selecting a suitable photovoltaic power generation simulator and configuring its operating environment, dependent libraries, and initial parameters in detail, the simulation environment can be ensured to be as close as possible to the actual photovoltaic system, thereby improving the prediction accuracy of power output and helping to more accurately evaluate the power generation capacity of the photovoltaic system, providing a reliable basis for the scheduling and planning of the power system; the scheme requires configuring various parameters of the photovoltaic system, including component parameters, installation parameters, and geographic location, and comprehensive parameter configuration can more realistically reflect the actual operation of the photovoltaic system, including the influence of different installation angles, azimuth angles, and geographic locations on power generation efficiency, thereby improving the comprehensiveness and accuracy of the prediction; inputting meteorological data of the photovoltaic power station location, including predicted solar radiation and near-real-time observed temperature, can fully consider the influence of meteorological conditions on the power generation efficiency of the photovoltaic system, and the use of multiple sources of data can more accurately simulate the power generation performance of photovoltaic components under different meteorological conditions, thereby improving the accuracy of the prediction results.

[0125] Embodiment 2:

[0126] To verify the performance of the application, the photovoltaic output prediction of a station in Zhenjiang, Jiangsu is taken as an example to verify the application.

[0127] 1) Data collection and processing: 5km resolution monthly solar radiation data is collected from the Pangaea website (https: / / rda.ucar.edu / datasets / ds609.0 / ), solar radiation and photovoltaic output observation data are collected from ground stations, satellite cloud image data is downloaded from the Fengyun meteorological remote sensing data service network, and the data is preprocessed and time-space paired. https: / / doi.org / 10.1594 / PANGAEA.904136 ) collects 5km resolution monthly solar radiation data, collects solar radiation and photovoltaic output observation data from ground stations, and downloads satellite cloud image data from the Fengyun meteorological remote sensing data service network, and pre-processes and time-space pairs the data.

[0128] 2) Spatial grid division: an infinite Gaussian mixture (IGMM) model is constructed, and the BIC criterion is used to adaptively infer that the optimal number of partitions in Zhenjiang area is 14, then the spatially continuous monthly solar radiation data is input into the model to predict the partition of each location, and the division result of the spatial grid is as shown in Figure 4 , pixels of the same color in the figure belong to the same partition.

[0129] 3) Partition prediction modeling: the ground observation radiation data collected in this application example belongs to Figure 4 the light green partition in space, so a prediction model is constructed for this area. The observation data covers the time range from January 1, 2022 to December 31, 2023. We construct the prediction model based on the data of 2022, and the data of 2023 is used for independent test of photovoltaic output prediction performance. Organize the samples according to the form of the solar radiation prediction model, divide them into training set, validation set and test set in the ratio of 3:1:1, and use the training data set to train the model to obtain the optimized model parameters, and the model with the smallest mean square error on the validation set is reserved for subsequent application and deployment.

[0130] 4) Solar radiation prediction: Based on the observation station, the visible light cloud image sequence of the 6 hours before the predicted time period is extracted from the satellite image, matched with the historical observation data to form a data pair that meets the input format requirements of the prediction model. The paired ground and satellite data are input into the prediction model to predict the solar radiation. As shown in Figure 5 , the prediction results at 12:00 on June 20, 2023 are shown. The model predicts the radiation change in the next 6 hours based on the real observation in the past 6 hours. By continuously rolling the time to be predicted, the prediction results for the whole year of 2023 can be obtained.

[0131] 5) Photovoltaic power output calculation: The PVLIB professional software is used to simulate the photovoltaic power generation process, and the power generation of a specific installed capacity photovoltaic station under the predicted solar radiation level can be obtained. The photovoltaic system of the ground photovoltaic station selected in the example of the present application is inclined at an angle of 25° south, uses polycrystalline silicon photovoltaic modules, has an installed capacity of 12.8 MW, the module temperature difference ΔT is about 3℃, and the inverter model is "PVP2500 240V[CEC 2006]". Figure 6 The comparison results of photovoltaic power output prediction results and actual recorded values are shown. When the preset time length increases from 1 hour to 6 hours, the prediction error gradually increases, and the accuracy gradually decreases. The average prediction error for 1 hour in advance is 0.77 MW, and the accuracy is 93.96%.

