Multi-source new energy output prediction method and device

Abstract

The embodiment of the invention provides a multi-source new energy output prediction method and device, and the method comprises the steps: carrying out the confidence interval evaluation and abnormality correction of multi-source meteorological data through an error time sequence model, and actively selecting an optimal meteorological source on the D-1 day through an AI algorithm; meteorological data space distribution features are extracted by adopting a color gradation feature recognition technology, and multi-modal accurate matching of numerical forecasting and geographic coordinates is realized in combination with a physical significance mapping model; constructing a hybrid prediction model of coupling data driving and a physical mechanism, and generating a new energy power generation power prediction curve; and iteratively optimizing meteorological source evaluation parameters and model weights through a dynamic feedback mechanism of a D + 1 day prediction result and actual output. According to the method, the technical bottleneck of multi-source data standardization fusion is broken through, the meteorological data utilization efficiency and the interpretability of the prediction model are improved, and high-reliability decision support is provided for new energy output prediction in a power transaction scene.

Description

Technical Field

[0001] The present application relates to the technical field of power prediction of renewable energy in power systems, and in particular to a method and device for predicting the output of renewable energy from multiple sources. Background Art

[0002] In modern power systems, renewable energy output forecasting is the core technical support for power trading, directly affecting the formulation of trading strategies and the security of grid dispatch. With the surge in wind and photovoltaic power generation, the high volatility of their output has led to a sharp increase in the risk of trading deviation penalties. Current forecasting systems are highly dependent on multi-source meteorological data (such as ECMWF, GFS, and regional numerical models), but the characteristics of meteorological sources vary significantly: the temporal and spatial resolutions are not uniform (hourly to minute-level), the forecast reliability fluctuates greatly, and the data formats are heterogeneous. In the context of electricity spot trading, existing forecasting methods are unable to meet the requirements of high accuracy, interpretability, and dynamic adaptability, resulting in rising losses and balancing costs due to wind and solar curtailment. There is an urgent need to overcome the technical bottlenecks of multi-source data fusion and physical mechanism embedding.

[0003] Currently, data-driven models are widely used in the field of renewable energy output forecasting. Most models employ purely data-driven algorithms such as LSTM and XGBoost to train historical weather-output mapping relationships. These models also rely on manual experience or static screening of meteorological sources based on historical average errors. This results in the following shortcomings: 1. Low meteorological utilization efficiency, as the time-varying reliability of meteorological sources is not quantified. Fixed weights or subjective optimization lead to amplified errors, increasing the failure rate of forecasts during extreme weather periods. 2. Model interpretability is poor; purely black-box models fail to correlate forecast results with geographic and physical boundaries, making it difficult for traders to assess risk. Summary of the Invention

[0004] In view of this, embodiments of the present application provide a multi-source renewable energy output forecasting method. One or more embodiments of the present application also involve a multi-source renewable energy output forecasting apparatus, a computing device, a computer-readable storage medium, and a computer program to address technical deficiencies in the prior art, such as significant differences in the characteristics of multi-source meteorological data in the context of power trading, inconsistent meteorological data granularity, insufficient forecast reliability, and poor interpretability of machine learning models.

[0005] According to a first aspect of an embodiment of the present application, a method for predicting the output of multi-source renewable energy is provided, comprising: Collect historical meteorological data from multiple meteorological sources in the same area and the same time period, and use the historical meteorological data of each meteorological source to obtain the dynamic error weight parameters of each meteorological source in different time periods; Using the dynamic error weight parameters of each meteorological source in different time periods, an AI reinforcement learning multi-objective optimization model for multiple meteorological sources is constructed, and the optimal meteorological source combination in each time period is obtained by solving the AI reinforcement learning multi-objective optimization model for multiple meteorological sources; Acquire meteorological pictures from the meteorological sources corresponding to each time period of the optimal meteorological source combination, and extract meteorological element information, meteorological element image coordinates, and meteorological element color scale parameters from each meteorological picture using color scale feature recognition technology; By spatially matching the meteorological element image coordinates in each time period of the optimal meteorological source combination with the target geographical area coordinates and the new energy installed capacity information, a multimodal fusion meteorological dataset in each time period of the optimal meteorological source combination is obtained; The hybrid prediction models corresponding to the meteorological element information in the multimodal fusion meteorological data set in each time period of the optimal meteorological source combination are determined respectively, and the output prediction results of the multi-source renewable energy are obtained by inputting the multimodal fusion meteorological data set in each time period of the optimal meteorological source combination into the corresponding hybrid prediction model.

[0006] Preferably, the method of using the historical meteorological data of each meteorological source to obtain the dynamic error weight parameter of each meteorological source in different time periods includes: The dynamic confidence intervals of each meteorological source in different time periods are constructed by using the forecast values and measured values of each meteorological element data in different time periods contained in the historical meteorological data of each meteorological source; The dynamic confidence intervals of each meteorological source in different time periods are used to quantitatively evaluate the error of each meteorological source, and the dynamic error weight parameters of each meteorological source in different time periods are obtained.

