Photovoltaic station data interpolation method and system based on sunset law

Abstract

The invention relates to the technical field of photovoltaic power generation, and particularly discloses a photovoltaic station data interpolation method and system based on a sunset law, and the method comprises the steps: obtaining the power generation power data of a photovoltaic station, carrying out the anomaly detection of the power generation power data of the photovoltaic station, screening out an abnormal value in the power generation power data, and carrying out the calculation of the abnormal value. Removing abnormal values of the generated power data; acquiring historical operation data of the photovoltaic station, determining a sunset rule interval according to the historical operation data of the photovoltaic station, and interpolating the abnormal value of the generated power data according to the sunset rule interval; and establishing a power generation power prediction model based on the interpolated power generation power data, and predicting the power generation power of the photovoltaic station according to the power generation power prediction model. According to the method, the abnormal value of the generated power data is interpolated through the sunset law of the photovoltaic station, the influence of different sunset laws on the generated power data can be adapted, and the accuracy of the generated power data is effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of photovoltaic power generation, and more specifically, to a photovoltaic station data interpolation method and system based on sunset law. Background Art

[0002] As a renewable and clean energy, solar energy has good application prospects, and the photovoltaic industry that uses solar energy to generate electricity has developed rapidly. However, photovoltaic power generation is affected by a variety of meteorological factors, and has obvious volatility and intermittency, which significantly affects the stability and security of the traditional power grid connected to it. Therefore, it is particularly important to quantitatively identify the key meteorological factors and their characteristic quantities that affect photovoltaic power generation. This can provide key basic support for energy meteorological services such as site selection of photovoltaic power stations, solar energy resource assessment, daily operation and maintenance, and photovoltaic power generation power prediction.

[0003] Photovoltaic power stations are intermittent power sources. The output of photovoltaic power generation is affected by many factors and has randomness and volatility. The large-scale grid connection of photovoltaic power stations affects the safe and economic operation of the power system. Accurately predicting photovoltaic power generation is an effective measure to solve a series of problems caused by high-proportion photovoltaic power generation access to the power grid. The existing technology for predicting photovoltaic power generation is usually based on simple statistical analysis of historical data, and cannot be dynamically adjusted according to the specific conditions of the site, resulting in the accuracy of power generation prediction data being less than ideal. Summary of the invention

[0004] The present invention provides a photovoltaic station data interpolation method and system based on sunset law, which are used to solve the problem that the accuracy of power generation prediction data in the prior art is not ideal.

[0005] A first aspect of the present invention provides a photovoltaic station data interpolation method based on sunset law, comprising:

[0006] Obtain the power generation data of the photovoltaic station, perform anomaly detection on the power generation data of the photovoltaic station, filter out the abnormal values ​​in the power generation data, and remove the abnormal values ​​in the power generation data;

[0007] Obtain historical operation data of photovoltaic stations, determine the sunset regularity interval based on the historical operation data of photovoltaic stations, and interpolate abnormal values ​​of power generation data based on the sunset regularity interval;

[0008] A power generation prediction model is established based on the interpolated power generation data, and the power generation of the photovoltaic station is predicted according to the power generation prediction model.

[0009] Furthermore, the power generation data of the photovoltaic station is detected for abnormalities, including:

[0010] Determine the first quartile and the third quartile of the power generation data based on the box plot algorithm, and calculate the interquartile range of the power generation data based on the first quartile and the third quartile;

[0011] Obtain the ambient temperature of the photovoltaic station, and determine the box plot parameters according to the ambient temperature of the photovoltaic station;

[0012] The upper and lower boundaries of the box plot are calculated according to the first quartile, the third quartile, the interquartile range and the box plot parameters, and the data in the power generation data that is greater than the upper boundary or less than the lower boundary is determined as an outlier.

