A Probabilistic Prediction Method and System for Photovoltaic Power Based on Copula Functions

Abstract

This invention discloses a photovoltaic power probability prediction method and system based on the copula function. The method includes: acquiring historical photovoltaic power data and historical meteorological data of centralized and distributed photovoltaic power plants; obtaining different weather types through clustering; obtaining the cumulative distribution of photovoltaic power based on the historical photovoltaic power data under different weather types, and constructing optimal copula function models that quantify and dynamically represent the spatial correlation of centralized and distributed photovoltaic power respectively; obtaining the predicted value of distributed photovoltaic power points based on the acquired photovoltaic power data of centralized photovoltaic power plants and the corresponding optimal copula function models; constructing a conditional probability model based on the quantile regression method, and obtaining the conditional probability prediction value corresponding to the predicted value of distributed photovoltaic power points through the conditional probability model, thereby achieving more accurate distributed photovoltaic power prediction and power probability prediction, and ensuring the reliability of power system operation.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of photovoltaic power generation, and particularly relates to a photovoltaic power probability prediction method and system based on a copula function. BACKGROUND

[0002] The statements in this section merely provide background information related to the present disclosure and do not necessarily constitute the prior art.

[0003] With the acceleration of industrialization and electrification, the demand for energy, especially for electricity, is increasing rapidly. Distributed photovoltaic has the advantages of rich developable resources, small development and construction difficulty, and significant energy-saving and environmental protection benefits, and is one of the important ways of photovoltaic development and utilization. However, due to the instability of weather conditions, photovoltaic power generation has strong intermittency and randomness, which poses challenges to the planning and operation of existing power systems. Therefore, the accuracy of photovoltaic power prediction is an important factor affecting the access of photovoltaic to the power system.

[0004] The inventors have found that existing photovoltaic power prediction methods include physical model-based methods and statistical model-based methods. The physical model-based method is complex, so the statistical model-based method is usually used for photovoltaic power prediction. Statistical models include neural network models, copula models, etc. Due to the difficulty of obtaining distributed photovoltaic data and the dispersion of installation, the copula model is usually used to better describe the uncertainty of photovoltaic power and meteorological variables, and the amount of data required for modeling is relatively small. However, the traditional copula function has limitations and cannot well fit the power data, resulting in low accuracy of photovoltaic power prediction.

[0005] In addition, traditional deterministic prediction cannot effectively describe the uncertainty of photovoltaic power prediction, and it is difficult to meet the needs of power grid dispatching decision and risk assessment. Probability prediction can provide probability information of future photovoltaic output and has more important engineering significance. Compared with point prediction of distributed photovoltaic power, photovoltaic power probability prediction can reflect more prediction information and has better practical reference value.

[0006] At present, photovoltaic power probability prediction methods can be divided into parametric methods and non-parametric methods. The parametric method assumes that the photovoltaic power follows a certain probability distribution, which is more subjective, and the prediction result often deviates greatly from the actual probability distribution; the non-parametric method does not depend on any prior knowledge and can better reflect the distribution characteristics of the sample itself, which makes the model have better generalization ability. However, the non-parametric method is usually based on quantile regression and combined with machine learning, which often has the problem of quantile crossing. SUMMARY

[0007] To solve the above problems of the prior art, the application provides a photovoltaic power probability prediction method and system based on a copula function, weather types are determined through cluster analysis, a Frank Copula function model and a vine copula function model are constructed, according to the different weather types, the optimal function model suitable for the corresponding weather type is selected to perform distributed photovoltaic power prediction, and based on the quantile regression idea, a photovoltaic power probability prediction model is constructed to avoid the quantile crossing problem, realize distributed photovoltaic power probability prediction, improve the accuracy of power prediction and the reliability of power system operation, reduce the power consumption cost, reduce the energy consumption, save energy and reduce emissions, and improve the economic benefit.

[0008] In a first aspect, the present disclosure provides a photovoltaic power probability prediction method based on a copula function.

