Day-ahead solar irradiance prediction method, system and device and storage medium

Abstract

The invention discloses a day-ahead solar irradiance prediction method, a day-ahead solar irradiance prediction system, a day-ahead solar irradiance prediction device and a storage medium. The day-ahead solar irradiance prediction method comprises the following steps of obtaining historical data of day-ahead solar irradiance and determining meteorological data corresponding to an optimal parameterization scheme in a WRF model; determining a weather label of the prediction day according to the historical data; according to historical data, meteorological data and weather labels, parameters of the day-ahead solar irradiance prediction network are updated, and a trained solar irradiance prediction model is obtained; and inputting the historical data into the solar irradiance prediction model to obtain the target day-ahead solar irradiance of the prediction day. The method can be widely applied to the technical field of artificial intelligence.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and in particular to a method, system, device and storage medium for predicting daily solar irradiance. Background Art

[0002] Existing WRF model configuration options are numerous, especially in aspects such as the radiation transfer model (RTM), microphysics scheme, and convection scheme, resulting in thousands of possible combinations for users when configuring the model. This not only increases the difficulty of model optimization but also may affect the accuracy of predictions. In addition, most studies focus on GHI prediction on sunny and cloudy days, and there is less research on improving the prediction accuracy under complex weather conditions such as rainy days and cloudy days. The prediction errors of existing methods are still relatively large under these complex weather conditions, especially in high-altitude and complex terrain areas. Therefore, there is currently a lack of adaptive optimization methods for different weather types, resulting in insufficient accuracy and stability of GHI prediction under variable weather conditions. Thus, there are still technical problems to be solved in the related art. Summary of the Invention

[0003] An object of this application is to solve at least to some extent one of the technical problems existing in the prior art.

[0004] To this end, an object of an embodiment of this application is to provide a method, system, device and storage medium for predicting daily solar irradiance, and this solution can improve the accuracy of GHI prediction.

[0005] To achieve the above technical object, the technical solutions adopted in the embodiments of this application include: A method for predicting daily solar irradiance, comprising the following steps: obtaining historical data of daily solar irradiance and determining meteorological data corresponding to the optimal parameterization scheme in the WRF model; determining the weather label of the prediction day according to the historical data; updating the parameters of the daily solar irradiance prediction network according to the historical data, the meteorological data, and the weather label to obtain a trained solar irradiance prediction model; inputting the historical data into the solar irradiance prediction model to obtain the target daily solar irradiance of the prediction day.

[0006] This application can obtain historical data of daily solar irradiance and determine meteorological data corresponding to the optimal parameterization scheme in the WRF model; determine the weather label of the prediction day according to the historical data; update the parameters of the daily solar irradiance prediction network according to the historical data, the meteorological data, and the weather label to obtain a trained solar irradiance prediction model; input the historical data into the solar irradiance prediction model to obtain the target daily solar irradiance of the prediction day. This application can improve the prediction accuracy by adding meteorological data.

[0007] In addition, for a method for predicting daily solar irradiance according to the above embodiments of the present invention, the following additional technical features may also be included:

[0008] Further, in the embodiments of the present application, the determination of the meteorological data corresponding to the optimal parameterization scheme in the WRF model specifically includes:

[0009] In the combined test and optimization stage of all parameterization scheme combinations of the WRF model, a systematic test is performed on the parameterization schemes of the surface and the planetary boundary layer to obtain a number of first optimized parameterization schemes;

[0010] On the basis of fixing the surface and planetary boundary layer schemes, test the combinations of fixed radiation transfer models, microphysical models, and convective parameterization schemes, and screen out the optimal parameterization scheme from the number of first optimized parameterization schemes; wherein, the number of first optimized parameterization schemes is less than the number of all parameterization schemes; the number of optimal parameterization schemes is less than the number of first optimized parameterization schemes;

[0011] The WRF model runs the optimal parameterization scheme to obtain the corresponding meteorological data.

[0012] Further, in the embodiments of the present application, the determination of the weather label of the prediction day according to the historical data specifically includes:

[0013] Determine the daily average cloud cover coefficient according to the historical data;

[0014] Based on a preset neural network model, determine the first weather label corresponding to each daily average cloud cover coefficient;

[0015] Use all the first weather labels as all the weather labels of the prediction day.

[0016] Further, in the embodiments of the present application, the determination of the daily average cloud cover coefficient according to the historical data specifically includes:

[0017] Extract the first irradiance from the historical data;

[0018] Based on the preset theoretical clear-sky irradiance and the first irradiance, determine the daily average cloud cover coefficient.