[0132] The present application proposes a method for predicting photovoltaic power based on spatial similarity grid division and integrating satellite and ground data, aiming at the defects of insufficient prediction accuracy of common photovoltaic power prediction technology under cloudy conditions and intensive calculation of satellite-based prediction method. By analyzing the spatio-temporal heterogeneity characteristics of meteorological factors affecting photovoltaic power output, the region is divided into grids, and only one solar radiation prediction model needs to be established for each grid. Then, combined with the specific location and parameters of the photovoltaic station, the photovoltaic power output is predicted, which greatly reduces the calculation consumption of the solar radiation prediction link. By simultaneously inputting satellite and ground observation sequences and using convolutional neural networks to extract spatio-temporal features and LSTM for time series reasoning, the spatio-temporal dynamic changes of clouds are effectively inferred and quantified, and the influence of clouds on photovoltaic power output is improved. The accuracy of photovoltaic power output prediction under cloudy weather conditions is improved. Therefore, the present application has the advantages of regional application efficiency, high model prediction accuracy, etc.

[0133] The present application also includes a prediction device for integrating satellite and ground data photovoltaic power output, which uses the method as described above, as shown in Figure 7 , comprising:

[0134] A data preprocessing unit for preprocessing data, including remote sensing inversion radiation data, ground observation radiation data and satellite remote sensing data;

[0135] a spatial grid division unit, configured to divide a spatial grid based on spatially continuous remote sensing inversion radiation data, the grid division being based on geographical positions and meteorological characteristics, and regions with different meteorological characteristics being divided into different grid units, each grid unit containing at least one ground observation station;

[0136] a partitioned prediction modeling unit, configured to establish a solar radiation prediction model for each grid unit, and train the model in combination with ground observation radiation data and satellite remote sensing data;

[0137] a solar radiation prediction unit, configured to calculate a solar radiation prediction value for each grid according to the trained solar radiation prediction model in combination with real-time satellite cloud images and ground observation radiation data;

[0138] a photovoltaic output calculation unit, configured to calculate the power output of a photovoltaic system in a target time period based on the solar radiation prediction value and by using a photovoltaic system simulation tool.

[0139] Please refer to Figure 8 The computer device provided by the embodiment of the present application is shown in the structural schematic diagram. The computer device 400 provided by the embodiment of the present application comprises a processor 410 and a memory 420, the memory 420 stores a computer program executable by the processor 410, and the computer program is executed by the processor 410 to perform the method as above.

[0140] The embodiment of the present application further provides a storage medium 430, the storage medium 430 stores a computer program, and the computer program is executed by the processor 410 to perform the method as above.

[0141] The storage medium 430 can be realized by any type of volatile or non-volatile storage device or a combination thereof, such as a static random access memory (SRAM), an electrically erasable programmable read-only memory (EEPROM), an erasable programmable read-only memory (EPROM), a programmable read-only memory (PROM), a read-only memory (ROM), a magnetic storage, a flash memory, a magnetic disk or an optical disk.

[0142] In the description of the application, the terms "first", "second", "third", etc. are used only for the purpose of description, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the technical features indicated. Therefore, the features defined as "first", "second", etc. can be explicitly or implicitly included one or more of the features. The meaning of "a plurality of" is two or more, unless otherwise explicitly specified and limited.

[0143] In the present application, unless otherwise explicitly specified and limited, the terms "mounting", "connecting", "connecting", "fixing" and the like should be understood in a broad sense, for example, it can be fixedly connected, or it can be detachably connected, or it can be integrated; it can be mechanically connected, or it can be electrically connected; it can be directly connected, or it can be indirectly connected through an intermediate medium; it can be the internal communication of two elements or the interaction relationship between two elements. For those skilled in the art, the specific meaning of the above-mentioned terms in the present application can be understood according to the specific circumstances.

[0144] In the description of the present application, the description of the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In the present application, the illustrative description of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art can combine and combine the different embodiments or examples described in the present application and the features of the different embodiments or examples without contradiction.

[0145] Any process or method descriptions in flow charts or otherwise described herein represents an example of executable instructions for performing a certain task or set of tasks and can be understood as representing a module, segment, or portion of code that comprises one or more executable instructions for performing the specified logic function or task. The scope of preferred embodiments of the present application includes additional implementation in which the functions described in the illustrated or discussed order are performed in a different order, including substantially simultaneously or in reverse order, as will be understood by those skilled in the art of the embodiments of the present application.

[0146] The logic and / or steps represented in flow diagrams or otherwise described herein, for example, can be considered as a sequence of instructions to implement logic functions, and can be realized in any computer-readable medium for use by an instruction execution system, apparatus, or device, such as a computer-based system, processor- containing system, or other system that can fetch the instructions from the instruction execution system, apparatus, or device and execute the instructions. In the context of this specification, a "computer-readable medium" can be any means that can contain, store, communicate, propagate or transport the program for use by or in connection with the instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of the computer-readable medium include an electronic connection (an electronic device), a portable computer diskette (a magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, and a portable compact disc read-only memory (CDROM). Further, the computer-readable medium can even be paper or other suitable medium upon which the program is printed, as the program can be electronically captured, for example, by optically scanning the paper or other suitable medium, then electronically converted into a form that is suitable for use by the instruction execution system, apparatus, or device, and then stored in computer memory.