[0009] A photovoltaic power probability prediction method based on a copula function comprises the following steps.

[0010] Obtaining historical photovoltaic power data and historical meteorological data of centralized and distributed photovoltaic power stations;

[0011] Pretreating the historical photovoltaic power data and the historical meteorological data, and obtaining different weather types through clustering;

[0012] According to the historical photovoltaic power data under different weather types, obtaining the cumulative distribution of photovoltaic power, and respectively constructing optimal copula function models quantitatively and dynamically representing the spatial correlation of centralized photovoltaic power and distributed photovoltaic power under different weather types;

[0013] According to the obtained photovoltaic power data of the centralized photovoltaic power station, combining the optimal copula function model under the corresponding weather type, and obtaining a distributed photovoltaic power point prediction value;

[0014] Based on the quantile regression method, a conditional probability model is constructed, and a conditional probability prediction value corresponding to the distributed photovoltaic power point prediction value is obtained through the conditional probability model.

[0015] In a second aspect, the present disclosure provides a photovoltaic power probability prediction system based on a copula function.

[0016] A photovoltaic power probability prediction system based on a copula function comprises the following steps.

[0017] A data acquisition module is configured to obtain historical photovoltaic power data and historical meteorological data of centralized and distributed photovoltaic power stations;

[0018] A data processing module is configured to pretreat the historical photovoltaic power data and the historical meteorological data, and obtain different weather types through clustering;

[0019] An optimal copula function model construction module is configured to obtain cumulative distribution of photovoltaic power according to historical photovoltaic power data under different weather types, and to construct optimal copula function models for quantitatively and dynamically representing spatial correlation of centralized photovoltaic power and distributed photovoltaic power under different weather types, respectively.

[0020] A photovoltaic power point prediction module is configured to obtain a distributed photovoltaic power point prediction value according to the photovoltaic power data of the centralized photovoltaic power station and the optimal copula function model under the corresponding weather type.

[0021] A photovoltaic power probability prediction module is configured to construct a conditional probability model based on a quantile regression method, and to obtain a conditional probability prediction value corresponding to the distributed photovoltaic power point prediction value through the conditional probability model.

[0022] In a third aspect, the present disclosure further provides an electronic device, which comprises a memory and a processor, and computer instructions stored in the memory and running on the processor, and when the computer instructions are run by the processor, the steps of the method of the first aspect are completed.

[0023] In a fourth aspect, the present disclosure further provides a computer readable storage medium for storing computer instructions, and when the computer instructions are executed by a processor, the steps of the method of the first aspect are completed.

[0024] The above one or more technical solutions have the following beneficial effects:

[0025] 1. The present application provides a photovoltaic power probability prediction method and system based on copula function, which determines the weather type through cluster analysis, constructs Frank Copula function model and Vine Copula function model, selects the optimal function model suitable for the corresponding weather type for distributed photovoltaic power prediction according to the difference of weather types, solves the problem of insufficient single Copula fitting, and improves the accuracy of power prediction.

[0026] 2. The present application is based on the idea of quantile regression, and gives a quantile regression algorithm based on copula, and constructs a photovoltaic power probability prediction model. Since there is an analytical expression, the problem of quantile crossing in existing non-parametric methods can be overcome, distributed photovoltaic power probability prediction is realized, more accurate reference data is provided for power system operation and control, the reliability of power system operation is ensured, the cost of electricity is reduced, energy consumption is reduced, energy saving and emission reduction are realized, and economic benefits are improved. BRIEF DESCRIPTION OF DRAWINGS

[0027] The drawings accompanying the specification of the present application form a part thereof, serve to provide further understanding of the present application, and together with the exemplary embodiments of the present application and their description, serve to explain the present application, and do not constitute improper limitations on the present application.