[0019] Further, in the embodiments of the present application, the determination of the daily average cloud cover coefficient based on the preset theoretical clear-sky irradiance and the first irradiance specifically includes:

[0020] Divide the first irradiance by the theoretical clear-sky irradiance to obtain a first quotient;

[0021] Use the first quotient as the daily average cloud cover coefficient.

[0022] Further, in the embodiment of the present application, the day-ahead solar irradiance prediction network includes a convolutional neural network, a long short-term memory network, and a Transformer model; updating the parameters of the day-ahead solar irradiance prediction network according to the historical data, the meteorological data, and the weather label to obtain a trained solar irradiance prediction model specifically includes:

[0023] Inputting the meteorological data into the convolutional neural network to obtain a spatial feature sequence;

[0024] Inputting the spatial feature sequence and the historical data into the long short-term memory network to obtain the first predicted solar irradiance on the prediction day;

[0025] Inputting the historical data, the first predicted solar irradiance, and the weather label into the Transformer model to obtain the second predicted solar irradiance;

[0026] Updating the parameters of the day-ahead solar irradiance prediction network for several rounds according to the second predicted solar irradiance to obtain a spatial feature sequence.

[0027] Further, in the embodiment of the present application, the method further includes:

[0028] Comparing the target day-ahead solar irradiance with a preset day-ahead solar irradiance and adjusting the parameters of the solar irradiance prediction model according to the comparison result.

[0029] On the other hand, the embodiment of the present application further provides a day-ahead solar irradiance prediction system, including:

[0030] A first processing unit, configured to obtain historical data of day-ahead solar irradiance and determine meteorological data corresponding to the optimal parameterization scheme in the WRF model;

[0031] A second processing unit, configured to determine the weather label on the prediction day according to the historical data;

[0032] A third processing unit, configured to update the parameters of the day-ahead solar irradiance prediction network according to the historical data, the meteorological data, and the weather label to obtain a trained solar irradiance prediction model;

[0033] A fourth processing unit, configured to input the historical data into the solar irradiance prediction model to obtain the target day-ahead solar irradiance on the prediction day.

[0034] On the other hand, the present application further provides a day-ahead solar irradiance prediction device, including:

[0035] At least one processor;

[0036] At least one memory for storing at least one program;

[0037] When the at least one program is executed by the at least one processor, the at least one processor implements a method for predicting the daily solar irradiance as described in any one of the invention contents.

[0038] In addition, the present application also provides a computer-readable storage medium, in which instructions executable by a processor are stored, and the instructions executable by the processor are used to execute a method for predicting the daily solar irradiance as described in any one of the above.

[0039] The advantages and beneficial effects of the present application will be partially given in the following description, partially will become obvious from the following description, or will be understood through the practice of the present application:

[0040] The present application can obtain historical data of the daily solar irradiance and determine the meteorological data corresponding to the optimal parameterization scheme in the WRF model; determine the weather label of the prediction day according to the historical data; update the parameters of the daily solar irradiance prediction network according to the historical data, the meteorological data, and the weather label to obtain a trained solar irradiance prediction model; input the historical data into the solar irradiance prediction model to obtain the target daily solar irradiance of the prediction day. The present application can improve the prediction accuracy by adding meteorological data. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 It is a schematic diagram of the steps of a method for predicting the daily solar irradiance in a specific embodiment of the present invention;

[0042] Figure 2 It is a schematic flowchart of a method for predicting the daily solar irradiance in a specific embodiment of the present invention

[0043] Figure 3 It is a schematic structural diagram of a system for predicting the daily solar irradiance in another specific embodiment of the present invention;

[0044] Figure 4 It is a schematic structural diagram of a device for predicting the daily solar irradiance in a specific embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0045] The following describes in detail the embodiments of the present invention. The principles and processes of the method, system, device, and storage medium for predicting the daily solar irradiance in the embodiments of the present invention are described as follows.

[0046] First, the terms of the present application are explained:

[0047] 1. Numerical Weather Prediction (NWP) Model: A computer simulation that uses current weather conditions to predict future weather. It is the basis of weather forecasting and uses mathematical models to describe processes in the atmosphere.

[0048] 2. WRF Model: Weather Research and Forecasting Model, a widely used regional numerical weather prediction model

[0049] 3. Global Horizontal Irradiance (GHI): The total amount of solar radiation received at the Earth's surface, including direct and diffuse radiation.

[0050] 4. Parameterization Scheme: In numerical models, a simplified method used to handle small-scale processes (such as cloud formation, radiation transfer) that cannot be explicitly resolved.

[0051] 5. Radiation Transfer Model (RTM): A model used to calculate radiation transfer in the atmosphere.