[0007] Preferably, the constructing of dynamic confidence intervals for each meteorological source in different time periods using the forecast values and measured values of each meteorological element data in different time periods contained in the historical meteorological data of each meteorological source comprises: Calculating the historical forecast error sequence of each meteorological source in different time periods based on the forecast value and the measured value of each meteorological element data contained in the historical meteorological data of each meteorological source in different time periods; The quantile regression method is used to model the distribution characteristics of the historical forecast error series of each meteorological source in different time periods, and the upper quantile and lower quantile of the historical forecast error of each meteorological source in different time periods at the confidence level are calculated respectively; According to the upper quantile and lower quantile of each meteorological source in different time periods, the dynamic confidence interval of each meteorological source in different time periods is constructed.

[0008] Preferably, the dynamic confidence interval of each meteorological source in different time periods is used to quantitatively evaluate the error of each meteorological source, and the dynamic error weight parameter of each meteorological source in different time periods is obtained, which includes: According to the dynamic confidence intervals of each meteorological source in different time periods, counting the number of times that the historical forecast error of each meteorological source in different time periods falls into the dynamic confidence interval of its corresponding time period; According to the number of times that the historical forecast error of each meteorological source in different time periods falls into the dynamic confidence interval of its corresponding time period, the actual coverage rate of each meteorological source is calculated, and based on the actual coverage rate of each meteorological source, the dynamic error weight parameter of each meteorological source in different time periods is obtained.

[0009] Preferably, the AI reinforcement learning multi-objective optimization model for multiple meteorological sources includes:

[0010] Where, Indicates the time period t The activation status of the jth historical meteorological data in the meteorological source; 、 、 is the weight coefficient and ; Indicates the time period t Dynamic error weight parameter of the jth historical meteorological data in a meteorological source; Indicates the time period t The actual coverage rate of the jth historical meteorological data in the meteorological source; Respectively represent the tth time period The mean and variance of the jth historical meteorological data in a meteorological source.

[0011] Preferably, the method of extracting meteorological element information, meteorological element image coordinates, and meteorological element color scale parameters from each meteorological image using color scale feature recognition technology includes: Dividing the color scale intervals for representing the RGB or HSV color value range of each color scale, and establishing a mapping function between the color scale and the meteorological element based on the color scale intervals; Extracting meteorological element information and meteorological element color scale parameters from each meteorological image according to the mapping function between color scale and meteorological element; Using color scale feature recognition technology, the coordinates within each color scale block are extracted from each meteorological image, and by defining the original image coordinate reference point of each meteorological image, a coordinate data set of each pixel point in each color scale block is generated, and the coordinate data set of each pixel point in each color scale block is used as the meteorological element image coordinates.

[0012] Preferably, obtaining the multimodal fusion meteorological dataset for each time period of the optimal meteorological source combination by spatially matching the meteorological element image coordinates for each time period of the optimal meteorological source combination with the target geographical area coordinates and the new energy installed capacity information includes: Define the conversion relationship between the meteorological element image coordinate system of each meteorological image and the coordinate system of the target geographical area; By using the conversion relationship of each meteorological image, spatially matching and locating the meteorological element image coordinates of each meteorological image with the corresponding target geographical area coordinates, to obtain the target geographical area coordinates of each meteorological image; By fusing the meteorological element information, meteorological element color scale parameters, meteorological element image coordinates and target geographical area coordinates of each meteorological image, a multimodal fusion meteorological dataset for each time period of the optimal meteorological source combination is obtained.

[0013] According to a second aspect of an embodiment of the present application, a multi-source new energy output prediction device is provided, comprising: The dynamic error weight parameter acquisition module is configured to collect historical meteorological data from multiple meteorological sources for the same area and the same time period, and obtain dynamic error weight parameters for each meteorological source in different time periods using the historical meteorological data of each meteorological source; The optimal meteorological source combination module is configured to use the dynamic error weight parameters of each meteorological source in different time periods to build an AI reinforcement learning multi-objective optimization model for multiple meteorological sources, and obtain the optimal meteorological source combination in each time period by solving the AI reinforcement learning multi-objective optimization model for multiple meteorological sources; an extraction module configured to obtain meteorological images from the meteorological sources corresponding to each time period of the optimal meteorological source combination, and extract meteorological element information, meteorological element image coordinates, and meteorological element color scale parameters from each meteorological image using color scale feature recognition technology; a spatial matching module configured to obtain a multimodal fusion meteorological dataset for each time period of the optimal meteorological source combination by spatially matching the meteorological element image coordinates for each time period of the optimal meteorological source combination with the target geographical area coordinates and the new energy installed capacity information; The output prediction module is configured to respectively determine the hybrid prediction models corresponding to the meteorological element information in the multimodal fusion meteorological data set in each time period of the optimal meteorological source combination, and obtain the output prediction results of the multi-source new energy by inputting the multimodal fusion meteorological data set in each time period of the optimal meteorological source combination into the corresponding hybrid prediction model.