[0028] Figure 1 Flow chart of the copula function-based photovoltaic power probability prediction method of embodiment one of the present application;

[0029] Figure 2 Weather clustering result chart of K-means clustering in embodiment one of the present application;

[0030] Figure 3 (a) is a frequency histogram of centralized and distributed photovoltaic output under cloudy weather type in embodiment one of the present application;

[0031] Figure 3 (b) is a frequency histogram of centralized and distributed photovoltaic output under sunny weather type in embodiment one of the present application;

[0032] Figure 3 (c) is a frequency histogram of centralized and distributed photovoltaic output under overcast weather type in embodiment one of the present application;

[0033] Figure 4 Structure chart of the Vine Copula function model in embodiment one of the present application;

[0034] Figure 5 (a) is point prediction and actual value of photovoltaic output under cloudy weather type in embodiment one of the present application;

[0035] Figure 5 (b) is point prediction and actual value of photovoltaic output under sunny weather type in embodiment one of the present application;

[0036] Figure 5 (c) is point prediction and actual value of photovoltaic output under overcast weather type in embodiment one of the present application;

[0037] Figure 6 (a) is point prediction, actual value and confidence interval of probability prediction of photovoltaic output under cloudy weather type in embodiment one of the present application;

[0038] Figure 6 (b) is point prediction, actual value and confidence interval of probability prediction of photovoltaic output under sunny weather type in embodiment one of the present application;

[0039] Figure 6 (c) is point prediction, actual value and confidence interval of probability prediction of photovoltaic output under overcast weather type in embodiment one of the present application. DETAILED DESCRIPTION

[0040] It should be noted that the following detailed description is exemplary in nature and is intended to provide further description of the application. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0041] It is to be understood that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of example embodiments in accordance with the present application. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "comprises" and / or "comprising," when used in this specification, specify the presence of stated features, steps, operations, elements, components, and / or groups thereof, but do not preclude the presence or addition of one or more other features, steps, operations, elements, components, and / or groups thereof.

[0042] To solve the problems in the background art, the application provides a photovoltaic power probability prediction method and system based on a copula function. First, weather types are classified by a clustering method. Then, a Frank Copula function model and a vine copula function model are constructed according to the characteristics of each weather attribute, to quantitatively represent the power space correlation of centralized photovoltaic and distributed photovoltaic. Finally, the real value of centralized photovoltaic power at a future time is used to obtain a distributed photovoltaic prediction value, and conditional prediction probability prediction of the distributed photovoltaic power is realized. The method can better reflect the space-time correlation of distributed photovoltaic output by adding k-means clustering and vine copula functions, and improve the prediction accuracy to a certain extent. The following will be described with specific examples.

[0043] Embodiment one

[0044] The embodiment provides a photovoltaic power probability prediction method based on a copula function, as shown in the following formula (1), which comprises the following steps. Figure 1

[0045] Step S1, obtaining historical photovoltaic power data and historical meteorological data of centralized and distributed photovoltaic power stations;

[0046] Step S2, preprocessing the historical photovoltaic power data and the historical meteorological data, and obtaining different weather types by clustering;

[0047] ​Step S3, according to the historical photovoltaic power data under different weather types, the cumulative distribution of photovoltaic power is obtained, and the optimal copula function model for quantifying the dynamic representation of the spatial correlation of centralized photovoltaic and distributed photovoltaic power under different weather types is constructed; wherein the constructed copula function model includes Frank Copula correlation function model and Vine Copula correlation function model, wherein the Vine Copula correlation function model is a combination of multiple binary Copula correlation function models;

[0048] Step S4, according to the obtained photovoltaic power data of the centralized photovoltaic power station, the optimal copula function model under the corresponding weather type is combined to obtain the distributed photovoltaic power point prediction value;

[0049] Step S5, based on the quantile regression method, a conditional probability model is constructed, and the conditional probability prediction value corresponding to the distributed photovoltaic power point prediction value is obtained through the conditional probability model.

[0050] Wherein, the Copula function describes the correlation between variables, which is actually a kind of function connecting the joint distribution function with its respective marginal distribution function, also known as connection function.

