Distributed photovoltaic anomaly monitoring method and system
By segmenting and determining the power value of distributed photovoltaic power stations, combining the generation of abnormal parameters and the calculation of alarm coefficients, the problem of low abnormal monitoring efficiency of distributed photovoltaic power stations is solved, and accurate power data prediction and real-time abnormal monitoring of distributed photovoltaic power stations are realized.
Patent Information
- Application Number
- CN202510213256.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-02-25
- Publication Date
- 2025-06-10
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The abnormal monitoring efficiency of distributed photovoltaic power plants is low and it is difficult to monitor abnormal events in a timely manner.
By acquiring image data of a distributed photovoltaic power station, dividing it into multiple photovoltaic blocks, obtaining historical power generation data to determine predicted power values, drawing a power change curve to determine abnormal parameters, and generating alarm coefficients based on these parameters for abnormal alarms.
Accurate power data prediction and real-time abnormality monitoring of distributed photovoltaic power stations are realized, improving the abnormality monitoring efficiency.
Smart Images
Figure CN120128082A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of photovoltaic power generation, and more specifically, to a method and system for monitoring abnormalities in distributed photovoltaics. Background Art
[0002] Distributed photovoltaic power generation specifically refers to photovoltaic power generation facilities constructed near user sites, operating in a way that users consume electricity generated by themselves on-site, and the excess electricity is fed into the grid, and characterized by balancing and regulating in the distribution system. Distributed photovoltaic power generation follows the principles of adapting to local conditions, being clean and efficient, having a decentralized layout, and being utilized nearby, making full use of local solar energy resources to replace and reduce fossil energy consumption.
[0003] In recent years, with the continuous progress of photovoltaic technology, the penetration rate of distributed photovoltaics in the power system has been increasing year by year. However, due to the characteristics of volatility and uncertainty of photovoltaic power output, it will impact the stability of the power system. Moreover, due to the difficulty in predicting the power data of distributed photovoltaic power stations, abnormal events cannot be monitored in a timely manner, resulting in low efficiency of abnormal monitoring. Summary of the Invention
[0004] The present invention provides a method and system for monitoring abnormalities in distributed photovoltaics to solve the problem of low efficiency of abnormal monitoring in existing distributed photovoltaic power stations, including: Obtain image data of a distributed photovoltaic power station, and divide the distributed photovoltaic power station into multiple photovoltaic blocks according to the image data of the distributed photovoltaic power station by a preset grid; Obtain historical power generation data of a photovoltaic power station, and determine the predicted power value of a photovoltaic block according to the historical power generation data of the photovoltaic power station; Draw a curve of the change in the predicted power value according to the predicted power value of the photovoltaic block, and determine the abnormal parameter of the photovoltaic block according to the curve of the change in the predicted power value; Generate an alarm coefficient according to the abnormal parameter of the photovoltaic block, and perform abnormal alarm on the photovoltaic block according to the alarm coefficient.
[0005] Further, the determining the predicted power value of a photovoltaic block according to the historical power generation data of the photovoltaic power station includes: Obtain meteorological data of the positions of each photovoltaic block in the distributed photovoltaic power station, and set the photovoltaic block with the best meteorological data in each preset time period as the benchmark photovoltaic block for that preset time period; Statistical distance values and corresponding historical power change values of the remaining photovoltaic blocks and the benchmark photovoltaic block in each preset time period, calculate the correlation coefficient of the distance value and the corresponding historical power change value in each preset time period, and determine the power change coefficient of each photovoltaic block according to the correlation coefficient of the distance value and the corresponding historical power change value in each preset time period; Obtain historical power generation data of a photovoltaic power station, and determine the predicted power value of each photovoltaic block according to the historical power generation data of the photovoltaic power station and the power change coefficient.
[0006] Further, determining the predicted power values of each photovoltaic block according to the historical power generation data of the photovoltaic power station and the power change coefficient includes: Determining the power change value of the historical benchmark photovoltaic block, the distance values between the remaining photovoltaic blocks and the benchmark photovoltaic block, and the power change values of the corresponding remaining photovoltaic blocks according to the historical power generation data of the photovoltaic power station; Establishing a training sample set according to the power change value of the historical benchmark photovoltaic block, the distance values between the remaining photovoltaic blocks and the benchmark photovoltaic block, and the power change values of the corresponding remaining photovoltaic blocks; Establishing a power prediction model according to the training sample set and training the power prediction model to obtain a trained power prediction model; Obtaining the power change value of the current benchmark photovoltaic block and the distance values between the remaining photovoltaic blocks and the benchmark photovoltaic block, and inputting the power change value of the current benchmark photovoltaic block and the distance values between the remaining photovoltaic blocks and the benchmark photovoltaic block into the trained power prediction model to obtain the initial predicted power values of each photovoltaic block; Obtaining the power change coefficient of each photovoltaic block, and multiplying the power change coefficient by the initial predicted power value to obtain the final predicted power value.