[0014] According to a third aspect of an embodiment of the present application, a computing device is provided, including: memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement any step of the multi-source new energy output prediction method.

[0015] According to a fourth aspect of the embodiments of the present application, a computer-readable storage medium is provided, which stores computer-executable instructions. When the instructions are executed by a processor, the steps of any one of the multi-source new energy output prediction methods are implemented.

[0016] According to a fifth aspect of the embodiments of the present application, a computer program is provided, wherein when the computer program is executed in a computer, the computer is caused to execute the steps of the above-mentioned multi-source new energy output prediction method.

[0017] The multi-source renewable energy output forecasting solution provided in the embodiment of the present application uses an error time series model to perform confidence interval assessment and anomaly correction on multi-source meteorological data, and uses an AI algorithm to actively select the best meteorological source on D-1 day; uses color scale feature recognition technology to extract the spatial distribution characteristics of meteorological data, and combines the physical meaning mapping model to achieve multi-modal precision matching of numerical forecasts and geographic coordinates; constructs a hybrid prediction model that couples data drive and physical mechanism to generate a new energy power generation power forecast curve; through a dynamic feedback mechanism between the D+1 day forecast results and the actual output, it iteratively optimizes the meteorological source assessment parameters and model weights. The present invention breaks through the technical bottleneck of standardized fusion of multi-source data, improves the efficiency of meteorological data utilization and the interpretability of the prediction model, and provides highly reliable decision support for the prediction of renewable energy output in power trading scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 This is a flow chart of a multi-source renewable energy output prediction method provided by one embodiment of the present application; Figure 2 This is a schematic diagram of a multi-source new energy output prediction device provided by one embodiment of the present application; Figure 3 This is an overall flow chart of a multi-source renewable energy output prediction method provided by an embodiment of the present application; Figure 4 This is a structural block diagram of a computing device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0019] The following description sets forth many specific details to facilitate a thorough understanding of the present application. However, the present application can be implemented in many other ways than those described herein, and those skilled in the art can make similar generalizations without violating the scope of the present application. Therefore, the present application is not limited to the specific implementations disclosed below.

[0020] The terms used in one or more embodiments of the present application are for the purpose of describing specific embodiments only and are not intended to limit one or more embodiments of the present application. The singular forms "a", "the" and "the" used in one or more embodiments of the present application and the appended claims are also intended to include plural forms, unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used in one or more embodiments of the present application refers to and includes any or all possible combinations of one or more associated listed items.

[0021] Figure 1 A flowchart of a multi-source renewable energy output prediction method provided according to an embodiment of the present application is shown, which specifically includes the following steps.

[0022] Step S101: collecting historical meteorological data from multiple meteorological sources in the same area and the same time period, and using the historical meteorological data of each meteorological source to obtain dynamic error weight parameters for each meteorological source in different time periods; In one embodiment of the present application, the use of historical meteorological data of each meteorological source to obtain dynamic error weight parameters for each meteorological source in different time periods includes: using the forecast value and measured value of each meteorological element data contained in the historical meteorological data of each meteorological source in different time periods to construct a dynamic confidence interval for each meteorological source in different time periods; using the dynamic confidence interval for each meteorological source in different time periods to quantitatively evaluate the error of each meteorological source, and obtain the dynamic error weight parameters for each meteorological source in different time periods.

[0023] In one embodiment of the present application, the use of the forecast value and measured value of each meteorological element data contained in the historical meteorological data of each meteorological source in different time periods to construct the dynamic confidence interval of each meteorological source in different time periods includes: calculating the historical forecast error sequence of each meteorological source in different time periods based on the forecast value and measured value of each meteorological element data contained in the historical meteorological data of each meteorological source in different time periods; using the quantile regression method to model the distribution characteristics of the historical forecast error sequence of each meteorological source in different time periods, and respectively calculating the upper quantile and lower quantile of the historical forecast error of each meteorological source in different time periods at the confidence level; and constructing the dynamic confidence interval of each meteorological source in different time periods based on the upper quantile and lower quantile of each meteorological source in different time periods.

[0024] In one embodiment of the present application, the dynamic confidence interval of each meteorological source in different time periods is used to quantitatively evaluate the error of each meteorological source to obtain the dynamic error weight parameter of each meteorological source in different time periods, including: according to the dynamic confidence interval of each meteorological source in different time periods, counting the number of times the historical forecast error of each meteorological source in different time periods falls into the dynamic confidence interval of its corresponding time period; calculating the actual coverage rate of each meteorological source according to the number of times the historical forecast error of each meteorological source in different time periods falls into the dynamic confidence interval of its corresponding time period, and obtaining the dynamic error weight parameter of each meteorological source in different time periods based on the actual coverage rate of each meteorological source.