[0052] 6. Microphysics Scheme: A parameterization scheme that describes processes of small particles (such as cloud droplets, raindrops) in the atmosphere.

[0053] 7. Convection Scheme: A parameterization scheme that simulates convection processes in the atmosphere.

[0054] 8. Planetary Boundary Layer (PBL) Scheme: A parameterization scheme that describes the turbulent exchange process between the Earth's surface and the upper atmosphere, affecting the thermal and momentum transfer in the lower atmosphere.

[0055] 9. Root Mean Square Error (RMSE): A statistical metric that measures the deviation between predicted values and actual observations, calculated as the square root of the average of the squared prediction errors.

[0056] 10. Convolutional Neural Network (CNN): A deep learning model that is good at processing data with grid structures (such as images). In GHI prediction, CNN is used to extract spatial features such as cloud distribution and changes, improving the spatial resolution and accuracy of predictions.

[0057] 11. Long Short-Term Memory (LSTM): A type of Recurrent Neural Network (RNN) that can capture long-term dependencies in time series data. In GHI prediction, LSTM is used to model the temporal dynamics of GHI, improving the accuracy of short-term and medium-term predictions.

[0058] 12. Transformer Model: A deep learning model based on self-attention mechanism, which is good at dealing with long-term dependencies and multi-variable input data. In GHI prediction, the Transformer model is used to capture complex spatio-temporal relationships and improve the accuracy of prediction.

[0059] 13. Reinforcement Learning (RL): A machine learning method that learns optimal strategies through interaction with the environment.

[0060] 14. Automated Machine Learning (AutoML): A technology for automatically searching and optimizing the hyperparameters and architectures of machine learning models, aiming to reduce the complexity and time cost of manual hyperparameter tuning.

[0061] There are a large number of existing WRF model configuration options, especially in aspects such as the Radiation Transfer Model (RTM), microphysics scheme, and convection scheme, resulting in thousands of possible combinations for users when configuring the model. This not only increases the difficulty of model optimization but also may affect the accuracy of prediction. In addition, most studies focus on GHI prediction for sunny and cloudy days, and there is less research on improving the prediction accuracy under complex weather conditions such as rainy days and cloudy days. The prediction errors of existing methods are still large under these complex weather conditions, especially in high-altitude and complex terrain areas. Therefore, there is currently a lack of adaptive optimization methods for different weather types, resulting in insufficient accuracy and stability of GHI prediction under variable weather conditions. Thus, there are still technical problems to be solved in the related technologies.

[0062] In view of the above defects of the existing technologies, referring to Figure 1 , this application provides a method for predicting daily solar irradiance. In Figure 1 , the method may include the following steps S101 - step S104.

[0063] S101. Obtain historical data of daily solar irradiance and determine the meteorological data corresponding to the optimal parameterization scheme in the WRF model.

[0064] S102. Determine the weather label of the prediction day according to the historical data.

[0065] S103. Update the parameters of the daily solar irradiance prediction network according to the historical data, meteorological data, and weather label to obtain a trained solar irradiance prediction model.

[0066] S104. Input the historical data into the solar irradiance prediction model to obtain the target daily solar irradiance of the prediction day.

[0067] In some possible embodiments of the present application, the processor may be connected to the acquisition unit through a wired or wireless connection. After establishing the connection, the processor may obtain the historical data of the daily solar irradiance from the acquisition unit and determine the meteorological data corresponding to the optimal parameterization scheme in the WRF model. According to the historical data, the processor may determine the weather label of the prediction day. According to the historical data, meteorological data, and weather label, the parameters of the daily solar irradiance prediction network are updated to obtain a trained solar irradiance prediction model. The historical data is input into the solar irradiance prediction model to obtain the target daily solar irradiance of the prediction day.

[0068] It should be noted that the above-mentioned wired connection methods may include the connection between the mobile device and the processing module, and may also include the connection between the processing module and the hardware device, as well as the wired connection between other currently known or future-developed devices and the processing module; and the above-mentioned wireless connection methods may include, but are not limited to, 3G / 4G / 5G connections, WiFi connections, Bluetooth connections, WiMAX connections, Zigbee connections, UWB (Ultra Wide Band) connections, and other currently known or future-developed wireless connection methods.

[0069] The present application can obtain the historical data of the daily solar irradiance and determine the meteorological data corresponding to the optimal parameterization scheme in the WRF model; according to the historical data, determine the weather label of the prediction day; according to the historical data, meteorological data, and weather label, update the parameters of the daily solar irradiance prediction network to obtain a trained solar irradiance prediction model; input the historical data into the solar irradiance prediction model to obtain the target daily solar irradiance of the prediction day. The present application can improve the prediction accuracy by adding meteorological data.