[0013] Further, the upper and lower boundaries of the box plot are calculated according to the first quartile, the third quartile, the interquartile range and the box plot parameters, including:

[0014] The upper and lower boundaries of the box plot are calculated according to the box plot calculation formula. The box plot calculation formula is as follows:

[0015] L up =Q3+k*IQR

[0016] L down =Q1-k*IQR

[0017] Among them, L up is the upper boundary, L down is the lower boundary, Q1 is the first quartile, Q3 is the third quartile, k is the box plot parameter, IQR is the interquartile range, IQR = Q3-Q1.

[0018] Furthermore, the box plot parameters are determined according to the ambient temperature of the photovoltaic station, including:

[0019] Obtaining preset initial box plot parameters, and calculating abnormal values ​​of power generation data according to the preset initial box plot parameters;

[0020] Counting the number of abnormal values, when the number of abnormal values ​​is greater than a first preset threshold, obtaining the current ambient temperature of the photovoltaic station, and determining whether the difference between the current ambient temperature and the preset allowable ambient temperature is greater than a second preset threshold;

[0021] If the difference between the current ambient temperature and the preset allowable ambient temperature is greater than a second preset threshold, setting the preset first box plot parameter as the box plot parameter of subsequent power generation data;

[0022] If the difference between the current ambient temperature and the preset allowable ambient temperature is less than or equal to the second preset threshold, determining whether the difference between the current ambient temperature and the preset allowable ambient temperature is greater than a third preset threshold;

[0023] If the difference between the current ambient temperature and the preset allowable ambient temperature is greater than a third preset threshold, setting the preset second box plot parameter as the box plot parameter of subsequent power generation data;

[0024] If the difference between the current ambient temperature and the preset allowable ambient temperature is less than or equal to the third preset threshold, the preset third box plot parameter is set as the box plot parameter of the subsequent power generation data.

[0025] Furthermore, the sunset regularity interval is determined based on the historical operation data of the photovoltaic station, including:

[0026] Determine the sunset period of the photovoltaic station based on the historical operation data of the photovoltaic station, and determine the change of solar radiation intensity during the sunset period based on the sunset period of the photovoltaic station;

[0027] Draw a solar radiation intensity variation curve according to the variation of solar radiation intensity, obtain a preset radiation cycle, and divide the solar radiation intensity variation curve into a plurality of sub-variation curves according to the preset radiation cycle;

[0028] Calculate the average value of the solar radiation intensity in the sub-change curve, draw the average value change curve according to the average value of the solar radiation intensity in the sub-change curve, calculate the absolute value of the slope of two adjacent points in the average value change curve, and determine the average value of the slope of the solar radiation intensity change curve according to the absolute value of the slope of two adjacent points in the average value change curve;

[0029] A sunset regularity data set was established according to the average and maximum slope values ​​of the solar radiation intensity variation curve. The historical operation data of photovoltaic stations were clustered according to the sunset regularity data set based on the k-means clustering algorithm, and the sunset regularity interval was determined according to the clustering results.

[0030] Furthermore, the historical operation data of the photovoltaic station is clustered according to the sunset law data set based on the k-means clustering algorithm, including:

[0031] Randomly select k initial cluster centers and calculate the Euclidean distance from the sunset regularity data in the sunset regularity data set to the initial cluster centers;

[0032] According to the Euclidean distance from the sunset regularity data to the initial cluster center, each sunset regularity data is divided into corresponding partitions, and the average value of all sunset regularity data in each partition is calculated;

[0033] Update the cluster center according to the average value of all sunset regularity data in each cluster, and repeat the iteration according to the new cluster center until the cluster center no longer changes, and obtain k final cluster centers;

[0034] The sunset rule data in the sunset rule data set are divided into k sunset rule intervals according to the Euclidean distance from the sunset rule data to the final cluster center.

[0035] Furthermore, the data abnormal values ​​of the photovoltaic station are interpolated according to the sunset regular interval, including:

[0036] Determine the sunset regularity interval of the power generation data of the current photovoltaic station according to the change of solar radiation of the current photovoltaic station, and draw the power generation data curve according to the power generation data;

[0037] Calculate the correlation between the current power generation data curve and each power generation data curve in the corresponding sunset regular interval, and select the power generation data curve with the highest correlation with the current power generation data curve in the sunset regular interval;

[0038] The data abnormal value of the current power generation data is interpolated according to the power generation data curve with the highest correlation with the current power generation data curve in the sunset regular interval.