[0025] Step S102: constructing an AI reinforcement learning multi-objective optimization model for multiple meteorological sources using the dynamic error weight parameters of each meteorological source in different time periods, and obtaining the optimal meteorological source combination in each time period by solving the AI reinforcement learning multi-objective optimization model for multiple meteorological sources; In one embodiment of the present application, the AI reinforcement learning multi-objective optimization model for multiple meteorological sources includes:

[0026] Where, Indicates the time period t The activation status of the jth historical meteorological data in the meteorological source; 、 、 is the weight coefficient and ; Indicates the time period t Dynamic error weight parameter of the jth historical meteorological data in a meteorological source; Indicates the time period t The actual coverage rate of the jth historical meteorological data in the meteorological source; Respectively represent the tth time period The mean and variance of the jth historical meteorological data in a meteorological source.

[0027] Step S103: obtaining meteorological pictures from the meteorological sources corresponding to each time period of the optimal meteorological source combination, and extracting meteorological element information, meteorological element image coordinates, and meteorological element color scale parameters from each meteorological picture using color scale feature recognition technology; In one embodiment of the present application, the use of color scale feature recognition technology to extract meteorological element information, meteorological element image coordinates and meteorological element color scale parameters from each meteorological picture includes: dividing the color scale intervals used to represent the RGB or HSV color value range of each color scale, and establishing a mapping function between color scales and meteorological elements based on the color scale intervals; extracting meteorological element information and meteorological element color scale parameters from each meteorological picture according to the mapping function between color scales and meteorological elements; using color scale feature recognition technology to extract the coordinates within each color scale block from each meteorological picture, and by defining the original image coordinate reference point of each meteorological picture, generating a coordinate data set for each pixel point in each color scale block, and using the coordinate data set for each pixel point in each color scale block as the meteorological element image coordinates.

[0028] Step S104: spatially matching the meteorological element image coordinates in each time period of the optimal meteorological source combination with the target geographical area coordinates and the new energy installed capacity information to obtain a multimodal fusion meteorological dataset in each time period of the optimal meteorological source combination; In one embodiment of the present application, the multimodal fusion meteorological dataset for each time period of the optimal meteorological source combination is obtained by spatially matching the meteorological element image coordinates in each time period of the optimal meteorological source combination with the target geographical area coordinates and the new energy installed capacity information, including: defining the conversion relationship between the meteorological element image coordinate system of each meteorological picture and the target geographical area coordinate system; spatially matching and positioning the meteorological element image coordinates of each meteorological picture with its corresponding target geographical area coordinates through the conversion relationship of each meteorological picture to obtain the target geographical area coordinates of each meteorological picture; and obtaining the multimodal fusion meteorological dataset for each time period of the optimal meteorological source combination by fusing the meteorological element information, meteorological element color scale parameters, meteorological element image coordinates and target geographical area coordinates of each meteorological picture.

[0029] Step S105: Determine the hybrid prediction models corresponding to the meteorological element information in the multimodal fusion meteorological data set in each time period of the optimal meteorological source combination, and obtain the output prediction results of the multi-source renewable energy by inputting the multimodal fusion meteorological data set in each time period of the optimal meteorological source combination into the corresponding hybrid prediction model.

[0030] The following combined Figure 2 Taking the application of the multi-source renewable energy output prediction method provided by this application in the field of renewable energy power prediction of power systems as an example, the multi-source renewable energy output prediction method is further explained. Figure 2 A flowchart of a processing process of a multi-source renewable energy output prediction method provided by an embodiment of the present application is shown, which specifically includes the following steps.

[0031] S1: Collect historical meteorological data from various meteorological sources, process outliers and missing values in the data, and build an error time series model. Evaluate and analyze the errors of each meteorological source within a certain confidence interval; specifically, it includes: S11: Collect historical meteorological data from multiple meteorological sources for the same region or area during the same period, including key meteorological elements such as wind speed, temperature, and irradiance; use a sliding window statistical method to detect local outliers; modify or fill in data through interpolation to generate standardized historical meteorological time series data; S12: For each meteorological source, define its historical forecast error sequence as

[0032] Where, For the Time period Among the meteorological sources Historical weather forecast values, For the Time period Among the meteorological sources Historical meteorological measurements; For the Time period Among the meteorological sources The forecast error of historical meteorological data.

[0033] The error series is analyzed using quantile regression method. The distribution characteristics of At the confidence level Upper and lower quantiles , .