[0007] Further, determining the abnormal parameters of the photovoltaic block according to the predicted power value change curve includes: Obtaining the real-time power change situation of each photovoltaic block, and drawing a real-time power change curve according to the real-time power change situation of the photovoltaic block; Obtaining a preset rolling time window, and dividing the real-time power change curve according to the preset rolling time window to obtain a plurality of sub-power change curves; Calculating the absolute value of the difference between the peak and the adjacent valley in the sub-power change curve, counting the number of peaks in the sub-power change curve, and multiplying the sum of the absolute values of the differences between the peak and the adjacent valley in the sub-power change curve by the number of peaks to obtain the curve form value of the determined sub-change curve; Determining the predicted power value of the photovoltaic block in the preset rolling time window according to the predicted power value change curve, calculating the difference between the real-time power change curve and the predicted power value within the same rolling time window, and determining the abnormal parameters of the photovoltaic block according to the difference between the real-time power change curve and the predicted power value within the same rolling time window and the corresponding curve form value.
[0008] Further, determining the abnormal parameters of the photovoltaic block according to the difference between the real-time power change curve and the predicted power value within the same rolling time window and the corresponding curve form value includes: Calculating the abnormal parameters of the photovoltaic block according to the abnormal parameter calculation formula, and the abnormal parameters are specifically
[0009] Wherein, is the abnormal parameter of the photovoltaic block in the preset rolling time window, is the difference between the real-time power change curve and the predicted power value within the same rolling time window, is the curve form value of the sub-change curve, is the preset standard form value, is the preset range coefficient.
[0010] Further, generating an alarm coefficient according to the abnormal parameter of the photovoltaic block includes: Obtaining the time series of abnormal parameters of the photovoltaic block, and clustering the abnormal parameters in the time series of abnormal parameters; Dividing a normal data set and an abnormal data set according to the clustering result of the abnormal parameters, and determining the alarm coefficient of the photovoltaic block according to the normal data set and the abnormal data set.
[0011] Further, clustering the abnormal parameters in the time series of abnormal parameters includes: Obtaining each abnormal parameter in the time series of abnormal parameters, establishing a sample data set according to each abnormal parameter in the time series of abnormal parameters, and randomly selecting k initial clustering centers of the sample data set; Calculating the Euclidean distance from the abnormal parameters in the sample data set to the initial clustering centers, and dividing each abnormal parameter into the corresponding clustering cluster according to the Euclidean distance from the abnormal parameters in the sample data set to the initial clustering centers; Calculating the average value of the abnormal parameters in each clustering cluster, and recalculating the clustering centers according to the average value of the abnormal parameters in each clustering cluster; Repeating and iterating the above steps until the clustering centers no longer change or the number of iterations reaches the preset maximum number of iterations, to obtain the clustering result of the abnormal parameters.
[0012] Further, dividing a normal data set and an abnormal data set according to the clustering result of the abnormal parameters, and determining the alarm coefficient of the photovoltaic block according to the normal data set and the abnormal data set includes: Obtaining the preset standard abnormal parameters, and calculating the difference between the clustering centers of each abnormal parameter clustering cluster and the preset standard abnormal parameters; Dividing the clustering clusters with differences less than the first preset threshold into the normal data set, and dividing the clustering clusters with differences greater than the first preset threshold into the abnormal data set; Respectively obtaining the average clustering center values of all clustering clusters in the normal data set and the abnormal data set, calculating the difference between the corresponding average clustering center values of the two data sets, and determining the alarm coefficient of the photovoltaic block according to the difference between the corresponding average clustering center values of the two data sets.