[0051] In this embodiment, the clustering algorithm and the weather classification vine copula function model are innovatively applied to the distributed photovoltaic power prediction, the weather clustering under three-dimensional scale is carried out for the historical meteorological data, the different weather types are determined based on the clustering results, and the optimal copula function model under different weather types is constructed, including Frank Copula correlation function model and Vine Copula correlation function model, which improves the accuracy of the prediction model and improves the accuracy of predicting distributed photovoltaic power through centralized photovoltaic data. Through the centralized photovoltaic prediction of distributed photovoltaic, the problem of unable to collect distributed photovoltaic data is solved, and more accurate reference data is provided for power system operation and regulation.

[0052] In the above step S1, the historical photovoltaic power data of centralized and distributed photovoltaic power stations is obtained, and the historical meteorological data corresponding to the historical photovoltaic power data is obtained. In this embodiment, the historical photovoltaic power data of centralized and distributed photovoltaic power stations in January, February and December can be collected and obtained, and the sampling interval is 10 minutes. The meteorological data includes data of various meteorological elements, such as temperature, wind speed, relative humidity, atmospheric pressure, shortwave radiation, etc.

[0053] In the above step S2, the historical photovoltaic power data and historical meteorological data are preprocessed, the historical meteorological data is clustered under three-dimensional scale, and the different weather types are obtained through clustering to make the data accurate for the prediction model.

[0054] Specifically, step S2.1, data cleaning of historical photovoltaic power data, including data cleaning of historical photovoltaic power data, excluding outliers and zero negative values.

[0055] Step S2.2, obtaining historical meteorological data corresponding to the historical photovoltaic power data in the period, determining clustering elements based on correlation analysis, and obtaining different weather types through clustering.

[0056] Specifically, k-means clustering is used in this embodiment, and the process includes the following:

[0057] Step S2.2.1, based on the correlation between meteorological elements and photovoltaic power, through correlation analysis, determine the meteorological elements that affect photovoltaic power generation, and determine the meteorological elements as clustering elements.

[0058] Select the meteorological data corresponding to the date of the historical photovoltaic power data, and after correlation analysis, as shown in Table 1 below, the finally determined meteorological elements include atmospheric pressure, relative humidity, and shortwave radiation, which are used as clustering elements. Among them, the correlation analysis is: by calculating and statistically analyzing the correlation coefficient (including Pearson, Spearman, Kendall) of a meteorological element and photovoltaic power to comprehensively measure the correlation between the meteorological element and photovoltaic power generation.

[0059] Table 1 Correlation coefficient of meteorological elements and photovoltaic power

[0060]

[0061] Among them, Pearson is Pearson correlation coefficient, Spearman is Spearman rank correlation coefficient, and Kendall is Kendall rank correlation coefficient.

[0062] In this embodiment, three factors with strong correlation are selected from five meteorological factors as clustering basis, namely relative humidity, air pressure, and shortwave radiation.

[0063] Step S2.2.2, according to the clustering elements, using k-means clustering algorithm for weather clustering, and the clustering result is as shown in Figure 2 According to the range of each meteorological element corresponding to the clustering result, different weather types are finally determined. Three weather types determined in this embodiment are cloudy, sunny, and overcast (i.e. R, B, and G shown in Figure 2

[0064] ​In step S3, based on historical photovoltaic power data under different weather types, the cumulative distribution of photovoltaic power is obtained, and optimal copula function models for quantifying the spatial correlation of centralized and distributed photovoltaic power under different weather types are constructed. These optimal copula function models include the Frank Copula correlation function model and the VineCopula correlation function model.

[0065] Step S3.1: Based on the cumulative distribution of photovoltaic power, obtain the correlation coefficient value λ for each weather type and establish the Frank Copula correlation function model.

[0066] By observing the frequency distribution, such as Figure 3 (a) Figure 3 (b) Figure 3 As shown in (c), the frequency distribution differs under different weather types, but overall it satisfies symmetrical tail correlation. The Frank Copula correlation function is used for modeling, and the Frank Copula correlation function model is shown below:

[0067]

[0068] In the formula, u and v are two marginal distribution variables, and λ is the correlation coefficient.