[0070] Further, in the embodiments of the present application, the step of determining the meteorological data corresponding to the optimal parameterization scheme in the WRF model specifically includes steps S201 - S203.

[0071] S201. In the combined test and optimization stage of all parameterization schemes of the WRF model, systematically test the parameterization schemes of the surface and the planetary boundary layer to obtain several first optimized parameterization schemes.

[0072] S202. On the basis of fixing the surface and planetary boundary layer schemes, test the combinations of fixed radiation transfer models, microphysical models, and convective parameterization schemes, and screen out the optimal parameterization scheme from several first optimized parameterization schemes; among them, the number of first optimized parameterization schemes is less than the number of all parameterization schemes; the number of optimal parameterization schemes is less than the number of first optimized parameterization schemes.

[0073] S203. Run the optimal parameterization scheme to obtain the corresponding meteorological data.

[0074] Further, in the embodiment of the present application, the step of determining the weather label of the prediction date according to historical data specifically includes steps S301 - S303.

[0075] S301. Determine the daily average cloud cover coefficient according to historical data.

[0076] S302. Based on a preset neural network model, determine the first weather label corresponding to each daily average cloud cover coefficient.

[0077] S303. Use all the first weather labels as all the weather labels of the prediction date.

[0078] Further, in the embodiment of the present application, the step of determining the daily average cloud cover coefficient according to historical data specifically includes steps S401 - S402.

[0079] S401. Extract the first irradiance in historical data;

[0080] S402. Based on the preset theoretical clear - sky irradiance and the first irradiance, determine the daily average cloud cover coefficient.

[0081] Further, in the embodiment of the present application, the step of determining the daily average cloud cover coefficient based on the preset theoretical clear - sky irradiance and the first irradiance specifically includes steps S501 - S502.

[0082] S501. Divide the first irradiance by the theoretical clear - sky irradiance to obtain the first quotient;

[0083] S502. Use the first quotient as the daily average cloud cover coefficient.

[0084] Further, in the embodiment of the present application, the day - ahead solar irradiance prediction network includes a convolutional neural network, a long short - term memory network, and a Transformer model; the step of updating the parameters of the day - ahead solar irradiance prediction network according to historical data, meteorological data, and weather labels to obtain a trained solar irradiance prediction model specifically includes steps S501 - S504.

[0085] S501. Input the meteorological data into the convolutional neural network to obtain a spatial feature sequence;

[0086] S502. Input the spatial feature sequence and historical data into the long short - term memory network to obtain the first predicted solar irradiance of the prediction date;

[0087] S503. Input the historical data, the first predicted solar irradiance, and the weather label into the Transformer model to obtain the second predicted solar irradiance;

[0088] S504. Update the parameters of the day-ahead solar irradiance prediction network for several rounds according to the second predicted solar irradiance to obtain a spatial feature sequence.

[0089] Further, in the embodiment of the present application, the day-ahead solar irradiance prediction method further includes step S105.

[0090] S105. Compare the target day-ahead solar irradiance with the preset day-ahead solar irradiance, and adjust the parameters of the solar irradiance prediction model according to the comparison result.

[0091] The following combines the attached Figure 2 to illustrate the principle of the present application.

[0092] Refer to Figure 2 , this embodiment proposes a day-ahead solar irradiance (GHI) prediction method based on the adaptive parameterization optimization of the numerical model WRF and the post-processing of advanced artificial intelligence algorithms. First, collect and organize high-resolution GHI observation data and numerical weather prediction (NWP) model data. The observation data is from the latest meteorological satellites and ground observation stations to ensure that the data has high spatio-temporal resolution (for example, 15-minute interval, 1-kilometer spatial resolution). The NWP model input uses the latest version of the WRF model (such as WRF 4.5), and combines the initial and boundary condition data provided by the European Centre for Medium-Range Weather Forecasts (ECMWF). Data preprocessing includes missing value filling, outlier detection and correction, and data synchronization processing to ensure the integrity and consistency of the input data.

[0093] In the parameterization scheme combination test and optimization stage, first systematically test the surface and planetary boundary layer (PBL) parameterization schemes, a total of 219 combination methods. By fixing the radiation transfer model (RTM), microphysical scheme, and convective scheme, evaluate the root mean square error (RMSE) and bias (Bias) of each combination during different test periods (such as 15 days), and finally select the surface and PBL scheme combination that performs best on most test days, such as the NCEP Global Forecast System surface layer scheme, RUC surface model, and UW (Park and Bretherton) PBL scheme. Subsequently, on the basis of fixing the surface and PBL schemes, test 187 combinations of fixed radiation transfer models, microphysical, and convective parameterization schemes, screen out 130 combinations that are successfully operated in practice, and determine the optimal parameterization combination by comparing the RMSE and Bias of these combinations under various weather conditions, which has a significant impact on the GHI prediction performance.