[0013] Further, performing abnormal alarm on the photovoltaic block according to the alarm coefficient includes: Obtain a preset standard alarm coefficient, and calculate the difference between the alarm coefficient of the photovoltaic block and the preset standard alarm coefficient; Determine whether the difference between the alarm coefficient of the photovoltaic block and the preset standard alarm coefficient is less than a second preset threshold. If the difference between the alarm coefficient of the photovoltaic block and the preset standard alarm coefficient is less than the second preset threshold, then perform an abnormal alarm for the photovoltaic block; If the difference between the alarm coefficient of the photovoltaic block and the preset standard alarm coefficient is greater than or equal to the second preset threshold, then do not perform an abnormal alarm for the photovoltaic block.
[0014] To achieve the above object, the present invention also provides a distributed photovoltaic anomaly monitoring system, including: A first module, configured to obtain image data of a distributed photovoltaic power station, and divide the distributed photovoltaic power station into a plurality of photovoltaic blocks according to the preset grid based on the image data of the distributed photovoltaic power station; A second module, configured to obtain historical power generation data of a photovoltaic power station, and determine a predicted power value of a photovoltaic block according to the historical power generation data of the photovoltaic power station; A third module, configured to draw a predicted power value change curve according to the predicted power value of the photovoltaic block, and determine an abnormal parameter of the photovoltaic block according to the predicted power value change curve; A fourth module, configured to generate an alarm coefficient according to the abnormal parameter of the photovoltaic block, and perform an abnormal alarm on the photovoltaic block according to the alarm coefficient.
[0015] The beneficial effects of the present invention are as follows: By applying the above technical solutions, the present invention can accurately predict the power data of each photovoltaic block by dividing the distributed photovoltaic power station, and at the same time monitor the abnormal photovoltaic blocks through the power data of each photovoltaic block and perform real-time alarms, which can effectively improve the efficiency of abnormal monitoring of the distributed photovoltaic power station. Description of the Drawings
[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without creative efforts.
[0017] Figure 1 Shows a schematic flowchart of a distributed photovoltaic anomaly monitoring method proposed by an embodiment of the present invention; Figure 2 Shows the overall structure diagram of a distributed photovoltaic anomaly monitoring system proposed by an embodiment of the present invention. Detailed Embodiments
[0018] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present application without creative efforts shall fall within the protection scope of the present application.
[0019] An embodiment of the present application provides a distributed photovoltaic anomaly monitoring method, as Figure 1 shown, including: S101, obtaining image data of a distributed photovoltaic power station, and dividing the distributed photovoltaic power station into multiple photovoltaic blocks according to the image data of the distributed photovoltaic power station by a preset grid; In this embodiment, the area value of the preset grid is set according to the scale of the distributed photovoltaic power station. The larger the scale, the larger the corresponding area value of the preset grid. The image data of the photovoltaic power station is evenly divided by the preset grid to obtain multiple photovoltaic blocks.
[0020] S102, obtaining historical power generation data of a photovoltaic power station, and determining the predicted power value of a photovoltaic block according to the historical power generation data of the photovoltaic power station; In some embodiments of the present application, the determining the predicted power value of a photovoltaic block according to the historical power generation data of the photovoltaic power station includes: obtaining meteorological data at the positions of each photovoltaic block in the distributed photovoltaic power station, and setting the photovoltaic block with the best meteorological data in each preset time period as the benchmark photovoltaic block of the preset time period; statistically calculating the distance values and corresponding historical power change values between the remaining photovoltaic blocks and the benchmark photovoltaic block in each preset time period, calculating the correlation coefficients between the distance values and the corresponding historical power change values in each preset time period, and determining the power change coefficients of each photovoltaic block according to the correlation coefficients between the distance values and the corresponding historical power change values in each preset time period; obtaining the historical power generation data of the photovoltaic power station, and determining the predicted power value of each photovoltaic block according to the historical power generation data of the photovoltaic power station and the power change coefficients.
[0021] In this embodiment, the photovoltaic block with the best meteorological conditions is selected as the benchmark photovoltaic block by detecting meteorological data such as light and temperature around each photovoltaic block. The power change coefficient is obtained through the correlation coefficients between the distance values and the corresponding historical power change values between the remaining photovoltaic blocks and the benchmark photovoltaic block. The range of the power change coefficient is set to [0.5, 2]. The larger the correlation coefficient, the higher the corresponding power change coefficient. The power generation data of the photovoltaic power station is specifically the power generation amount of the photovoltaic power station. The power values of each photovoltaic block are predicted through the power change coefficients and the historical power generation data of the photovoltaic power station.
