Distributed photovoltaic power generation load prediction method and system
By acquiring and analyzing photovoltaic power generation data and environmental data, and combining image processing and cluster analysis, load correction coefficients are calculated, which solves the problem of low accuracy in photovoltaic power generation load forecasting, achieves more accurate load forecasting, and meets the needs of power grid management.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- MIAN COUNTY HUANENG ELECTRIC POWER CO LTD
- Filing Date
- 2025-12-01
- Publication Date
- 2026-04-17
AI Technical Summary
Existing photovoltaic power generation load forecasting methods lack in-depth utilization of historical operating data, making it difficult to adapt to complex and ever-changing operating environments. This results in low forecast accuracy and fails to meet the needs of refined scheduling and grid load management.
By acquiring historical photovoltaic power generation data and environmental data, the set of power generation characteristic coefficients and shadow state coefficients are determined. Cluster analysis and image processing techniques are used to calculate load correction coefficients to adjust the predicted power generation load, and a load prediction model is established for accurate prediction.
It enables intelligent assessment and analysis of distributed photovoltaic power stations, improves the accuracy of photovoltaic power generation load forecasting, and meets the refined requirements of grid load management.
Smart Images
Figure CN121886336A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of photovoltaic power generation technology, and more specifically, to a method and system for predicting distributed photovoltaic power generation load. Background Technology
[0002] With the rapid development of global industrialization, conventional energy sources such as oil, coal and natural gas are becoming increasingly depleted. The energy crisis and environmental pressures have led the world to focus on renewable energy sources such as photovoltaics and wind power.
[0003] Existing photovoltaic power generation load methods lack in-depth utilization of historical operating data, making it difficult to adapt to the complex and ever-changing operating environment in practical applications. This results in low accuracy of prediction results, failing to meet the needs of refined scheduling and grid load management. Summary of the Invention
[0004] This invention provides a method and system for predicting distributed photovoltaic (PV) power generation load, to solve the problem of low accuracy in PV power generation load prediction results in the prior art, including: The process involves: acquiring historical photovoltaic (PV) power generation data; determining the power generation characteristic coefficient set for the current PV power plant based on the historical PV power generation data; acquiring current environmental data; determining the predicted power generation load based on the current environmental data and the power generation characteristic coefficient set; acquiring PV panel images during the power generation process; determining the shading state coefficient based on the shading state coefficient; and adjusting the predicted power generation load based on the load correction coefficient.
[0005] Furthermore, the step of determining the power generation characteristic coefficient set of the current photovoltaic power station based on historical photovoltaic power generation data includes: determining the changes in power generation data of each photovoltaic region based on historical photovoltaic power generation data, and determining the power generation quality coefficient of each photovoltaic region based on the changes in power generation data; clustering each photovoltaic region based on the power generation quality coefficient, and dividing each photovoltaic region into corresponding cluster partitions; obtaining the cluster center value of each cluster partition, and establishing the power generation characteristic coefficient set based on the cluster center value of each cluster partition.
[0006] Furthermore, the clustering of each photovoltaic region based on the power generation quality coefficient includes: establishing a sample dataset based on the power generation quality coefficient corresponding to each photovoltaic region; randomly selecting k initial cluster centers from the sample dataset; calculating the Manhattan distance from the power generation quality coefficient in the sample dataset to the initial cluster centers; dividing the photovoltaic region corresponding to the power generation quality coefficient into the corresponding cluster partition based on the Manhattan distance from the power generation quality coefficient in the sample dataset to the initial cluster centers; calculating the average value of the power generation quality coefficient in each cluster partition; recalculating the cluster centers based on the average value of the power generation quality coefficient in each cluster partition; repeating the iterative clustering steps until the cluster centers no longer change or the number of iterations reaches the preset maximum number of iterations, thereby obtaining the clustering result of the photovoltaic region.
[0007] Furthermore, the step of determining the predicted power generation load based on the current environmental data and power generation characteristic coefficient set includes: determining historical environmental data, power generation characteristic coefficient set, and corresponding power generation load based on historical photovoltaic power generation data; establishing a training sample set based on the historical environmental data, power generation characteristic coefficient set, and corresponding power generation load; establishing an initial load prediction model based on the training sample set and training the initial load prediction model to obtain a trained load prediction model; and inputting the current photovoltaic power station's environmental data and power generation characteristic coefficient set into the trained load prediction model to obtain the predicted power generation load of the current photovoltaic power station.
[0008] Furthermore, the environmental data of the photovoltaic power station specifically includes sunlight data and meteorological data.
[0009] Furthermore, determining the shadow state coefficient based on the photovoltaic panel image during the power generation process includes: performing grayscale processing on the photovoltaic panel image during the power generation process to obtain a photovoltaic grayscale image; The photovoltaic grayscale image is segmented to obtain several sub-photovoltaic grayscale images; the average grayscale value of each sub-photovoltaic grayscale image is calculated, and the shadow state coefficient is determined based on the change of the average grayscale value of each sub-photovoltaic grayscale image.