[0094] Next, based on the observation data and the output of the WRF model (Weather Research and Forecasting model), use advanced artificial intelligence algorithms to accurately classify the weather types. Specifically, first calculate the daily average cloud cover coefficient (Kt ), i.e., where I observed is the irradiance observed or predicted by the model, and I clearsky is the theoretical clear-sky irradiance. Then, a deep learning model such as a convolutional neural network (CNN) is used to classify the K t value, and the predicted days are divided into four categories: sunny days, cloudy days, overcast days, and rainy days.

[0095] Through training, the AI model can automatically identify and classify complex weather patterns, improving the accuracy and robustness of classification. According to the classification results, the optimal WRF parameterization combination corresponding to the weather type is dynamically selected. For example, parameterization scheme A is selected for sunny days, scheme B for cloudy days, scheme C for overcast days, and scheme D for rainy days. This selection process is optimized by a reinforcement learning algorithm to achieve real-time and autonomous adjustment of the parameterization scheme.

[0096] In the post-processing stage of the artificial intelligence algorithm, combined with the latest AI technology, the preliminary GHI prediction results of the WRF model (Weather Research and Forecasting model) are further optimized. First, feature engineering is carried out to extract various meteorological variables related to GHI as input features, including temperature, humidity, wind speed, cloud cover, precipitation, etc., and a multi-dimensional feature set is constructed by combining historical GHI data and real-time observation data. Then, a convolutional neural network (CNN), a long short-term memory network (LSTM), and a Transformer model are used to model spatial features, temporal dynamics, and multivariate associations respectively. Through multi-level feature extraction and spatio-temporal relationship capture, the prediction ability is improved. At the same time, automatic machine learning (AutoML) technology is used to automatically search for and optimize the hyperparameters and architectures of the AI model to ensure the best performance of the model under different weather types. Finally, the trained AI model is applied to real-time prediction data to post-process the output of the WRF model and generate the final GHI prediction results, significantly reducing RMSE and Bias and improving the accuracy and stability of the prediction.

[0097] In the performance evaluation and feedback optimization stage, by comparing with high-resolution observation data, indicators such as root mean square error (RMSE), mean absolute error (MAE), and bias (Bias) are used to evaluate the prediction performance, and the prediction performance of the adaptive parameterization combination combined with AI post-processing is compared with that of a single best configuration, a traditional hybrid model method, and a global model (such as ECMWF) to ensure that the adaptive method can improve the prediction accuracy under different weather types. According to the evaluation results, a reinforcement learning algorithm and an AutoML framework are used to continuously optimize the parameter selection strategy and the training process of the AI model to form a closed-loop feedback mechanism to ensure that the system continuously improves the prediction performance in practical applications.

[0098] Finally, the entire GHI prediction method is integrated into the energy management and power grid dispatching system to support real-time prediction and dynamic adjustment. System integration includes establishing real-time data interfaces with meteorological data sources (such as satellites, ground observation stations, ECMWF) to ensure timely acquisition and update of data; using cloud computing platforms and distributed computing resources to deploy WRF model instances and AI algorithms to support the testing and optimization of large-scale parameterization schemes; and developing a user-friendly interface to provide visual display of real-time GHI prediction results and support user-defined parameter configuration and result analysis. Through the above technical solutions, the present invention can achieve high-precision and strong-stability GHI prediction, meet the high requirements of modern power systems for renewable energy prediction, and has broad application prospects and significant economic benefits.

[0099] In another embodiment, taking the Qinghai region of China as an example, the specific application process of the technical solution of the present invention is shown in detail.

[0100] First, high-resolution GHI observation data and numerical weather prediction (NWP) model data from 2018 to 2023 were collected and sorted. The observation data was sourced from the latest meteorological satellites and ground observation stations to ensure high spatio-temporal resolution (e.g., 15-minute interval, 1-kilometer spatial resolution). The NWP model input used the latest version of the WRF model (such as WRF4.5) and combined the initial and boundary condition data provided by the European Centre for Medium-Range Weather Forecasts (ECMWF). Data preprocessing included missing value filling, outlier detection and correction, and data synchronization processing to ensure the integrity and consistency of the input data.

[0101] In the stage of parametric scheme combination testing and optimization, first, a systematic test of surface and planetary boundary layer (PBL) parametric schemes was carried out, with a total of 219 combination methods. By fixing the radiation transfer model (RTM), microphysical scheme, and convective scheme, the root mean square error (RMSE) and bias (Bias) of each combination during different test periods (such as 15 days) were evaluated. Finally, the combination of surface and planetary boundary layer PBL schemes that performed best on most test days was selected, such as the NCEP Global Forecast System surface layer scheme, RUC surface model, and UW (Park and Bretherton) PBL scheme.