[0022] In some embodiments of the present application, determining the predicted power values of each photovoltaic block according to the historical power generation data of the photovoltaic power station and the power change coefficient includes: determining the power change value of the historical benchmark photovoltaic block according to the historical power generation data of the photovoltaic power station, the distance values between the remaining photovoltaic blocks and the benchmark photovoltaic block, and the corresponding power change values of the remaining photovoltaic blocks; establishing a training sample set according to the power change value of the historical benchmark photovoltaic block, the distance values between the remaining photovoltaic blocks and the benchmark photovoltaic block, and the corresponding power change values of the remaining photovoltaic blocks; establishing a power prediction model according to the training sample set and training the power prediction model to obtain a trained power prediction model; obtaining the power change value of the current benchmark photovoltaic block and the distance values between the remaining photovoltaic blocks and the benchmark photovoltaic block, and inputting the power change value of the current benchmark photovoltaic block and the distance values between the remaining photovoltaic blocks and the benchmark photovoltaic block into the trained power prediction model to obtain the initial predicted power values of each photovoltaic block; obtaining the power change coefficient of each photovoltaic block, and multiplying the power change coefficient by the initial predicted power value to obtain the final predicted power value.
[0023] In this embodiment, the power of the current photovoltaic block is predicted by training the power prediction model based on the power change value of the benchmark photovoltaic block and the distance values between the remaining photovoltaic blocks and the benchmark photovoltaic block. At the same time, the predicted power value is corrected by the power change coefficient to more accurately obtain the final predicted power value.
[0024] S103. Draw a predicted power value change curve according to the predicted power values of the photovoltaic blocks, and determine the abnormal parameters of the photovoltaic blocks according to the predicted power value change curve; In some embodiments of the present application, determining the abnormal parameters of the photovoltaic blocks according to the predicted power value change curve includes: obtaining the real-time power change situation of each photovoltaic block, and drawing a real-time power change curve according to the real-time power change situation of the photovoltaic blocks; obtaining a preset rolling time window, and dividing the real-time power change curve according to the preset rolling time window to obtain a plurality of sub-power change curves; calculating the absolute value of the difference between the peak and the adjacent valley in the sub-power change curve, counting the number of peaks in the sub-power change curve, and multiplying the sum of the absolute values of the differences between the peak and the adjacent valley in the sub-power change curve by the number of peaks to obtain the curve form value of the determined sub-change curve; determining the predicted power value of the photovoltaic block in the preset rolling time window according to the predicted power value change curve, calculating the difference between the real-time power change curve and the predicted power value within the same rolling time window, and determining the abnormal parameters of the photovoltaic block according to the difference between the real-time power change curve and the predicted power value within the same rolling time window and the corresponding curve form value.
[0025] In some embodiments of the present application, determining the abnormal parameters of the photovoltaic blocks according to the difference between the real-time power change curve and the predicted power value within the same rolling time window and the corresponding curve form value includes: calculating the abnormal parameters of the photovoltaic blocks according to the abnormal parameter calculation formula, and the abnormal parameters are specifically
[0026] Among them, is the abnormal parameter of the photovoltaic block in the preset rolling time window, is the difference between the real-time power change curve and the predicted power value within the same rolling time window, is the curve form value of the sub-change curve, is the preset standard form value, is the preset range coefficient.
[0027] In this embodiment, by analyzing the curve form value of the predicted power value change curve and the deviation value between the real-time power and the predicted power, the abnormal parameter of the photovoltaic block is calculated. By combining the power fluctuation and power deviation of the photovoltaic block, the abnormal situation of the photovoltaic block can be accurately reflected.
[0028] S104. Generate an alarm coefficient according to the abnormal parameter of the photovoltaic block, and perform an abnormal alarm on the photovoltaic block according to the alarm coefficient.
[0029] In some embodiments of the present application, the generating an alarm coefficient according to the abnormal parameter of the photovoltaic block includes: obtaining the time series of the abnormal parameter of the photovoltaic block, and clustering the abnormal parameters in the time series of the abnormal parameter; dividing the normal data set and the abnormal data set according to the clustering result of the abnormal parameter, and determining the alarm coefficient of the photovoltaic block according to the normal data set and the abnormal data set.
[0030] In this embodiment, by collecting the abnormal parameters of the photovoltaic block in real time and counting the time series of the abnormal parameters, the normal data set and the abnormal data set of the time series of the abnormal parameters are divided, and then the alarm coefficient is determined.