[0147] It should be understood that aspects of the application can be implemented in hardware, software, firmware or combinations thereof. In the above embodiments, the various steps or methods can be implemented in software or firmware that is stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, any of the following technologies, known in the art, or their combinations can be used: discrete logic circuitry having logic gates for implementing logic functions on data signals, application specific integrated circuits having appropriate combinational logic gates, programmable gate arrays (PGA), field programmable gate arrays (FPGA), and the like.

[0148] Those skilled in the art can understand that all or part of the steps carried out by the above-mentioned embodiment method can be completed by a program instructing related hardware, and the program can be stored in a computer-readable storage medium. When the program is executed, it includes one of the steps of the method embodiment or a combination thereof.

[0149] The storage medium mentioned above can be a read-only memory, a magnetic disk or an optical disk, etc. Although the embodiments of the present application have been shown and described above, it should be understood that the above-mentioned embodiments are exemplary and cannot be understood as limiting the present application, and those skilled in the art can make changes, modifications, replacements and variations to the above-mentioned embodiments within the scope of the present application.

Claims

1. A method of integrating satellite and terrestrial data photovoltaic output forecasts, characterized in that, The method comprises the following steps: S10: preprocessing data, including remote sensing inversion radiation data, ground observation radiation data and satellite remote sensing data; S20: spatial grid division based on the spatially continuous remote sensing inversion radiation data, the grid division is based on geographical location and meteorological characteristics, and regions with different meteorological characteristics are divided into different grid units, each grid unit containing at least one ground observation station; S30: establishing a solar radiation prediction model for each grid unit, and training in combination with the ground observation radiation data and the satellite remote sensing data; S40: calculating the solar radiation prediction value of each grid according to the trained solar radiation prediction model in combination with real-time satellite cloud images and the ground observation radiation data; S50: calculating the power output of a photovoltaic system in a target time period based on the solar radiation prediction value and using a photovoltaic system simulation tool.

2. The method of claim 1, wherein the integrated satellite and terrestrial data photovoltaic power output prediction is based on a satellite and terrestrial data photovoltaic power output prediction model. In step S10, the data is preprocessed, including: S11: quality control inspection of the remote sensing inversion radiation data, comparison with ground observation true data, elimination of data records with an absolute difference greater than 20%, and filling of the eliminated data with inversion multi-year average values; S12: quality detection of the ground observation radiation data, elimination of noise and missing values by filtering and interpolation preprocessing technology, detection and elimination of abnormal values by statistical methods and physical models, and calibration and verification of the data to ensure its accuracy and consistency; S13: geometric and radiation correction of satellite remote sensing data to eliminate errors caused by instruments and observation angles, application of quality evaluation indicators to screen contaminated or distorted data, and filling of contaminated areas based on adjacent observation images; S14: matching of the processed remote sensing inversion radiation data, ground observation radiation data and satellite remote sensing data through time and space information for coordinates and time information to form a paired sequence of satellite data and ground data.

3. The method of claim 1, wherein the method further comprises: In step S20, the spatial grid division is based on the spatially continuous remote sensing inversion radiation data, and the step comprises: S21: constructing an infinite Gaussian mixture model, initializing the mean and covariance matrix of the infinite Gaussian mixture model, estimating the parameters by Gibbs sampling, calculating the posterior probability of each data point belonging to each Gaussian distribution, and updating the parameter values to convergence; S22: using an adaptive method to determine the optimal cluster number in the data based on the output results of the infinite Gaussian mixture model; S23: inputting the preprocessed data into the infinite Gaussian mixture model fitted with the optimal cluster number to predict the partition to which each pixel belongs, forming a spatially continuous partition prediction result; S24: edge detection of the partition prediction result, extraction of the boundary line and smoothing processing, determination of the accurate boundary and range of each partition, and conversion of the same into spatial coordinates for storage.

4. The method of claim 3, wherein the integrated satellite and terrestrial data photovoltaic power output prediction is based on a satellite and terrestrial data photovoltaic power output prediction model. In step S22, the adaptive method is used to determine the optimal cluster number in the data, and the adaptive method is realized by a Bayesian information criterion, and the model of the Bayesian information criterion comprises: where BIC is the Bayesian Information Criterion, n is the number of samples, and k is the number of parameters of the model, is the maximum likelihood estimate of the model.