[0010] Further, determining the shadow state coefficient based on the average gray value change of each sub-photovoltaic grayscale image includes: plotting an average gray value change curve based on the average gray value change of each sub-photovoltaic grayscale image; performing curve fitting on the average gray value change curve to obtain an average gray value fitting curve; determining the time required for the average gray value to reach a value greater than a preset shadow average gray value based on the average gray value fitting curve; determining the sub-shadow state coefficient based on the time required for the average gray value to reach a value greater than the preset shadow average gray value; and calculating the average sub-shadow state coefficient of all sub-photovoltaic grayscale images to obtain the shadow state coefficient.
[0011] Furthermore, the determination of the load correction factor based on the shadow state factor includes: determining the load correction factor according to the load correction factor calculation formula, wherein the load correction factor calculation formula is specifically as follows: , in, This is the load correction factor. The preset standard load correction factor, To preset the standard shadow state coefficient, This is the shadow state coefficient. For the preset range coefficient, It is a natural exponential function.
[0012] Furthermore, adjusting the predicted power generation load according to the load correction factor includes multiplying the load correction factor by the predicted power generation load to obtain the adjusted predicted power generation load.
[0013] To achieve the above objectives, the present invention also provides a distributed photovoltaic power generation load forecasting system, comprising: The first module is used to acquire historical photovoltaic power generation data and determine the power generation characteristic coefficient set of the current photovoltaic power station based on the historical photovoltaic power generation data; the second module is used to acquire current environmental data and determine the predicted power generation load based on the current environmental data and the power generation characteristic coefficient set; the third module is used to acquire photovoltaic panel images during the power generation process and determine the shading state coefficient based on the photovoltaic panel images during the power generation process; the fourth module is used to determine the load correction coefficient based on the shading state coefficient and adjust the predicted power generation load based on the load correction coefficient.
[0014] The beneficial effects of this invention are as follows: By applying the above technical solutions, this invention acquires historical photovoltaic (PV) power generation data and determines the current PV power generation characteristic coefficient set based on this data; acquires current environmental data and determines the predicted power generation load based on this data and the PV power generation characteristic coefficient set; acquires PV panel images during the power generation process and determines the shading state coefficient based on these images; determines the load correction coefficient based on the shading state coefficient; and adjusts the predicted power generation load based on the load correction coefficient. This invention, through intelligent evaluation and analysis of relevant data from distributed PV power stations, enables more accurate PV power generation load prediction. Attached Figure Description
[0015] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0016] Figure 1 A general flowchart of a distributed photovoltaic power generation load forecasting method proposed in an embodiment of the present invention is shown; Figure 2 A schematic diagram of a distributed photovoltaic power generation load forecasting system proposed in an embodiment of the present invention is shown. Detailed Implementation
[0017] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0018] This application provides a method for predicting distributed photovoltaic power generation load, such as... Figure 1 As shown, it includes: S101, Obtain historical photovoltaic power generation data, and determine the power generation characteristic coefficient set of the current photovoltaic power station based on the historical photovoltaic power generation data; In some embodiments of this application, determining the power generation characteristic coefficient set of the current photovoltaic power station based on historical photovoltaic power generation data includes: determining the changes in power generation data of each photovoltaic region based on historical photovoltaic power generation data, determining the power generation quality coefficient of each photovoltaic region based on the changes in power generation data; clustering each photovoltaic region based on the power generation quality coefficient, dividing each photovoltaic region into corresponding clustering partitions; obtaining the cluster center value of each clustering partition, and establishing a power generation characteristic coefficient set based on the cluster center value of each clustering partition.
[0019] In some embodiments of this application, the step of clustering each photovoltaic region based on the power generation quality coefficient includes: establishing a sample dataset based on the power generation quality coefficient corresponding to each photovoltaic region, and randomly selecting k initial cluster centers from the sample dataset; calculating the Manhattan distance from the power generation quality coefficient in the sample dataset to the initial cluster centers, and dividing the photovoltaic regions corresponding to the power generation quality coefficients into corresponding cluster partitions based on the Manhattan distance from the power generation quality coefficients in the sample dataset to the initial cluster centers; calculating the average value of the power generation quality coefficients in each cluster partition, and recalculating the cluster centers based on the average value of the power generation quality coefficients in each cluster partition; repeating the clustering steps iteratively until the cluster centers no longer change or the number of iterations reaches the preset maximum number of iterations, thereby obtaining the clustering results of the photovoltaic regions.