[0031] In some embodiments of the present application, the clustering of the abnormal parameters in the time series of the abnormal parameter includes: obtaining each abnormal parameter in the time series of the abnormal parameter, establishing a sample data set according to each abnormal parameter in the time series of the abnormal parameter, and randomly selecting k initial clustering centers of the sample data set; calculating the Euclidean distance from the abnormal parameters in the sample data set to the initial clustering centers, and dividing each abnormal parameter into the corresponding clustering cluster according to the Euclidean distance from the abnormal parameters in the sample data set to the initial clustering centers; calculating the average value of the abnormal parameters in each clustering cluster, and recalculating the clustering centers according to the average value of the abnormal parameters in each clustering cluster; repeating the above steps iteratively until the clustering centers no longer change or the number of iterations reaches the preset maximum number of iterations, and obtaining the clustering result of the abnormal parameters.
[0032] In this embodiment, the k-means clustering algorithm is used to cluster the abnormal parameters in the time series of the abnormal parameter. The selection of the k value is determined by the sample quantity of the abnormal parameter. The more the quantity, the larger the corresponding k value.
[0033] In some embodiments of the present application, dividing the normal data set and the abnormal data set according to the abnormal parameter clustering result, and determining the alarm coefficient of the photovoltaic block according to the normal data set and the abnormal data set, includes: obtaining a preset standard abnormal parameter, and calculating the difference between the clustering center of each abnormal parameter clustering cluster and the preset standard abnormal parameter; dividing the clustering clusters with the difference less than the first preset threshold into the normal data set, and dividing the clustering clusters with the difference greater than the first preset threshold into the abnormal data set; respectively obtaining the average clustering center values of all clustering clusters in the normal data set and the abnormal data set, calculating the difference between the corresponding average clustering center values of the two data sets, and determining the alarm coefficient of the photovoltaic block according to the difference between the corresponding average clustering center values of the two data sets.
[0034] In this embodiment, the preset standard abnormal parameter is set by analyzing the historical experience data of the photovoltaic power station, the normal data set and the abnormal data set are screened out by the difference between the clustering center of each abnormal parameter clustering cluster and the preset standard abnormal parameter, and the difference between the corresponding average clustering center values of the two data sets is determined as the alarm coefficient of the photovoltaic block.
[0035] In some embodiments of the present application, performing abnormal alarm on the photovoltaic block according to the alarm coefficient, includes: obtaining a preset standard alarm coefficient, and calculating the difference between the alarm coefficient of the photovoltaic block and the preset standard alarm coefficient; determining whether the difference between the alarm coefficient of the photovoltaic block and the preset standard alarm coefficient is less than a second preset threshold, if the difference between the alarm coefficient of the photovoltaic block and the preset standard alarm coefficient is less than the second preset threshold, then performing abnormal alarm on the photovoltaic block; if the difference between the alarm coefficient of the photovoltaic block and the preset standard alarm coefficient is greater than or equal to the second preset threshold, then not performing abnormal alarm on the photovoltaic block.
[0036] Based on the same technical concept, as Figure 2 shown, the present invention also provides a distributed photovoltaic abnormal monitoring system, including: A first module, configured to obtain image data of a distributed photovoltaic power station, and divide the distributed photovoltaic power station into a plurality of photovoltaic blocks according to the preset grid according to the image data of the distributed photovoltaic power station; a second module, configured to obtain historical photovoltaic power station power generation data, and determine the predicted power value of the photovoltaic block according to the historical photovoltaic power station power generation data; a third module, configured to draw a predicted power value change curve according to the predicted power value of the photovoltaic block, and determine the abnormal parameter of the photovoltaic block according to the predicted power value change curve; a fourth module, configured to generate an alarm coefficient according to the abnormal parameter of the photovoltaic block, and perform abnormal alarm on the photovoltaic block according to the alarm coefficient.
[0037] By applying the above technical solution, the present invention obtains the image data of a distributed photovoltaic power station, divides the distributed photovoltaic power station into multiple photovoltaic blocks according to the preset grid based on the image data of the distributed photovoltaic power station; obtains the historical power generation data of the photovoltaic power station, and determines the predicted power value of the photovoltaic block according to the historical power generation data of the photovoltaic power station; draws a curve of the change of the predicted power value according to the predicted power value of the photovoltaic block, and determines the abnormal parameter of the photovoltaic block according to the curve of the change of the predicted power value; generates an alarm coefficient according to the abnormal parameter of the photovoltaic block, and performs abnormal alarm on the photovoltaic block according to the alarm coefficient. The present invention can accurately predict the power data of the distributed photovoltaic power station, and can monitor the photovoltaic blocks with anomalies in real time and perform abnormal alarm, improving the efficiency of abnormal monitoring of the distributed photovoltaic power station.