[0039] Furthermore, a power generation prediction model is established based on the interpolated power generation data, and the power generation of the photovoltaic station is predicted according to the power generation prediction model, including:

[0040] Part of the interpolated power generation data is used as training sample data, and the other part is used as test sample data;

[0041] Training a regional convolutional neural network model according to the training sample data and the features of the training sample data, inputting the test sample data into the trained regional convolutional neural network model, and extracting the features of the test sample data;

[0042] A support vector regression model is trained according to the features of the training sample data and the test sample data, and the features of the test sample data are input into the trained support vector regression model to obtain the predicted power generation data of the photovoltaic station.

[0043] In order to achieve the above-mentioned purpose, the second aspect of the present invention further provides a photovoltaic station data interpolation system based on sunset law, comprising:

[0044] The first module is used to obtain the power generation data of the photovoltaic station, perform anomaly detection on the power generation data of the photovoltaic station, filter out abnormal values ​​in the power generation data, and remove abnormal values ​​in the power generation data;

[0045] The second module is used to obtain the historical operation data of the photovoltaic station, determine the sunset regularity interval according to the historical operation data of the photovoltaic station, and interpolate the abnormal values ​​of the power generation data according to the sunset regularity interval;

[0046] The third module is used to establish a power generation prediction model based on the interpolated power generation data, and predict the power generation of the photovoltaic station according to the power generation prediction model.

[0047] The beneficial effects of the present invention are:

[0048] By applying the above technical scheme, the present invention eliminates outliers in the power generation data through box plots, and interpolates the eliminated outliers based on the sunset law, which effectively improves the accuracy of the power generation data, ensures the integrity of the data, and improves the accuracy of subsequent power generation prediction values ​​for photovoltaic stations. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below 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 work.

[0050] Figure 1 The overall flow chart of a photovoltaic station data interpolation method based on sunset law proposed in an embodiment of the present invention is shown;

[0051] Figure 2 A schematic structural diagram of a photovoltaic station data interpolation method based on sunset law proposed in an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0052] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0053] The present application embodiment provides a photovoltaic station data interpolation method based on the sunset law, such as Figure 1 As shown, including:

[0054] S101, obtaining power generation data of a photovoltaic station, performing anomaly detection on the power generation data of the photovoltaic station, screening out abnormal values ​​in the power generation data, and removing the abnormal values ​​in the power generation data;

[0055] In some embodiments of the present application, anomaly detection is performed on the power generation data of a photovoltaic station, including: determining the first quartile and the third quartile in the power generation data based on a box plot algorithm, and calculating the interquartile range of the power generation data according to the first quartile and the third quartile; obtaining the ambient temperature of the photovoltaic station, and determining the box plot parameters according to the ambient temperature of the photovoltaic station; calculating the upper and lower boundaries of the box plot according to the first quartile, the third quartile, the interquartile range and the box plot parameters, and determining the data in the power generation data that is greater than the upper boundary or less than the lower boundary as an abnormal value.

[0056] It should be noted that the quartile of power generation data refers to the value of dividing the power generation data into four equal parts after arranging the data in ascending order. Specifically, the quartile includes three key points: the first quartile (Q1), the second quartile (Q2, i.e. the median) and the third quartile (Q3). These values ​​can help us understand the distribution of power generation data, such as the central tendency and dispersion of the data.

[0057] In some embodiments of the present application, the upper boundary and the lower boundary of the box plot are calculated according to the first quartile, the third quartile, the interquartile range and the box plot parameters, including: calculating the upper boundary and the lower boundary of the box plot according to the box plot calculation formula, and the box plot calculation formula is specifically,

[0058] L up =Q3+k*IQR

[0059] L down =Q1-k*IQR

[0060] Among them, L up is the upper boundary, L down is the lower boundary, Q1 is the first quartile, Q3 is the third quartile, k is the box plot parameter, IQR is the interquartile range, IQR = Q3-Q1.