[0102] Subsequently, based on the fixed surface and PBL schemes, 187 combinations of RTM radiation transfer models, microphysical, and convective parametric schemes were tested, and 130 combinations that were successfully implemented in practice were screened out. By comparing the RMSE and Bias of these combinations under various weather conditions, the optimal parametric combination was determined, which had a significant impact on the GHI prediction performance.

[0103] Next, based on the observed data (data measured under the above model) and the WRF model output, advanced artificial intelligence algorithms are used to accurately classify weather types. Specifically, first calculate the daily average cloud cover coefficient (K t ), that is where I observed is the irradiance observed or predicted by the model, and I clearsky is the theoretical clear-sky irradiance.

[0104] Then, use the deep learning model Convolutional Neural Network (CNN) to classify the values, and divide the predicted days into four categories: sunny days, cloudy days, overcast days, and rainy days.

[0105] Through training, the AI model can automatically identify and classify complex weather patterns, significantly improving the accuracy and robustness of classification. In the post-processing stage of the artificial intelligence algorithm, combined with the latest AI technology, the preliminary GHI prediction results of the WRF model are further optimized.

[0106] First, perform feature engineering, extract various meteorological variables related to GHI as input features, including temperature, humidity, wind speed, cloud cover, precipitation, etc., and combine historical GHI data and real-time observation data to construct a multi-dimensional feature set. These feature sets will be used as the input of the subsequent AI model to ensure that the model can fully capture various factors affecting GHI.

[0107] Next, use the Convolutional Neural Network (CNN) to model spatial features. Through multiple layers of convolution and pooling operations, CNN extracts local spatial features in high-resolution spatial meteorological data, such as cloud layer distribution and terrain effects, so as to capture the spatial dependence between geographical locations. These spatial features are processed by multiple layers of convolutional kernels to form high-dimensional feature vectors, providing a solid foundation for subsequent time series modeling.

[0108] Subsequently, use the Long Short-Term Memory Network (LSTM) to model the time dynamics. Through its unique gating mechanism, LSTM effectively captures the trends and patterns of GHI changes over time, especially performing well in short-term and medium-term predictions. LSTM receives the sequence of spatial features from CNN, gradually updates the hidden state, retains historical information and combines the current input to generate predictions of future GHI changes.

[0109] Finally, a Transformer model is used to model the multivariate associations. Through its self-attention mechanism, the Transformer can simultaneously handle the complex relationships among multiple meteorological variables, identify the variables that have a key impact on GHI prediction, and dynamically adjust their weights. Positional encoding ensures that the model understands the order and temporal dependencies of the data, and the feed-forward neural network further extracts deep features. Through the multi-head attention mechanism, the Transformer can integrate information from multiple aspects and improve the overall performance of GHI prediction. After integrating the CNN, LSTM, and Transformer models, a comprehensive feature representation is formed. Through a fully connected layer or other fusion mechanisms, the spatial, temporal, and multivariate features extracted by each model are integrated to generate the final GHI prediction result.

[0110] It can be understood that the GHI output by the LSTM is a short-term prediction based on time series data, which focuses on capturing the temporal dependence relationships of historical GHI and meteorological data. The finally predicted GHI is the result of further modeling and optimization of the spatial, temporal, and multivariate aspects by combining the CNN and Transformer models based on the LSTM prediction result.

[0111] In addition, during the forward propagation of each model for 100 times, automated machine learning (AutoML) techniques are used to automatically search for and optimize the hyperparameters and architectures of each model to ensure optimal performance under different weather types. Through this multi-level and multi-model collaborative working method, the method of the present invention can comprehensively capture various factors affecting GHI, significantly improve the prediction accuracy and robustness, and ensure efficient and accurate GHI prediction under different weather conditions.

[0112] In addition, with reference to Figure 3 , corresponding to the method of Figure 1 , an intraday solar irradiance prediction system is also provided in the embodiments of the present application. The system may include a first processing unit 1001, a second processing unit 1002, a third processing unit 1003, and a fourth processing unit 1004. Among them, the first processing unit 1001 may be used to obtain historical data of intraday solar irradiance and determine the meteorological data corresponding to the optimal parameterization scheme in the WRF model. The second processing unit 1002 may be used to determine the weather label of the prediction day according to the historical data. The third processing unit 1003 may be used to update the parameters of the intraday solar irradiance prediction network according to the historical data, meteorological data, and weather label to obtain a trained solar irradiance prediction model; the fourth processing unit 1004 may be used to input the historical data into the solar irradiance prediction model to obtain the target intraday solar irradiance of the prediction day.