[0038] Through the description of the above embodiments, those skilled in the art can clearly understand that the present invention can be implemented by hardware or by means of software plus a necessary general hardware platform. Based on such an 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 (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.), including several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in various implementation scenarios of the present invention.
[0039] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the various embodiments of the present application.
Claims
1. A distributed photovoltaic abnormality monitoring method, characterized in that: The method comprises: Acquire image data of a distributed photovoltaic power station, and divide the distributed photovoltaic power station into a plurality of photovoltaic blocks according to a preset grid according to the image data of the distributed photovoltaic power station; Obtain historical photovoltaic power station power generation data, and determine the predicted power value of the photovoltaic block based on the historical photovoltaic power station power generation data; Draw a predicted power value change curve according to the predicted power value of the photovoltaic block, and determine abnormal parameters of the photovoltaic block according to the predicted power value change curve; An alarm coefficient is generated according to the abnormal parameters of the photovoltaic block, and an abnormal alarm is issued for the photovoltaic block according to the alarm coefficient.
2. The distributed photovoltaic abnormality monitoring method according to claim 1 is characterized in that: The step of determining the predicted power value of the photovoltaic block according to the historical photovoltaic power station power generation data includes: Obtaining meteorological data of each photovoltaic block location in the distributed photovoltaic power station, and setting the photovoltaic block with the best meteorological data in each preset period as the benchmark photovoltaic block in the preset period; Count the distance values of the remaining photovoltaic blocks and the benchmark photovoltaic block in each preset time period and the corresponding historical power change values, calculate the correlation coefficient between the distance value in each preset time period and the corresponding historical power change value, and determine the power change coefficient of each photovoltaic block according to the correlation coefficient between the distance value in each preset time period and the corresponding historical power change value; Obtain historical photovoltaic power station power generation data, and determine the predicted power value of each photovoltaic block based on the historical photovoltaic power station power generation data and power variation coefficient.
3. The distributed photovoltaic abnormality monitoring method according to claim 2 is characterized in that: Determining the predicted power value of each photovoltaic block according to the historical photovoltaic power station power generation data and the power variation coefficient includes: Determine the power change value of the historical benchmark photovoltaic block and the distance value between the remaining photovoltaic blocks and the benchmark photovoltaic block and the power change value of the corresponding remaining photovoltaic blocks according to the historical photovoltaic power station power generation data; A training sample set is established according to the power change value of the historical benchmark photovoltaic block, the distance values between the remaining photovoltaic blocks and the benchmark photovoltaic block, and the power change values corresponding to the remaining photovoltaic blocks; Establishing a power prediction model according to the training sample set and training the power prediction model to obtain a trained power prediction model; Obtain the power change value of the current benchmark photovoltaic block and the distance values between the remaining photovoltaic blocks and the benchmark photovoltaic block, input the power change value of the current benchmark photovoltaic block and the distance values between the remaining photovoltaic blocks and the benchmark photovoltaic block into the trained power prediction model, and obtain the initial predicted power value of each photovoltaic block; The power variation coefficient of each photovoltaic block is obtained, and the power variation coefficient is multiplied by the initial predicted power value to obtain the final predicted power value.
4. The distributed photovoltaic abnormality monitoring method according to claim 1 is characterized in that: The step of determining the abnormal parameters of the photovoltaic block according to the predicted power value change curve includes: Obtain the real-time power change of each photovoltaic block, and draw a real-time power change curve according to the real-time power change of the photovoltaic block; Obtain a preset rolling time window, and divide the real-time power change curve according to the preset rolling time window to obtain a plurality of sub-power change curves; Calculate the absolute value of the difference between the peak and the adjacent trough in the sub-power change curve, count the number of peaks in the sub-power change curve, multiply the sum of the absolute values of the difference between the peak and the adjacent trough in the sub-power change curve by the number of peaks to determine the curve shape value of the sub-change curve; The predicted power value of the photovoltaic block in the preset rolling time window is determined according to the predicted power value change curve, the difference between the real-time power change curve and the predicted power value in the same rolling time window is calculated, and the abnormal parameters of the photovoltaic block are determined according to the difference between the real-time power change curve and the predicted power value in the same rolling time window and the corresponding curve shape value.