[0061] In this embodiment, abnormal values ​​of the power generation data are detected and eliminated according to the box plot algorithm, and the box plot parameters are adjusted according to the ambient temperature of the photovoltaic station so that the power generation data can adapt to changes in the ambient temperature of the photovoltaic station, thereby ensuring the accuracy of the power generation data.

[0062] In some embodiments of the present application, box plot parameters are determined according to the ambient temperature of the photovoltaic station, including: obtaining preset initial box plot parameters, calculating abnormal values ​​of power generation data according to the preset initial box plot parameters; counting the number of abnormal values, when the number of abnormal values ​​is greater than a first preset threshold, obtaining the ambient temperature of the current photovoltaic station, and judging whether the difference between the current ambient temperature and the preset allowable ambient temperature is greater than a second preset threshold; if the difference between the current ambient temperature and the preset allowable ambient temperature is greater than the second preset threshold, setting the preset first box plot parameters as the box plot parameters of subsequent power generation data; if the difference between the current ambient temperature and the preset allowable ambient temperature is less than or equal to the second preset threshold, judging whether the difference between the current ambient temperature and the preset allowable ambient temperature is greater than a third preset threshold; if the difference between the current ambient temperature and the preset allowable ambient temperature is greater than the third preset threshold, setting the preset second box plot parameters as the box plot parameters of subsequent power generation data; if the difference between the current ambient temperature and the preset allowable ambient temperature is less than or equal to the third preset threshold, setting the preset third box plot parameters as the box plot parameters of subsequent power generation data.

[0063] In this embodiment, the preset initial box plot parameter is 1.5. If the number of outliers is less than or equal to the first preset threshold, the box plot parameter is not adjusted. The first box plot parameter is set to be greater than the second box plot parameter and greater than the third box plot parameter to be greater than 1.5. The higher the ambient temperature, the larger the corresponding box plot parameter, and the larger the range of the calculated upper and lower boundaries, so that the power generation data can better adapt to the ambient temperature.

[0064] S102, acquiring historical operation data of the photovoltaic station, determining a sunset regularity interval according to the historical operation data of the photovoltaic station, and interpolating abnormal values ​​of the power generation data according to the sunset regularity interval;

[0065] In some embodiments of the present application, a sunset regularity interval is determined based on historical operation data of a photovoltaic station, including: determining a sunset period of the photovoltaic station based on the historical operation data of the photovoltaic station, and determining changes in solar radiation intensity during the sunset period based on the sunset period of the photovoltaic station; drawing a solar radiation intensity change curve based on the changes in solar radiation intensity, obtaining a preset radiation period, and dividing the solar radiation intensity change curve into a number of sub-change curves based on the preset radiation period; calculating the average value of the solar radiation intensity in the sub-change curves, drawing an average value change curve based on the average value of the solar radiation intensity in the sub-change curves, calculating the absolute value of the slope of two adjacent points in the average value change curve, and determining the average value of the slope of the solar radiation intensity change curve based on the absolute value of the slope of two adjacent points in the average value change curve; establishing a sunset regularity data set based on the average value and the maximum value of the slope of the solar radiation intensity change curve, clustering the historical operation data of the photovoltaic station based on the sunset regularity data set based on the k-means clustering algorithm, and determining a sunset regularity interval based on the clustering result.

[0066] In this embodiment, the historical operation data of the photovoltaic station includes the sunset cycle of the photovoltaic station, the change in solar radiation intensity corresponding to the sunset cycle, and the power generation data curve corresponding to the change in solar radiation intensity. The slope average and maximum values ​​are calculated through the change in solar radiation intensity to form a sunset regularity data set. The historical operation data of the photovoltaic station are clustered according to the sunset regularity data set, and the sunset regularity interval of each sunset regularity data is set according to the clustering results.