[0034]

[0035]

[0036] Set different confidence levels according to different error distributions in different time periods , corresponding to the generation of different quantile intervals , which is different from the fixed interval method and can dynamically adapt to the time-varying characteristics of the error.

[0037] For example, the confidence level is 95%, that is ,but 1-5% / 2=97.5%, It is 5% / 2=2.5%.

[0038] S13: For each meteorological source, use the time point backtracking method to verify its historical error value Whether it falls into the dynamic confidence interval at the corresponding moment , count the number of times it falls , calculate the actual coverage:

[0039] Where, For the During the period Meteorological source The total number of times the error value of a historical meteorological event falls into its corresponding confidence interval, For the During the period Meteorological source The total number of data for each historical weather event under this weather data; Indicates the time period t The actual coverage rate of the jth historical weather in the weather sources.

[0040] Construct a dynamic weight function:

[0041] Where, 、 For the The mean and variance of the error series within the time period; is the smoothing coefficient to avoid the denominator being zero.

[0042] right The weights are normalized to obtain the normalized dynamic weight parameters :

[0043] S14: The normalized dynamic weight parameters It is stored in association with meteorological sources and meteorological data to form a library of optimal parameters. The higher the weight, the better the overall reliability.

[0044]

[0045] S2: On D-1, AI is used to proactively select the best weather source based on the original D-day weather information and evaluation and analysis results. On D-1, based on the original weather forecast data on D-day and the weather data error weight parameters of each weather source, the AI optimization algorithm is used to dynamically screen and adapt weather sources to output the optimal weather source combination.

[0046] Furthermore, AI proactive selection of meteorological sources includes: S21: Construct an AI reinforcement learning multi-objective optimization model for the required meteorological data from multiple meteorological sources:

[0047] Where, express Meteorological sources in the period The enabled state within 、 、 is the weight coefficient; and .

[0048] S22: Select at least Weather sources , through the Q-Learning algorithm, iteratively solves the required meteorological data in 24 time periods of a day to obtain the optimal meteorological source combination

[0049] in, For the Under the historical meteorological Optimal meteorological source for time period selection .

[0050] S3: Preprocess and store meteorological data; Perform standardized format conversion, outlier removal, and storage management on historical meteorological forecast values in meteorological data to generate standardized meteorological data sets; S4: Extracting original meteorological data and corresponding image coordinates using color-scale feature recognition method; Using color scale feature recognition technology, the spatial distribution characteristics of meteorological elements in the optimal meteorological source combination are analyzed and processed to extract the original image coordinates and corresponding color scale parameters. The color scale parameters are the colors corresponding to meteorological data such as wind speed, temperature, and solar radiation intensity. For example, the analyzed color scale parameter is wind speed, with 10m / s corresponding to blue and 20m / s corresponding to red. In other words, the colors in the image are extracted into specific meteorological data and the coordinate values of these colors. Furthermore, each image corresponds to a specific type of meteorological data.

[0051] Furthermore, the spatial distribution characteristics of meteorological data are analyzed and processed including: S41: Divide the color range ,in Indicates the The RGB or HSV color value range of the color level.

[0052] Based on the optimal meteorological source combination forecast data, a color scale-element mapping function is established:

[0053] Where, For color levels The corresponding numerical range of meteorological elements.

[0054] S42: Use image recognition technology to extract the coordinates of each color level block, define the original image coordinate reference point, and generate the first Block coordinate dataset ; Where, For the In the block The horizontal coordinate of the point, For the In the block The vertical coordinate of a point.

[0055] Preset target province Boundary coordinates of counties .

[0056] Where, For the In the block The horizontal coordinate of the point, For the In the block The vertical coordinate of a point.

[0057] S43: Define the conversion relationship between image coordinate system and geographic coordinate system Assume that the original meteorological image resolution is Pixel

[0058]

[0059] Where, 、 They are 、 The corresponding horizontal and vertical coordinates in the geographic coordinate system. 、 It is the maximum and minimum value of the horizontal coordinate in the geographic coordinate system; 、 It is the maximum and minimum value of the vertical coordinate in the geographic coordinate system.

[0060] Through the conversion relationship, the color scale coordinate points are matched and located with the geographical boundaries of the target districts and counties.

[0061] S5: Construct a physical meaning mapping model between image coordinates and provincial and county regional coordinates to complete the fusion of multimodal meteorological data; Construct a physical meaning mapping model between original image coordinates and geographic area coordinates, spatially match the numerical forecast results of meteorological data with the geographic information and new energy installed capacity information of the target province and county area, and generate a multimodal fusion meteorological dataset; S6: Based on the physical correlation analysis of meteorological factors and historical output data, a new energy power generation prediction model that integrates data-driven and physical mechanisms (i.e., a hybrid prediction model) is constructed to achieve the prediction of power generation curves; Analyze the physical correlation between meteorological factors and historical renewable energy output data, build a hybrid prediction model, integrate data-driven characteristics with physical mechanism constraints, and output renewable energy power generation prediction curves; S7: On D+1, compare and analyze the D-day forecast results with the actual output data, and update and improve the data of each meteorological source assessment.