[0069] Based on the cumulative distribution of photovoltaic power under each weather type, the correlation coefficient value λ is calculated, and then the Frank Copula correlation function model under each weather type is constructed.

[0070] Step S3.2: Based on the cumulative distribution of photovoltaic power, obtain the correlation coefficient value λ for each weather type and establish a Vine Copula correlation function model. The Vine Copula model includes not only the Frank Copula model mentioned above, but also the Clayton Copula model, the Gumbel Copula model, and the t Copula model. The connection structure of the Vine Copula correlation function model is as follows: Figure 4 As shown. Figure 4 In the diagram, U1, U2, and U3 represent marginal distribution variables, and C is obtained using the copula function based on U1 and U2. 21 C is obtained based on U2 and U3 using the copula function. 31 According to C 21 and C 31 Finally, we get C. 231 The selected copula function and its parameters are shown in the following formula and Table 2.

[0071] The Clayton Copula function model, Gumbel Copula function model, and t Copula function model are shown below:

[0072]

[0073]

[0074]

[0075] The parameters of the above model are shown in Table 2 below.

[0076] Table 2 Parameter values ​​for different models

[0077]

[0078] Where T1, T2, and T3 represent different Copula correlation function types used for each node, and λ1, λ2, and λ3 represent the correlation coefficient values ​​used in the corresponding Copula correlation function models. Furthermore, the correlation coefficient is calculated using the maximum likelihood estimation method based on the formulas (1), (2), (3), and (4) described above. This is generally transformed into solving the maximum / minimum problem of the above equations, which can be solved using differential equations.

[0079] In this embodiment, the Vine Copula correlation function model is used as the photovoltaic power prediction model, which improves the prediction accuracy to a certain extent and can better fit the power data. The Vine Copula model better describes the correlation of high-dimensional variables, overcomes the limitation of low-dimensional variables in the ordinary Copula model, and is suitable for fitting photovoltaic power data, making the model more flexible to build.

[0080] To further improve the accuracy of prediction, in this embodiment, for different weather types, the copula function model most suitable for that weather type is selected as the optimal copula function model, and power prediction is performed using the optimal copula function model.

[0081] Therefore, in step S3.3, the correlation coefficients and error evaluation indices of different copula function models under each weather type are compared, and the optimal prediction model corresponding to each weather type is selected from the Frank Copula correlation function model and the Vine Copula correlation function model.

[0082] In this embodiment, the correlation coefficients selected include the Pearson correlation coefficient and the coefficient of determination R. 2 The selected error evaluation index is the root mean square error (RMSE).

[0083] Specifically, based on historical photovoltaic power data of centralized photovoltaic power plants under different weather types (cloudy, sunny, overcast), photovoltaic power prediction for distributed photovoltaic power plants was performed using different copula function models. Correlation coefficients and error evaluation indices were calculated based on the predicted photovoltaic power data and historical photovoltaic power data of distributed photovoltaic power plants. The results are shown in Table 3 below. Based on the correlation coefficients and error evaluation indices, the optimal prediction model for each weather type was selected. The first selected error evaluation index was RMSE, and the correlation coefficient was R0. 2 Pearson, based on the error relationship and similarity between the actual values ​​and the model predictions, and according to the above evaluation indicators, selects the copula function model with the best indicators for each weather type as the optimal prediction model.

[0084] Table 3 Evaluation indicators of the Frank Copula and Vine Copula models under different weather types

[0085]

[0086] The Pearson correlation coefficient indicates the direction and degree of change between two variables, ranging from -1 to +1, where 0 represents no correlation, a positive value represents a positive correlation, and a negative value represents a negative correlation; the larger the value, the stronger the correlation. R0 2 The coefficient of determination, also known as the goodness of fit, is the square of the correlation coefficient r. It represents the portion of the variation in the dependent variable that can be explained by the variation in the independent variable. The magnitude of the coefficient of determination determines the strength of the correlation; a larger goodness of fit indicates a higher degree of explanation by the independent variable for the dependent variable, a higher percentage of the variation caused by the independent variable, and a denser concentration of observation points near the regression line. RMSE is the root mean square error.