[0067] In some embodiments of the present application, historical operating data of a photovoltaic station is clustered according to a sunset law data set based on a k-means clustering algorithm, including: randomly selecting k initial clustering centers, and calculating the Euclidean distance of the sunset law data in the sunset law data set to the initial clustering center; dividing each sunset law data into corresponding partitions according to the Euclidean distance from the sunset law data to the initial clustering center, and calculating the average value of all sunset law data in each partition; updating the clustering center according to the average value of all sunset law data in each cluster, and repeatedly iterating according to the new clustering center until the clustering center no longer changes, thereby obtaining k final clustering centers; and dividing the sunset law data in the sunset law data set into k sunset law intervals according to the Euclidean distance from the sunset law data to the final clustering center.

[0068] In this embodiment, the k value is set to 5, and the sunset regularity data is divided into k sunset regularity intervals according to the Euclidean distance from the sunset regularity data to the initial cluster center.

[0069] In some embodiments of the present application, data anomalies of a photovoltaic station are interpolated according to a sunset regular interval, including: determining a sunset regular interval of the power generation data of the current photovoltaic station according to the change in solar radiation of the current photovoltaic station, and drawing a power generation data curve according to the power generation data; calculating the correlation between the current power generation data curve and each power generation data curve in the corresponding sunset regular interval, and screening out the power generation data curve with the highest correlation with the current power generation data curve in the sunset regular interval; and interpolating data anomalies of the current power generation data according to the power generation data curve with the highest correlation with the current power generation data curve in the sunset regular interval.

[0070] In this embodiment, the average slope value and the maximum value of the current solar irradiation change are calculated by the solar irradiation change of the current photovoltaic station, and the Euclidean distance between the average slope value and the maximum value of the current solar irradiation change and the cluster center corresponding to each sunset regularity interval is calculated, and the sunset regularity interval with the closest Euclidean distance is selected as the sunset regularity interval of the power generation data of the current photovoltaic station;

[0071] In this embodiment, the curve equation of the current power generation data curve and the corresponding power generation data curves in the sunset regular interval can be calculated by a curve fitting method, and the correlation between the current power generation data curve and the corresponding power generation data curves in the sunset regular interval can be determined according to the constant part of the two curve equations;

[0072] In this embodiment, the data outliers of the current power generation data are interpolated by using the points corresponding to the data outliers removed from the power generation data in the power generation data curve with the highest correlation with the current power generation data curve in the sunset regular interval.

[0073] S103, establishing a power generation prediction model based on the interpolated power generation data, and predicting the power generation of the photovoltaic station according to the power generation prediction model.

[0074] In some embodiments of the present application, a power generation prediction model is established based on the interpolated power generation data, and the power generation of the photovoltaic station is predicted according to the power generation prediction model, including: using part of the interpolated power generation data as training sample data, and the other part as test sample data; training a regional convolutional neural network model according to the training sample data and the characteristics of the training sample data, inputting the test sample data into the trained regional convolutional neural network model, and extracting the characteristics of the test sample data; training a support vector regression model according to the training sample data and the characteristics of the test sample data, and inputting the characteristics of the test sample data into the trained support vector regression model to obtain the predicted power generation data of the photovoltaic station.

[0075] In this embodiment, the regional convolutional neural network model includes 2 convolutional layers, 2 maximum pooling layers, 1 activation layer and 1 normalization layer. The regional convolutional neural network model is used to extract features to train a support vector regression classifier; the support vector regression classifier is used to predict photovoltaic power generation to improve prediction accuracy.

[0076] Based on the same technical concept, such as Figure 2 As shown, the present invention also provides a photovoltaic station data interpolation system based on sunset law, which is characterized by comprising:

[0077] The first module is used to obtain the power generation data of the photovoltaic station, perform anomaly detection on the power generation data of the photovoltaic station, screen out the abnormal values ​​in the power generation data, and eliminate the abnormal values ​​of the power generation data; the second module is used to obtain the historical operation data of the photovoltaic station, determine the sunset regular interval according to the historical operation data of the photovoltaic station, and interpolate the abnormal values ​​of the power generation data according to the sunset regular interval; the third module is used to establish a power generation prediction model based on the interpolated power generation data, and predict the power generation of the photovoltaic station according to the power generation prediction model.