5. The distributed photovoltaic abnormality monitoring method according to claim 4 is characterized in that: The method of determining the abnormal parameters of the photovoltaic block according to the difference between the real-time power change curve and the predicted power value in the same rolling time window and the corresponding curve shape value includes: The abnormal parameters of the photovoltaic block are calculated according to the abnormal parameter calculation formula. The abnormal parameters are specifically: in, is the abnormal parameter of the PV block in the preset rolling time window, is the difference between the real-time power change curve and the predicted power value in the same rolling time window. is the curve shape value of the sub-variation curve, is the preset standard shape value, is the preset range factor.
6. The distributed photovoltaic abnormality monitoring method according to claim 5 is characterized in that: The generating of the alarm coefficient according to the abnormal parameters of the photovoltaic block includes: Obtain the abnormal parameter time series of the photovoltaic block, and cluster the abnormal parameters in the abnormal parameter time series; A normal data set and an abnormal data set are divided according to the abnormal parameter clustering result, and the alarm coefficient of the photovoltaic block is determined according to the normal data set and the abnormal data set.
7. The distributed photovoltaic abnormality monitoring method according to claim 6 is characterized in that: The clustering of abnormal parameters in the abnormal parameter time series includes: Obtain each abnormal parameter in the abnormal parameter time series, establish a sample data set according to each abnormal parameter in the abnormal parameter time series, and randomly select k initial clustering centers of the sample data set; Calculate the Euclidean distance from the abnormal parameters in the sample data set to the initial cluster center, and divide each abnormal parameter into a corresponding cluster cluster according to the Euclidean distance from the abnormal parameters in the sample data set to the initial cluster center; Calculate the average value of the abnormal parameters in each cluster, and recalculate the cluster center according to the average value of the abnormal parameters in each cluster; Repeat the above steps until the cluster center no longer changes or the number of iterations reaches the preset maximum number of iterations, and obtain the clustering result of the abnormal parameters.
8. The distributed photovoltaic abnormality monitoring method according to claim 7 is characterized in that: The method of dividing a normal data set and an abnormal data set according to the abnormal parameter clustering result, and determining the alarm coefficient of the photovoltaic block according to the normal data set and the abnormal data set, includes: Obtaining preset standard abnormal parameters, and calculating the difference between the cluster center of each abnormal parameter cluster and the preset standard abnormal parameters; The clusters whose difference is less than a first preset threshold are classified into a normal data set, and the clusters whose difference is greater than the first preset threshold are classified into an abnormal data set; The average cluster center values of all clusters in the normal data set and the abnormal data set are obtained respectively, the difference between the average cluster center values corresponding to the two data sets is calculated, and the alarm coefficient of the photovoltaic block is determined according to the difference between the average cluster center values corresponding to the two data sets.
9. The distributed photovoltaic abnormality monitoring method according to claim 8, characterized in that: The abnormal alarm of the photovoltaic block according to the alarm coefficient includes: Obtaining a preset standard alarm coefficient, and calculating the difference between the alarm coefficient of the photovoltaic block and the preset standard alarm coefficient; Determine whether the difference between the alarm coefficient of the photovoltaic block and the preset standard alarm coefficient is less than a second preset threshold value, and if the difference between the alarm coefficient of the photovoltaic block and the preset standard alarm coefficient is less than the second preset threshold value, issue an abnormal alarm for the photovoltaic block; If the difference between the alarm coefficient of the photovoltaic block and the preset standard alarm coefficient is greater than or equal to the second preset threshold, no abnormal alarm of the photovoltaic block is issued.
10. A distributed photovoltaic abnormality monitoring system, characterized in that: include: The first module is used to obtain image data of the distributed photovoltaic power station, and divide the distributed photovoltaic power station into a plurality of photovoltaic blocks according to a preset grid according to the image data of the distributed photovoltaic power station; The second module is used to obtain historical photovoltaic power station power generation data and determine the predicted power value of the photovoltaic block based on the historical photovoltaic power station power generation data; The third module is used to draw a predicted power value change curve according to the predicted power value of the photovoltaic block, and determine the abnormal parameters of the photovoltaic block according to the predicted power value change curve; The fourth module is used to generate an alarm coefficient according to abnormal parameters of the photovoltaic block, and to issue an abnormal alarm to the photovoltaic block according to the alarm coefficient.