[0113] It should be noted that the first processing unit can be any integrated circuit unit or microprocessor unit obtained by integrating a chip with processing functions and its peripheral circuits through existing integration technologies. The first processing unit and the second processing unit can also be any integrated circuit module or microprocessor module obtained by integrating a chip with processing functions and its peripheral circuits through existing integration technologies. The first processing unit and the second processing unit can further include one or more memories.

[0114] It should be noted that the content in the above embodiments of the method for predicting the daily solar irradiance is applicable to the embodiments of the present system for predicting the daily solar irradiance. The functions specifically implemented in the embodiments of the present system for predicting the daily solar irradiance are the same as those in the above embodiments of the method for predicting the daily solar irradiance, and the beneficial effects achieved are also the same as those achieved in the above embodiments of the method for predicting the daily solar irradiance.

[0115] Corresponding to Figure 1 the method of Figure 4 , an embodiment of the present application further provides a device for predicting the daily solar irradiance, and its specific structure can be referred to

[0116] at least one processor 1011;

[0117] at least one memory 1012, configured to store at least one program;

[0118] When the at least one program is executed by the at least one processor, the at least one processor implements the method for predicting the daily solar irradiance.

[0119] The content in the above method embodiments is applicable to the embodiments of the present device. The functions specifically implemented in the embodiments of the present device are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0120] Corresponding to Figure 1 the method of

[0121] , an embodiment of the present application further provides a computer-readable storage medium, in which processor-executable instructions are stored, and the processor-executable instructions are used to execute the method for predicting the daily solar irradiance when executed by a processor.

[0122] In some alternative embodiments, the functions / operations recited in the block diagrams may not occur in the order noted in the operational illustrations. For example, depending on the functionality / operation involved, two blocks shown in succession may actually be executed substantially concurrently or the blocks may sometimes be executed in reverse order. Further, the embodiments presented and described in the flowcharts of the present application are provided by way of example in order to provide a more thorough understanding of the technology. The disclosed methods are not limited to the operations and logic flows presented herein. Alternative embodiments are contemplated in which the order of various operations is altered and in which sub-operations described as part of a larger operation are performed independently.

[0123] In addition, although the present application has been described in the context of functional modules, it should be understood that, unless otherwise stated to the contrary, one or more of the functions and / or features may be integrated in a single physical device and / or software module, or one or more functions and / or features may be implemented in separate physical devices or software modules. It should also be understood that a detailed discussion of the actual implementation of each module is not necessary for an understanding of the present application. Rather, given the attributes, functions, and internal relationships of the various functional modules in the devices disclosed herein, the actual implementation of the modules will be understood within the ordinary skill of an engineer. Thus, those skilled in the art will be able to implement the present application as set forth in the claims without undue experimentation. It should also be understood that the specific concepts disclosed are merely illustrative and are not intended to limit the scope of the present application, which is determined by the full scope of the appended claims and their equivalents.

[0124] If the functions are implemented in the form of software functional units and sold or used as independent products, they may be stored in a computer-readable storage medium. Based on such understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art or a part of the technical solution may be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several programs for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application. The foregoing storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disc.

[0125] Logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a definitional list of executable programs for implementing logical functions, which can be embodied in any computer-readable medium for use by or in connection with a program execution system, apparatus, or device, such as a computer-based system, a system including a processor, or other systems that can fetch and execute programs from the program execution system, apparatus, or device. As used in this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport the program for use by or in connection with the program execution system, apparatus, or device.

[0126] More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection part (electronic device) having one or more wirings, a portable computer diskette (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 device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium can even be paper or other suitable media on which the program can be printed, as the program can be obtained electronically, for example, by optically scanning the paper or other media, followed by editing, interpretation, or otherwise processing as appropriate, and then storing it in a computer memory.

[0127] It should be understood that various parts of the present application can be implemented by hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable program execution system. For example, if implemented in hardware, as in another embodiment, any one or a combination of the following techniques known in the art can be used: discrete logic circuits having logic gate circuits for implementing logical functions on data signals, application-specific integrated circuits having appropriate combinational logic gate circuits, programmable gate arrays (PGAs), field programmable gate arrays (FPGAs), and the like.

[0128] In the foregoing description of this specification, the descriptions referring to the terms "one embodiment / example", "another embodiment / example", or "certain embodiments / examples", etc., mean that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do 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.

[0129] Although embodiments of the present application have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present application. The scope of the present application is defined by the claims and their equivalents.