[0087] In this embodiment, the optimal prediction model is selected: for cloudy and sunny weather types, the Frank Copula function model outperforms the Vine Copula function model in all error metrics, and the Frank Copula function model is preferred under these weather conditions; for overcast weather type, the Vine Copula function model outperforms the Frank Copula function model in all metrics, and the Vine Copula function model is preferred under these weather conditions. This embodiment uses clustering based on three weather types; the same method can be used to select the optimal prediction model for other weather types.

[0088] In step S4, the optimal copula function model under different weather types is used, with the obtained photovoltaic power prediction results of centralized photovoltaic power plants as input, to obtain the photovoltaic power point prediction results of distributed photovoltaic power plants through the corresponding model. Partial point prediction results are selected for each weather type, such as... Figure 5 (a) Figure 5 (b) Figure 5 As shown in (c), the point prediction model established in this embodiment yields the effect diagrams of point prediction under different weather types. Furthermore, the diagrams show differences in prediction results between different models, necessitating the selection of the optimal point prediction scheme for each weather type according to the aforementioned rules.

[0089] In step S5, a conditional probability model is constructed based on the quantile regression method, and the conditional probability prediction value corresponding to the predicted value of the distributed photovoltaic power point is obtained through the conditional probability model.

[0090] Finally, taking the optimal copula function model under the current weather type as the Vine Copula model as an example, the dependency structure between photovoltaic power and its conditional variables (such as total irradiance) is expressed flexibly and accurately in analytical form using the Vine Copula model. Considering that the Vine Copula function has stronger stability and can establish a multidimensional correlation model, a conditional probability model is constructed based on the Vine Copula function model using the idea of ​​quantile regression. The prediction results of centralized photovoltaic power and irradiance are used as inputs, and the probability prediction results of distributed photovoltaic power are obtained from the corresponding models. The conditional probability prediction values ​​corresponding to the point prediction values ​​are obtained through the conditional probability model, which is shown in formula (5).

[0091] Specifically, let the radiance value be x1, the photovoltaic power of the centralized power station be x2, and the photovoltaic power of the distributed power station be x3. Then the conditional probability distribution function of F(x3|x2,x1) is shown in equation (5) below:

[0092]

[0093] Where h(u,v) is related to the Copula function as follows: It should be noted that the Copula function here applies to all the Copula functions mentioned above.

[0094] Based on this, assuming the values ​​of each condition variable are x1 = R1 and x2 = R2, the conditional probability value of distributed photovoltaic power is calculated using the following equation:

[0095]

[0096] In the formula, h-1 (·) is the inverse function of h(·). This is the predicted value at the α quantile level, where α represents the quantile level and can be set to a custom value.

[0097] To further illustrate the superior performance of the Vine Copula model in this embodiment, simulation experiments were conducted. Based on the average width of different prediction intervals, the conditional probabilities of the predictions were evaluated and tested in practice. The prediction evaluation results are shown in Table 4.

[0098] Table 4. Prediction and evaluation results of different models

[0099]

[0100] Confidence intervals for point prediction, actual value, and probability prediction of distributed photovoltaic power under different weather types are as follows: Figure 6 (a) Figure 6 (b) Figure 6 As shown in (c), conditional probability prediction can be performed on all point predictions, and prediction information can be obtained on different confidence intervals. Table 4 shows the probability prediction results of the Vine Copula model, which is superior to the Frank Copula model, using only the average interval width.

[0101] The method described in this embodiment enables more accurate distributed photovoltaic power prediction and probabilistic prediction of distributed photovoltaic power, thereby improving the accuracy of power prediction and the reliability of power system operation, reducing electricity costs, reducing energy consumption, saving energy and reducing emissions, and improving economic benefits.