5. The method of claim 1, wherein the method further comprises: In step S30, a solar radiation prediction model is established for each grid cell, and the step includes: S31: In the unit of each spatial grid cell, the ground observation radiation data and the satellite remote sensing data are matched according to the geographic coordinates and the time stamp of the subarea to form paired data; S32: Based on the paired data, a solar radiation prediction model integrating the ground observation radiation data and the satellite remote sensing data is constructed; S33: Training samples including input features and target variables are extracted from historical observation data, the training samples are divided into a training set, a validation set and a test set in a ratio of 3:1:1, and the training set is used to train the solar radiation prediction model to optimize model parameters; S34: The solar radiation prediction model is evaluated using the validation set and the test set, and the performance indicators of the solar radiation prediction model are calculated to ensure that the performance of the solar radiation prediction model on the validation set and the test set is optimal.

6. The method of claim 5, wherein the method further comprises: In step S32, the solar radiation prediction model integrating the ground observation radiation data and the satellite remote sensing data is constructed, including constructing the solar radiation prediction model as: wherein, represents the solar radiation to be predicted; t represents the time at which the latest historical observation is located; m represents the maximum prediction length; f represents the deep learning prediction model; D represents the model input; x i is the solar radiation observation data at the i-th time point; is the visible light satellite image block of size h x w at the i-th time point; T is the length of the input historical sequence.

7. The method of claim 1, wherein the method further comprises: In step S40, the predicted value of solar radiation of each grid is calculated, and the step includes: S41: The geographic coordinates and time period of the point to be predicted are determined, the geographic coordinates include longitude and latitude, and the time period includes year, month, day, hour and minute information; S42: The point to be predicted is mapped into the grid cell that has been divided according to the geographic coordinates, and is matched to the solar radiation prediction model corresponding to the grid cell; S43: The historical observation data of the ground observation station nearest to the geographic position of the point to be predicted in the grid cell corresponding to the point to be predicted is extracted; S44: A sequence of visible light cloud images of T hours before the prediction time period is extracted from the satellite image with the point to be predicted as the center, and is paired with the historical observation data to form a data pair meeting the input format requirements of the prediction model; S45: The paired data pair is input into the solar radiation prediction model to generate the prediction result of the solar radiation of the point to be predicted.

8. The method of claim 1, wherein the method further comprises: In step S50, a photovoltaic system simulation tool is used to calculate the power output of the photovoltaic system in the target time period, and the step includes: S51: A photovoltaic power generation simulator is selected; the operating environment of the photovoltaic power generation simulator is configured, including dependent libraries and initial parameters; and the basic configuration of the photovoltaic power generation simulator is set by calling an initialization function; S52: The parameters of the photovoltaic system are configured, including component parameters, installation parameters and geographic location; S53: Meteorological data of the photovoltaic power station location is input, and the meteorological data at least includes predicted solar radiation and near-real-time observed temperature; S54: The direct current output of the photovoltaic component in the target time period is calculated by the photovoltaic power generation simulator using the configured parameters of the photovoltaic system and the input meteorological data; S55: configuring parameters of a photovoltaic power station inverter model, the parameters of the photovoltaic power station inverter model at least including maximum direct current input power, maximum alternating current output power, and efficiency; S56: inputting the calculated direct current power generation of the photovoltaic module into the inverter model, performing direct current to alternating current conversion calculation through the inverter model, and obtaining alternating current output of the photovoltaic system in a target time period; S57: outputting the calculated alternating current output and performing necessary verification and evaluation.

9. A device for integrating satellite and terrestrial data photovoltaic output forecasts, characterized in that, The method comprises the following steps: a data preprocessing unit configured to preprocess data, the data comprising remote sensing inversion radiation data, ground observation radiation data, and satellite remote sensing data; a spatial grid division unit configured to divide a spatial grid based on the spatially continuous remote sensing inversion radiation data, the grid division being based on geographical location and meteorological characteristics, and dividing regions with different meteorological characteristics into different grid units, each grid unit containing at least one ground observation station; a subarea prediction modeling unit configured to establish a solar radiation prediction model for each grid unit, and train the model in combination with the ground observation radiation data and satellite remote sensing data; a solar radiation prediction unit configured to calculate a solar radiation prediction value for each grid based on the trained solar radiation prediction model, in combination with real-time satellite cloud images and the ground observation radiation data; a photovoltaic output calculation unit configured to calculate power output of a photovoltaic system in a target time period based on the solar radiation prediction value and using a photovoltaic system simulation tool.

10. A computer device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor executes the computer program to implement the method of any one of claims 1-8.

11. A storage medium having stored thereon a computer program, characterized in that The computer program is executed by the processor to implement the method of any one of claims 1-8.

Images