[0078] By applying the above technical solutions, the present invention obtains the power generation data of the photovoltaic station, performs anomaly detection on the power generation data of the photovoltaic station, screens out the abnormal values ​​in the power generation data, and removes the abnormal values ​​of the power generation data; obtains the historical operation data of the photovoltaic station, determines the sunset regularity interval according to the historical operation data of the photovoltaic station, and interpolates the abnormal values ​​of the power generation data according to the sunset regularity interval; establishes a power generation prediction model based on the interpolated power generation data, and predicts the power generation of the photovoltaic station according to the power generation prediction model. The present invention interpolates the abnormal values ​​of the power generation data according to the sunset regularity of the photovoltaic station, can adapt to the influence of different sunset regularities on the power generation data, and effectively improves the accuracy of the power generation data.

[0079] Through the description of the above implementation methods, 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 a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each implementation scenario of the present invention.

[0080] 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 it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, 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 embodiments of the present application.

Claims

1. A photovoltaic station data interpolation method based on sunset law, characterized in that: The method comprises: Obtain the power generation data of the photovoltaic station, perform anomaly detection on the power generation data of the photovoltaic station, filter out the abnormal values ​​in the power generation data, and remove the abnormal values ​​in the power generation data; Obtain historical operation data of photovoltaic stations, determine the sunset regularity interval based on the historical operation data of photovoltaic stations, and interpolate abnormal values ​​of power generation data based on the sunset regularity interval; A power generation prediction model is established based on the interpolated power generation data, and the power generation of the photovoltaic station is predicted according to the power generation prediction model.

2. The photovoltaic station data interpolation method based on sunset law according to claim 1 is characterized in that: Perform anomaly detection on the power generation data of photovoltaic stations, including: Determine the first quartile and the third quartile of the power generation data based on the box plot algorithm, and calculate the interquartile range of the power generation data based on the first quartile and the third quartile; Obtain the ambient temperature of the photovoltaic station, and determine the box plot parameters according to the ambient temperature of the photovoltaic station; The upper and lower boundaries of the box plot are calculated according to the first quartile, the third quartile, the interquartile range and the box plot parameters, and the data in the power generation data that is greater than the upper boundary or less than the lower boundary is determined as an outlier.

3. The photovoltaic station data interpolation method based on sunset law according to claim 2 is characterized in that: The upper and lower boundaries of the box plot are calculated based on the first quartile, the third quartile, the interquartile range and the box plot parameters, including: The upper and lower boundaries of the box plot are calculated according to the box plot calculation formula. The box plot calculation formula is as follows: L up =Q3+k*IQR L down =Q1-k*IQR Among them, L up is the upper boundary, L down is the lower boundary, Q1 is the first quartile, Q3 is the third quartile, k is the box plot parameter, IQR is the interquartile range, IQR = Q3-Q1.

4. The photovoltaic station data interpolation method based on sunset law according to claim 2 is characterized in that: The box plot parameters are determined according to the ambient temperature of the photovoltaic station, including: Obtaining preset initial box plot parameters, and calculating abnormal values ​​of power generation data according to the preset initial box plot parameters; Counting the number of abnormal values, when the number of abnormal values ​​is greater than a first preset threshold, obtaining the current ambient temperature of the photovoltaic station, and determining whether the difference between the current ambient temperature and the preset allowable ambient temperature is greater than a second preset threshold; If the difference between the current ambient temperature and the preset allowable ambient temperature is greater than a second preset threshold, setting the preset first box plot parameter as the box plot parameter of subsequent power generation data; If the difference between the current ambient temperature and the preset allowable ambient temperature is less than or equal to the second preset threshold, determining whether the difference between the current ambient temperature and the preset allowable ambient temperature is greater than a third preset threshold; If the difference between the current ambient temperature and the preset allowable ambient temperature is greater than a third preset threshold, setting the preset second box plot parameter as the box plot parameter of subsequent power generation data; If the difference between the current ambient temperature and the preset allowable ambient temperature is less than or equal to the third preset threshold, the preset third box plot parameter is set as the box plot parameter of the subsequent power generation data.