[0062] On D+1, the D-day forecast results are compared and analyzed with the actual output data. The meteorological source assessment parameters and model weights are updated based on error feedback to complete the closed-loop optimization of the forecast system.

[0063] Among them, the output prediction of coupling data drive and physical mechanism includes: S51: Calculate theoretical output power using the installed capacity and conversion coefficient of the corresponding province;

[0064] Where, For the region In time The conversion coefficient is related to the lower limit of the new energy output startup; For the region installed capacity.

[0065]

[0066] Where, for The predicted output value of new energy at all times.

[0067] Figure 3 FIG1 shows a schematic diagram of a multi-source new energy output prediction device provided by an embodiment of the present application. Figure 3 As shown, the device includes: The dynamic error weight parameter acquisition module is configured to collect historical meteorological data from multiple meteorological sources for the same area and the same time period, and obtain dynamic error weight parameters for each meteorological source in different time periods using the historical meteorological data of each meteorological source; The optimal meteorological source combination module is configured to use the dynamic error weight parameters of each meteorological source in different time periods to build an AI reinforcement learning multi-objective optimization model for multiple meteorological sources, and obtain the optimal meteorological source combination in each time period by solving the AI reinforcement learning multi-objective optimization model for multiple meteorological sources; an extraction module configured to obtain meteorological images from the meteorological sources corresponding to each time period of the optimal meteorological source combination, and extract meteorological element information, meteorological element image coordinates, and meteorological element color scale parameters from each meteorological image using color scale feature recognition technology; a spatial matching module configured to obtain a multimodal fusion meteorological dataset for each time period of the optimal meteorological source combination by spatially matching the meteorological element image coordinates for each time period of the optimal meteorological source combination with the target geographical area coordinates and the new energy installed capacity information; The output prediction module is configured to respectively determine the hybrid prediction models corresponding to the meteorological element information in the multimodal fusion meteorological data set in each time period of the optimal meteorological source combination, and obtain the output prediction results of the multi-source new energy by inputting the multimodal fusion meteorological data set in each time period of the optimal meteorological source combination into the corresponding hybrid prediction model.

[0068] The above is a schematic diagram of a multi-source renewable energy output prediction device according to this embodiment. It should be noted that the technical solution of this multi-source renewable energy output prediction device and the technical solution of the multi-source renewable energy output prediction method described above are based on the same concept. For details not described in detail in the technical solution of the multi-source renewable energy output prediction device, please refer to the description of the technical solution of the multi-source renewable energy output prediction method described above.

[0069] Figure 4 The block diagram shows a structure of a computing device 400 according to one embodiment of the present application. The components of the computing device 400 include, but are not limited to, a memory 410 and a processor 420. The processor 420 is connected to the memory 410 via a bus 430, and a database 450 is used to store data.

[0070] Computing device 400 also includes an access device 440 that enables computing device 400 to communicate via one or more networks 460. Examples of such networks include a public switched telephone network (PSTN), a local area network (LAN), a wide area network (WAN), a personal area network (PAN), or a combination of communication networks such as the Internet. Access device 440 may include one or more of any type of network interface (e.g., a network interface card (NIC)), whether wired or wireless, such as an IEEE 802.11 wireless local area network (WLAN) wireless interface, a Worldwide Interoperability for Microwave Access (Wi-MAX) interface, an Ethernet interface, a universal serial bus (USB) interface, a cellular network interface, a Bluetooth interface, a near field communication (NFC) interface, and the like.

[0071] In one embodiment of the present application, the above components of the computing device 400 and Figure 4 Other components not shown in the figure may also be connected to each other, for example, via a bus. Figure 4 The computing device structure block diagram shown is for illustrative purposes only and is not intended to limit the scope of the present application. Those skilled in the art may add or replace other components as needed.

[0072] Computing device 400 can be any type of stationary or mobile computing device, including a mobile computer or mobile computing device (e.g., a tablet computer, personal digital assistant, laptop computer, notebook computer, netbook computer, etc.), a mobile phone (e.g., a smartphone), a wearable computing device (e.g., a smartwatch, smart glasses, etc.), or other types of mobile devices, or a stationary computing device such as a desktop computer or PC. Computing device 400 can also be a mobile or stationary server.

[0073] The processor 420 is configured to execute the following computer-executable instructions, which, when executed by the processor, implement the steps of the multi-source renewable energy output prediction method.