[0102] Example 2

[0103] This embodiment provides a photovoltaic power probability prediction system based on the copula function, including:

[0104] The data acquisition module is used to acquire historical photovoltaic power data and historical meteorological data for centralized and distributed photovoltaic power plants;

[0105] The data processing module is used to preprocess historical photovoltaic power data and historical meteorological data, and obtain different weather types through clustering;

[0106] The optimal copula function model construction module is used to obtain the cumulative distribution of photovoltaic power based on historical photovoltaic power data under different weather types, and to construct the optimal copula function model that quantifies the spatial correlation of centralized photovoltaic and distributed photovoltaic power under different weather types.

[0107] The photovoltaic power point prediction module is used to obtain the predicted value of distributed photovoltaic power points based on the photovoltaic power data of the centralized photovoltaic power station and the optimal copula function model under the corresponding weather type.

[0108] The photovoltaic power probability prediction module is used to construct a conditional probability model based on the quantile regression method, and obtain the conditional probability prediction value corresponding to the predicted value of distributed photovoltaic power points through the conditional probability model.

[0109] Example 3

[0110] This embodiment provides an electronic device, including a memory and a processor, as well as computer instructions stored in the memory and running on the processor. When the processor executes the computer instructions, it completes the steps in the photovoltaic power probability prediction method based on the copula function as described above.

[0111] Example 4

[0112] This embodiment also provides a computer-readable storage medium for storing computer instructions, which, when executed by a processor, complete the steps in the photovoltaic power probability prediction method based on the copula function as described above.

[0113] The steps and methods involved in Embodiments 2 to 4 above correspond to those in Embodiment 1. For specific implementation details, please refer to the relevant description section of Embodiment 1. The term "computer-readable storage medium" should be understood as a single medium or multiple media including one or more instruction sets; it should also be understood as including any medium capable of storing, encoding, or carrying an instruction set for execution by a processor and enabling the processor to perform any of the methods in this invention.

[0114] Those skilled in the art will understand that the modules or steps of the present invention described above can be implemented using general-purpose computer devices. Optionally, they can be implemented using computer-executable program code, thereby allowing them to be stored in a storage device for execution by a computer device, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. The present invention is not limited to any particular combination of hardware and software.

[0115] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0116] While the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. A photovoltaic power probabilistic forecasting method based on copula functions, characterized in that, The application comprises the following steps: acquiring historical photovoltaic power data and historical meteorological data of centralized and distributed photovoltaic power stations; preprocessing the historical photovoltaic power data and historical meteorological data, and obtaining different weather types through clustering; obtaining the cumulative distribution of photovoltaic power according to the historical photovoltaic power data under different weather types, and respectively constructing optimal copula function models quantitatively and dynamically representing the spatial correlation of centralized photovoltaic power and distributed photovoltaic power under different weather types; the constructed copula function models comprise Frank Copula correlation function models and Vine Copula correlation function models; the Vine Copula correlation function model is a combination of multiple binary Copula correlation function models; obtaining a distributed photovoltaic power point prediction value according to the acquired photovoltaic power data of the centralized photovoltaic power station and the optimal copula function model under the corresponding weather type; constructing a conditional probability model based on a quantile regression method, and obtaining a conditional probability prediction value corresponding to the distributed photovoltaic power point prediction value through the conditional probability model; Specifically, let the radiance value be , the photovoltaic power value of the centralized power station be , and the photovoltaic power value of the distributed power station be , then the conditional probability distribution function is as shown in the following formula: wherein The relationship with the Copula function is ; Based on this, let the value of each condition variable be given. , The conditional probability value of distributed photovoltaic power can be calculated using the following equation: wherein is the inverse function of is the prediction value at quantile level, parameter denotes the quantile level, which parameter takes a user-defined value.