5. The photovoltaic station data interpolation method based on sunset law according to claim 1 is characterized in that: Determine the sunset regularity interval based on the historical operation data of the photovoltaic station, including: Determine the sunset period of the photovoltaic station based on the historical operation data of the photovoltaic station, and determine the change of solar radiation intensity during the sunset period based on the sunset period of the photovoltaic station; Draw a solar radiation intensity variation curve according to the variation of solar radiation intensity, obtain a preset radiation cycle, and divide the solar radiation intensity variation curve into a plurality of sub-variation curves according to the preset radiation cycle; Calculate the average value of the solar radiation intensity in the sub-change curve, draw the average value change curve according to the average value of the solar radiation intensity in the sub-change curve, calculate the absolute value of the slope of two adjacent points in the average value change curve, and determine the average value of the slope of the solar radiation intensity change curve according to the absolute value of the slope of two adjacent points in the average value change curve; A sunset regularity data set was established according to the average and maximum slope values ​​of the solar radiation intensity variation curve. The historical operation data of photovoltaic stations were clustered according to the sunset regularity data set based on the k-means clustering algorithm, and the sunset regularity interval was determined according to the clustering results.

6. The photovoltaic station data interpolation method based on sunset law according to claim 5 is characterized in that: Based on the K-means clustering algorithm, the historical operation data of the photovoltaic station is clustered according to the sunset law data set, including: Randomly select k initial cluster centers and calculate the Euclidean distance from the sunset regularity data in the sunset regularity data set to the initial cluster centers; According to the Euclidean distance from the sunset regularity data to the initial cluster center, each sunset regularity data is divided into corresponding partitions, and the average value of all sunset regularity data in each partition is calculated; Update the cluster center according to the average value of all sunset regularity data in each cluster, and repeat the iteration according to the new cluster center until the cluster center no longer changes, and obtain k final cluster centers; The sunset rule data in the sunset rule data set are divided into k sunset rule intervals according to the Euclidean distance from the sunset rule data to the final cluster center.

7. The photovoltaic station data interpolation method based on sunset law according to claim 6 is characterized in that: The data outliers of the photovoltaic station are interpolated according to the sunset regular interval, including: Determine the sunset regularity interval of the power generation data of the current photovoltaic station according to the change of solar radiation of the current photovoltaic station, and draw the power generation data curve according to the power generation data; Calculate the correlation between the current power generation data curve and each power generation data curve in the corresponding sunset regular interval, and select the power generation data curve with the highest correlation with the current power generation data curve in the sunset regular interval; The data abnormal value of the current power generation data is interpolated according to the power generation data curve with the highest correlation with the current power generation data curve in the sunset regular interval.

8. The photovoltaic station data interpolation method based on sunset law according to claim 1 is characterized in that: A power generation prediction model is established based on the interpolated power generation data, and the power generation of the photovoltaic station is predicted according to the power generation prediction model, including: Part of the interpolated power generation data is used as training sample data, and the other part is used as test sample data; Training a regional convolutional neural network model according to the training sample data and the features of the training sample data, inputting the test sample data into the trained regional convolutional neural network model, and extracting the features of the test sample data; A support vector regression model is trained according to the features of the training sample data and the test sample data, and the features of the test sample data are input into the trained support vector regression model to obtain the predicted power generation data of the photovoltaic station.

9. A photovoltaic station data interpolation system based on sunset law, characterized in that: include: The first module is used to obtain the power generation data of the photovoltaic station, perform anomaly detection on the power generation data of the photovoltaic station, filter out abnormal values ​​in the power generation data, and remove abnormal values ​​in the power generation data; The second module is used to obtain the historical operation data of the photovoltaic station, determine the sunset regularity interval according to the historical operation data of the photovoltaic station, and interpolate the abnormal values ​​of the power generation data according to the sunset regularity interval; The third module is used to establish a power generation prediction model based on the interpolated power generation data, and predict the power generation of the photovoltaic station according to the power generation prediction model.

Images