[0074] The above is a schematic diagram of a computing device according to this embodiment. It should be noted that the technical solution of this computing device is based on the same concept as the technical solution of the aforementioned multi-source renewable energy output prediction method. For details not described in detail in the technical solution of the computing device, please refer to the description of the technical solution of the aforementioned multi-source renewable energy output prediction method.

[0075] An embodiment of the present application further provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the multi-source renewable energy output prediction method.

[0076] The above is a schematic diagram of a computer-readable storage medium according to this embodiment. It should be noted that the technical solution of this storage medium is based on the same concept as the technical solution of the aforementioned multi-source renewable energy output prediction method. For details not described in detail in the technical solution of the storage medium, please refer to the description of the technical solution of the aforementioned multi-source renewable energy output prediction method.

[0077] An embodiment of the present application further provides a computer program, wherein when the computer program is executed in a computer, the computer is caused to execute the steps of the above-mentioned multi-source new energy output prediction method.

[0078] The above is an illustrative embodiment of a computer program. It should be noted that the technical solution of this computer program is based on the same concept as the technical solution of the multi-source renewable energy output prediction method described above. For details not described in detail in the technical solution of the computer program, please refer to the description of the technical solution of the multi-source renewable energy output prediction method described above.

[0079] The foregoing description describes specific embodiments of the present application. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims can be performed in an order different from that described in the embodiments and still achieve the desired results. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order shown or the sequential order to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0080] The computer instructions include computer program code, which may be in source code form, object code form, executable file, or some intermediate form. The computer-readable medium may include any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard drive, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal, and software distribution medium. It should be noted that the content of the computer-readable medium may be appropriately increased or decreased based on the requirements of legislation and patent practice within a jurisdiction. For example, in some jurisdictions, based on legislation and patent practice, computer-readable media does not include electric carrier signals and telecommunication signals.

[0081] It should be noted that for the aforementioned method embodiments, for the sake of simplicity of description, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the embodiments of the present application are not limited by the order of the actions described, because according to the embodiments of the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the embodiments of the present application.

[0082] In the above embodiments, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant description of other embodiments.

[0083] The preferred embodiments of the present application disclosed above are intended only to help illustrate the present application. The optional embodiments do not describe all details in detail, nor do they limit the invention to the specific embodiments described. Obviously, many modifications and variations can be made based on the content of the embodiments of the present application. This application selects and specifically describes these embodiments in order to better explain the principles and practical applications of the embodiments of the present application, so that those skilled in the art can better understand and utilize the present application. This application is limited only by the claims and their full scope and equivalents.

Claims

1. A method for predicting the output of multi-source renewable energy, characterized in that: include: Collect historical meteorological data from multiple meteorological sources in the same area and the same time period, and use the historical meteorological data of each meteorological source to obtain the dynamic error weight parameters of each meteorological source in different time periods; Using the dynamic error weight parameters of each meteorological source in different time periods, an AI reinforcement learning multi-objective optimization model for multiple meteorological sources is constructed, and the optimal meteorological source combination in each time period is obtained by solving the AI reinforcement learning multi-objective optimization model for multiple meteorological sources; Acquire meteorological pictures from the meteorological sources corresponding to each time period of the optimal meteorological source combination, and extract meteorological element information, meteorological element image coordinates, and meteorological element color scale parameters from each meteorological picture using color scale feature recognition technology; By spatially matching the meteorological element image coordinates in each time period of the optimal meteorological source combination with the target geographical area coordinates and the new energy installed capacity information, a multimodal fusion meteorological dataset in each time period of the optimal meteorological source combination is obtained; The hybrid prediction models corresponding to the meteorological element information in the multimodal fusion meteorological data set in each time period of the optimal meteorological source combination are determined respectively, and the output prediction results of the multi-source renewable energy are obtained by inputting the multimodal fusion meteorological data set in each time period of the optimal meteorological source combination into the corresponding hybrid prediction model.

2. The method according to claim 1, characterized in that The method of using the historical meteorological data of each meteorological source to obtain the dynamic error weight parameters of each meteorological source in different time periods includes: The dynamic confidence intervals of each meteorological source in different time periods are constructed by using the forecast values and measured values of each meteorological element data in different time periods contained in the historical meteorological data of each meteorological source; The dynamic confidence intervals of each meteorological source in different time periods are used to quantitatively evaluate the error of each meteorological source, and the dynamic error weight parameters of each meteorological source in different time periods are obtained.

3. The method according to claim 2, characterized in that The method of constructing a dynamic confidence interval for each meteorological source in different time periods by using the forecast value and the measured value of each meteorological element data in different time periods contained in the historical meteorological data of each meteorological source includes: Calculating the historical forecast error sequence of each meteorological source in different time periods based on the forecast value and the measured value of each meteorological element data contained in the historical meteorological data of each meteorological source in different time periods; The quantile regression method is used to model the distribution characteristics of the historical forecast error series of each meteorological source in different time periods, and the upper quantile and lower quantile of the historical forecast error of each meteorological source in different time periods at the confidence level are calculated respectively; According to the upper quantile and lower quantile of each meteorological source in different time periods, the dynamic confidence interval of each meteorological source in different time periods is constructed.