[0020] S102, Obtain current environmental data, and determine the predicted power generation load based on the current environmental data and the power generation characteristic coefficient set; In some embodiments of this application, determining the predicted power generation load based on current environmental data and a set of power generation characteristic coefficients includes: determining historical environmental data, a set of power generation characteristic coefficients, and the corresponding power generation load based on historical photovoltaic power generation data; establishing a training sample set based on the historical environmental data, the set of power generation characteristic coefficients, and the corresponding power generation load; establishing an initial load prediction model based on the training sample set and training the initial load prediction model to obtain a trained load prediction model; and inputting the current environmental data and the set of power generation characteristic coefficients of the photovoltaic power station into the trained load prediction model to obtain the predicted power generation load of the current photovoltaic power station.
[0021] In some embodiments of this application, the environmental data of the photovoltaic power station specifically includes sunlight data and meteorological data.
[0022] S103, acquire images of the photovoltaic panel during the power generation process, and determine the shading state coefficient based on the images of the photovoltaic panel during the power generation process; In some embodiments of this application, determining the shadow state coefficient based on the photovoltaic panel image during the power generation process includes: performing grayscale processing on the photovoltaic panel image during the power generation process to obtain a photovoltaic grayscale image; performing image segmentation on the photovoltaic grayscale image to obtain several sub-photovoltaic grayscale images; calculating the average grayscale value of each sub-photovoltaic grayscale image; and determining the shadow state coefficient based on the change in the average grayscale value of each sub-photovoltaic grayscale image.
[0023] In some embodiments of this application, determining the shadow state coefficient based on the average gray value change of each sub-photovoltaic grayscale image includes: plotting an average gray value change curve based on the average gray value change of each sub-photovoltaic grayscale image; performing curve fitting on the average gray value change curve to obtain an average gray value fitting curve; determining the time required for the average gray value to reach a value greater than a preset average shadow gray value based on the average gray value fitting curve; determining the sub-shadow state coefficient based on the time required for the average gray value to reach a value greater than the preset average shadow gray value; and calculating the average sub-shadow state coefficient of all sub-photovoltaic grayscale images to obtain the shadow state coefficient.
[0024] S104, determine the load correction factor based on the shading state factor, and adjust the predicted power generation load based on the load correction factor.
[0025] In some embodiments of this application, the determination of the load correction factor by the shading state factor includes: determining the load correction factor according to the load correction factor calculation formula, wherein the load correction factor calculation formula is specifically as follows: , in, This is the load correction factor. The preset standard load correction factor, To preset the standard shadow state coefficient, This is the shadow state coefficient. For the preset range coefficient, It is a natural exponential function.
[0026] In some embodiments of this application, adjusting the predicted power generation load according to the load correction factor includes: multiplying the load correction factor by the predicted power generation load to obtain the adjusted predicted power generation load.
[0027] Based on the same technological concept, such as Figure 2 As shown, the present invention also provides a distributed photovoltaic power generation load forecasting system, comprising: The first module is used to acquire historical photovoltaic power generation data and determine the power generation characteristic coefficient set of the current photovoltaic power station based on the historical photovoltaic power generation data; the second module is used to acquire current environmental data and determine the predicted power generation load based on the current environmental data and the power generation characteristic coefficient set; the third module is used to acquire photovoltaic panel images during the power generation process and determine the shading state coefficient based on the photovoltaic panel images during the power generation process; the fourth module is used to determine the load correction coefficient based on the shading state coefficient and adjust the predicted power generation load based on the load correction coefficient.
[0028] By applying the above technical solutions, this invention acquires historical photovoltaic (PV) power generation data and determines the current PV power generation characteristic coefficient set based on this data; acquires current environmental data and determines the predicted power generation load based on this data and the PV power generation characteristic coefficient set; acquires PV panel images during the power generation process and determines the shading state coefficient based on these images; determines the load correction coefficient based on the shading state coefficient; and adjusts the predicted power generation load based on the load correction coefficient. This invention, through intelligent evaluation and analysis of relevant data from distributed PV power stations, enables more accurate PV power generation load prediction.
[0029] Through the above description of the embodiments, those skilled in the art can clearly understand that the present invention can be implemented in hardware or by means of software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solution of the present invention can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) and includes several instructions to cause a computer device (such as a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.
[0030] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.
Claims
1. A method for predicting distributed photovoltaic power generation load, characterized in that, include: Obtain historical photovoltaic power generation data, and determine the current photovoltaic power generation characteristic coefficient set based on the historical photovoltaic power generation data; Acquire current environmental data and determine the predicted power generation load based on the current environmental data and the power generation characteristic coefficient set; Acquire images of photovoltaic panels during the power generation process, and determine the shading state coefficient based on the images of photovoltaic panels during the power generation process; The load correction factor is determined based on the shadow state factor, and the predicted power generation load is adjusted based on the load correction factor.