2. The copula function based photovoltaic power probabilistic forecasting method according to claim 1, characterized in that, the Vine Copula correlation function model comprises Frank Copula function models, Clayton Copula function models, Gumbel Copula function models and t Copula function models.

3. The copula function based photovoltaic power probabilistic forecasting method of claim 1, wherein, acquiring historical meteorological data corresponding to the historical photovoltaic power data in a period, determining clustering elements based on correlation analysis, and obtaining different weather types through clustering, comprising the following steps: determining meteorological elements affecting photovoltaic power generation based on the correlation coefficient of meteorological elements and photovoltaic power, and taking the determined meteorological elements as clustering elements; performing weather clustering by using a k-means clustering algorithm according to the clustering elements, and determining different weather types according to the range of each meteorological element corresponding to the clustering result.

4. The copula function based photovoltaic power probabilistic forecasting method of claim 1, wherein, respectively constructing optimal copula function models quantitatively and dynamically representing the spatial correlation of centralized photovoltaic power and distributed photovoltaic power under different weather types, comprising the following steps: obtaining the correlation coefficient value under each weather type according to the cumulative distribution of photovoltaic power, and establishing a Frank Copula correlation function model; obtaining the correlation coefficient value under each weather type according to the cumulative distribution of photovoltaic power, and establishing a Vine Copula correlation function model; selecting the optimal prediction model corresponding to each weather type from the Frank Copula correlation function model and the Vine Copula correlation function model according to the correlation coefficient and error evaluation index of different copula function models under each weather type.

5. The copula function based photovoltaic power probabilistic forecasting method according to claim 4, characterized in that, The correlation coefficient includes a Pearson correlation coefficient and a determination coefficient R 2 The error evaluation index is a root mean square error RMSE.

6. A photovoltaic power probabilistic forecasting system based on copula functions, characterized in that, The application comprises the following steps: a data acquisition module for acquiring historical photovoltaic power data and historical meteorological data of centralized and distributed photovoltaic power stations; a data processing module for preprocessing the historical photovoltaic power data and historical meteorological data, and obtaining different weather types through clustering; The optimal copula function model construction module is configured to obtain cumulative distribution of photovoltaic power according to historical photovoltaic power data under different weather types, and to construct optimal copula function models for quantitatively representing dynamic spatial correlation of centralized photovoltaic power and distributed photovoltaic power under different weather types, respectively. The photovoltaic power point prediction module is configured to obtain a distributed photovoltaic power point prediction value according to the photovoltaic power data of the centralized photovoltaic power station and the optimal copula function model under the corresponding weather type. The constructed copula function models include a Frank Copula correlation function model and a Vine Copula correlation function model. The photovoltaic power probability prediction module is configured to construct a conditional probability model based on a quantile regression method, and to obtain a conditional probability prediction value corresponding to the distributed photovoltaic power point prediction value through the conditional probability model. Specifically, let the radiance value be , the photovoltaic power value of the centralized power station be , and the photovoltaic power value of the distributed power station be , then the conditional probability distribution function is as shown in the following formula: wherein The relationship with the Copula function is ; On this basis, given the value of each condition variable , , the conditional probability value of the distributed photovoltaic power is calculated by the following equation: wherein is the inverse function of is the prediction value at quantile level denotes the quantile level, which is a user-defined parameter.​​ 7. The copula function-based photovoltaic power probability prediction system of claim 6, wherein the Vine Copula function model includes a Frank Copula function model, a Clayton Copula function model, a Gumbel Copula function model, and a t Copula function model. The computer program product comprises a memory and a processor, and computer instructions stored in the memory and run on the processor, and when the computer instructions are run by the processor, the steps of the copula function-based photovoltaic power probability prediction method of any one of claims 1-5 are completed.

8. An electronic device, characterized by comprising: The computer program product is configured to store computer instructions, and when the computer instructions are executed by the processor, the steps of the copula function-based photovoltaic power probability prediction method of any one of claims 1-5 are completed.

9. A computer readable storage medium characterized by, ​

Images