4. The method according to claim 3, characterized in that The dynamic confidence interval of each meteorological source in different time periods is used to quantitatively evaluate the error of each meteorological source, and the dynamic error weight parameters of each meteorological source in different time periods are obtained, including: According to the dynamic confidence intervals of each meteorological source in different time periods, counting the number of times that the historical forecast error of each meteorological source in different time periods falls into the dynamic confidence interval of its corresponding time period; According to the number of times that the historical forecast error of each meteorological source in different time periods falls into the dynamic confidence interval of its corresponding time period, the actual coverage rate of each meteorological source is calculated, and based on the actual coverage rate of each meteorological source, the dynamic error weight parameter of each meteorological source in different time periods is obtained.

5. The method according to claim 4, characterized in that The AI reinforcement learning multi-objective optimization model for multiple meteorological sources includes: ； Where, Indicates the time period t The activation status of the jth historical meteorological data in the meteorological source; 、 、 is the weight coefficient and ; Indicates the time period t Dynamic error weight parameter of the jth historical meteorological data in a meteorological source; Indicates the time period t The actual coverage rate of the jth historical meteorological data in the meteorological source; Respectively represent the tth time period The mean and variance of the jth historical meteorological data in a meteorological source.

6. The method according to claim 5, characterized in that The method of using color scale feature recognition technology to extract meteorological element information, meteorological element image coordinates and meteorological element color scale parameters from each meteorological image includes: Dividing the color scale intervals for representing the RGB or HSV color value range of each color scale, and establishing a mapping function between the color scale and the meteorological element based on the color scale intervals; Extracting meteorological element information and meteorological element color scale parameters from each meteorological image according to the mapping function between color scale and meteorological element; Using color scale feature recognition technology, the coordinates within each color scale block are extracted from each meteorological image, and by defining the original image coordinate reference point of each meteorological image, a coordinate data set of each pixel point in each color scale block is generated, and the coordinate data set of each pixel point in each color scale block is used as the meteorological element image coordinates.

7. The method according to claim 6, characterized in that The multimodal fusion meteorological dataset for each time period of the optimal meteorological source combination is obtained by spatially matching the meteorological element image coordinates for each time period of the optimal meteorological source combination with the target geographical area coordinates and the new energy installed capacity information. The multimodal fusion meteorological dataset for each time period of the optimal meteorological source combination includes: Define the conversion relationship between the meteorological element image coordinate system of each meteorological image and the coordinate system of the target geographical area; By using the conversion relationship of each meteorological image, spatially matching and locating the meteorological element image coordinates of each meteorological image with the corresponding target geographical area coordinates, to obtain the target geographical area coordinates of each meteorological image; By fusing the meteorological element information, meteorological element color scale parameters, meteorological element image coordinates and target geographical area coordinates of each meteorological image, a multimodal fusion meteorological dataset for each time period of the optimal meteorological source combination is obtained.

8. A multi-source new energy output prediction device, characterized in that: include: The dynamic error weight parameter acquisition module is configured to collect historical meteorological data from multiple meteorological sources for the same area and the same time period, and obtain dynamic error weight parameters for each meteorological source in different time periods using the historical meteorological data of each meteorological source; The optimal meteorological source combination module is configured to use the dynamic error weight parameters of each meteorological source in different time periods to build an AI reinforcement learning multi-objective optimization model for multiple meteorological sources, and obtain the optimal meteorological source combination in each time period by solving the AI reinforcement learning multi-objective optimization model for multiple meteorological sources; an extraction module configured to obtain meteorological images from the meteorological sources corresponding to each time period of the optimal meteorological source combination, and extract meteorological element information, meteorological element image coordinates, and meteorological element color scale parameters from each meteorological image using color scale feature recognition technology; a spatial matching module configured to obtain a multimodal fusion meteorological dataset for each time period of the optimal meteorological source combination by spatially matching the meteorological element image coordinates for each time period of the optimal meteorological source combination with the target geographical area coordinates and the new energy installed capacity information; The output prediction module is configured to respectively determine the hybrid prediction models corresponding to the meteorological element information in the multimodal fusion meteorological data set in each time period of the optimal meteorological source combination, and obtain the output prediction results of the multi-source new energy by inputting the multimodal fusion meteorological data set in each time period of the optimal meteorological source combination into the corresponding hybrid prediction model.

9. A computing device comprising: memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the multi-source new energy output prediction method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing computer-executable instructions, wherein the computer-executable instructions, when executed by a processor, implement the steps of the multi-source renewable energy output prediction method according to any one of claims 1 to 7.

Images