2. The distributed photovoltaic power generation load forecasting method according to claim 1, characterized in that, The process of determining the power generation characteristic coefficient set of the current photovoltaic power station based on historical photovoltaic power generation data includes: Based on historical photovoltaic power generation data, determine the changes in power generation data for each photovoltaic region, and based on the changes in power generation data, determine the power generation quality coefficient for each photovoltaic region; Based on the power generation quality coefficient, each photovoltaic region is clustered and assigned to a corresponding cluster partition. Obtain the cluster center values of each cluster partition, and establish a set of power generation characteristic coefficients based on the cluster center values of each cluster partition.
3. The distributed photovoltaic power generation load forecasting method according to claim 2, characterized in that, The clustering of photovoltaic regions based on power generation quality coefficient includes: A sample dataset is established based on the power generation quality coefficient corresponding to each photovoltaic region, and k initial cluster centers are randomly selected from the sample dataset. Calculate the Manhattan distance from the power generation quality coefficient in the sample dataset to the initial cluster center, and divide the photovoltaic region corresponding to the power generation quality coefficient into the corresponding cluster partition based on the Manhattan distance from the power generation quality coefficient to the initial cluster center in the sample dataset; Calculate the average power generation quality coefficient within each cluster partition, and recalculate the cluster centers based on the average power generation quality coefficient within each cluster partition; Repeat the clustering steps until the cluster centers no longer change or the number of iterations reaches the preset maximum number of iterations to obtain the clustering results of the photovoltaic region.
4. The distributed photovoltaic power generation load forecasting method according to claim 1, characterized in that, The process of determining the predicted power generation load based on current environmental data and a set of power generation characteristic coefficients includes: Based on historical photovoltaic power generation data, determine historical environmental data, power generation characteristic coefficient set, and corresponding power generation load; and establish a training sample set based on historical environmental data, power generation characteristic coefficient set, and corresponding power generation load. An initial load prediction model is established based on the training sample set, and the initial load prediction model is trained to obtain a trained load prediction model. By inputting the current environmental data and power generation characteristic coefficient set of the photovoltaic power station into the trained load prediction model, the predicted power generation load of the current photovoltaic power station can be obtained.
5. The distributed photovoltaic power generation load forecasting method according to claim 4, characterized in that, The environmental data for the photovoltaic power station specifically includes solar radiation data and meteorological data.
6. The distributed photovoltaic power generation load forecasting method according to claim 1, characterized in that, The determination of the shading state coefficient based on the photovoltaic panel image during the power generation process includes: The photovoltaic panel images during the power generation process are converted to grayscale to obtain photovoltaic grayscale images; The photovoltaic grayscale image is segmented to obtain several sub-photovoltaic grayscale images; Calculate the average gray value of each sub-photovoltaic grayscale image, and determine the shadow state coefficient based on the change of the average gray value of each sub-photovoltaic grayscale image.
7. The distributed photovoltaic power generation load forecasting method according to claim 6, characterized in that, The step of determining the shadow state coefficient based on the average grayscale value change of each sub-photovoltaic grayscale image includes: Based on the changes in the average gray value of each sub-photovoltaic grayscale image, an average gray value change curve is plotted, and curve fitting is performed on the average gray value change curve to obtain the average gray value fitting curve. The time required for the average gray value to reach a value greater than the preset average gray value of the shadow is determined based on the fitting curve of the average gray value, and the sub-shadow state coefficient is determined based on the time required for the average gray value to reach a value greater than the preset average gray value of the shadow. The shadow state coefficient is obtained by calculating the average value of the sub-shadow state coefficients of all sub-photovoltaic grayscale images.
8. The distributed photovoltaic power generation load forecasting method according to claim 1, characterized in that, The shadowing state factor determines the load correction factor, including: The load correction factor is determined according to the formula for calculating the load correction factor, which is as follows: ， in, This is the load correction factor. The preset standard load correction factor, To preset the standard shadow state coefficient, This is the shadow state coefficient. For the preset range coefficient, It is a natural exponential function.
9. The distributed photovoltaic power generation load forecasting method according to claim 8, characterized in that, The adjustment of the predicted power generation load based on the load correction factor includes: The adjusted forecasted power generation load is obtained by multiplying the load correction factor by the forecasted power generation load.
10. A distributed photovoltaic power generation load forecasting system, characterized in that, include: The first module is used to acquire historical photovoltaic power generation data and determine the current photovoltaic power station's power generation characteristic coefficient set based on the historical photovoltaic power generation data; The second module is used to acquire current environmental data and determine the predicted power generation load based on the current environmental data and the power generation characteristic coefficient set. The third module is used to acquire images of photovoltaic panels during the power generation process and determine the shading state coefficient based on the images of photovoltaic panels during the power generation process. The fourth module is used to determine the load correction factor based on the shading state factor, and to adjust the predicted power generation load based on the load correction factor.