[0130] The above has specifically described the preferred embodiments of the present application, but the present application is not limited to the described embodiments. Those skilled in the art can also make various equivalent deformations or substitutions without violating the spirit of the present application, and these equivalent deformations or substitutions are all included within the scope defined by the claims of the present application.

Claims

1. A method for predicting solar irradiance before the day, characterized in that: The following steps are involved: Obtain historical data of solar irradiance in the previous day and determine the meteorological data corresponding to the optimal parameterization scheme in the WRF model; Determine the weather label of the forecast day based on the historical data; According to the historical data, the meteorological data and the weather label, the parameters of the solar irradiance prediction network of the day before are updated to obtain a trained solar irradiance prediction model; The historical data is input into the solar irradiance prediction model to obtain the target solar irradiance before the prediction day.

2. The method for predicting solar irradiance before the day according to claim 1, characterized in that: The determination of meteorological data corresponding to the optimal parameterization scheme in the WRF model specifically includes: In the stage of combined testing and optimization of all parameterization schemes of the WRF model, the parameterization schemes of the surface and the planetary boundary layer are systematically tested to obtain several first optimized parameterization schemes; On the basis of the fixed surface and planetary boundary layer schemes, a combination of a fixed radiation transfer model, a microphysical model and a convection parameterization scheme is tested, and an optimal parameterization scheme is screened out from the plurality of first optimized parameterization schemes; wherein the number of the first optimized parameterization schemes is less than the number of all the parameterization schemes; and the number of the optimal parameterization schemes is less than the number of the first optimized parameterization schemes; The optimal parameterization scheme is run to obtain corresponding meteorological data.

3. The method for predicting solar irradiance before the day according to claim 1, characterized in that: Determining the weather label of the forecast day according to the historical data specifically includes: Determine the daily average cloud cover factor based on the historical data; Determine a first weather label corresponding to each of the daily average cloud cover coefficients based on a preset neural network model; All of the first weather labels are used as all of the weather labels for the predicted day.

4. A method for predicting solar irradiance before the day according to claim 3, characterized in that: Determining the daily average cloud cover coefficient according to the historical data specifically includes: Extracting a first irradiance in the historical data; The daily average cloud cover coefficient is determined based on a preset theoretical clear sky irradiance and the first irradiance.

5. A method for predicting solar irradiance before the day according to claim 4, characterized in that: The determining the daily average cloud cover coefficient based on the preset theoretical clear sky irradiance and the first irradiance specifically includes: Taking the first irradiance and the theoretical clear sky irradiance as a quotient to obtain a first quotient; The first quotient is used as the daily average cloud cover coefficient.

6. The method for predicting solar irradiance before the day according to claim 1, characterized in that: The day-ahead solar irradiance prediction network includes a convolutional neural network, a long short-term memory network, and a Transformer model; the day-ahead solar irradiance prediction network parameters are updated according to the historical data, the meteorological data, and the weather label to obtain a trained solar irradiance prediction model, specifically including: Inputting the meteorological data into the convolutional neural network to obtain a spatial feature sequence; Inputting the spatial feature sequence and the historical data into the long short-term memory network to obtain the first predicted solar irradiance of the predicted day; Inputting the historical data, the first predicted solar irradiance, and the weather label into the Transformer model to obtain a second predicted solar irradiance; According to the second predicted solar irradiance, the parameters of the day-ahead solar irradiance prediction network are updated for several rounds to obtain a spatial feature sequence.

7. The method for predicting solar irradiance before the day according to claim 1, characterized in that: The method further comprises: The target day-ahead solar irradiance is compared with a preset day-ahead solar irradiance, and the parameters of the solar irradiance prediction model are adjusted according to the comparison result.

8. A day-ahead solar irradiance prediction system, characterized in that: include: The first processing unit is used to obtain the historical data of solar irradiance in the previous day and determine the meteorological data corresponding to the optimal parameterization scheme in the WRF model; A second processing unit, configured to determine a weather tag for a forecast day based on the historical data; A third processing unit is used to update the parameters of the day-ahead solar irradiance prediction network according to the historical data, the meteorological data and the weather label to obtain a trained solar irradiance prediction model; The fourth processing unit is used to input the historical data into the solar irradiance prediction model to obtain the solar irradiance of the target day before the prediction day.

9. A device for predicting solar irradiance before the day, characterized in that include: at least one processor; at least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements the day-ahead solar irradiance prediction method as described in any one of claims 1-7.

10. A computer-readable storage medium storing instructions executable by a processor, characterized in that: The processor executable instructions are used to execute a day-ahead solar irradiance prediction method as described in any one of claims 1-7 when executed by the processor.

Images