Medium-term power generation prediction method, system, electronic device and readable storage medium for large-scale distributed photovoltaic power stations

Abstract

The present invention provides a medium-term power generation prediction method, system, electronic device and readable storage medium for large-scale distributed photovoltaic power stations, wherein the medium-term power generation prediction method includes: obtaining historical operating condition data of each independent photovoltaic power station; judging whether the sample quantity of the historical operating condition data meets a preset prediction condition: when the sample quantity meets the prediction condition, predicting the power generation of the independent photovoltaic power station in the next month based on the historical operating condition data according to a pre-trained multivariate time series model; when the sample quantity does not meet the prediction condition, predicting the power generation of the independent photovoltaic power station in the next month according to the predicted values ​​of several neighboring photovoltaic power stations of the independent photovoltaic power station.

Description

Technical Field

[0001] The present invention relates to the field of distributed system consensus technology, and specifically discloses a medium-term power generation prediction method, system, electronic device and readable storage medium applied to a large-scale distributed photovoltaic power station. Background Art

[0002] my country is currently comprehensively promoting its new energy strategy. Following the completion of large-scale distributed photovoltaic power stations, a comprehensive management solution based on medium-term power generation forecasting technology for photovoltaic projects is directly related to the tangible economic benefits of power station owners, funders, and operation and maintenance managers. Accurately estimating the medium-term power generation of photovoltaic power stations is also crucial for grid planning improvements, power dispatch optimization, and power station operation and maintenance management. However, medium-term power generation forecasting is affected by factors such as the large number of civilian power stations, the widespread and scattered distribution of power stations, the lack of sufficient historical data samples for newly installed power stations, low weather forecast accuracy, and low-quality data from civilian power station equipment. This results in significant differences in technical methods between medium-term and short-term power generation forecasting, making short-term power generation forecasting techniques unsuitable for medium-term forecasting. Therefore, establishing an effective medium-term power generation forecasting method for distributed photovoltaic power stations is of great significance. Summary of the Invention

[0003] In view of the above-mentioned defects in the prior art, the present invention provides a medium-term power generation prediction method, system, electronic device and readable storage medium for large-scale distributed photovoltaic power stations.

[0004] Specifically, in a first aspect of the present application, a medium-term power generation forecasting method for a large-scale distributed photovoltaic power station is provided, wherein the large-scale distributed photovoltaic power station includes a plurality of independent photovoltaic power stations randomly distributed within a predetermined range;

[0005] The medium-term power generation forecasting method applied to large-scale distributed photovoltaic power stations can include:

[0006] Obtain historical operating data of each independent photovoltaic power station;

[0007] Determine whether the sample size of historical operating condition data meets the preset prediction conditions:

[0008] When the number of samples meets the prediction conditions, the pre-trained multivariate time series model is used to predict the next month's power generation of the independent photovoltaic power station based on historical operating data;

[0009] When the number of samples does not meet the prediction conditions, the power generation of the independent photovoltaic power station in the next month is predicted based on the predicted values ​​of several neighboring photovoltaic power stations of the independent photovoltaic power station.

[0010] In a possible implementation of the first aspect, the historical operating condition data includes historical monthly power generation and monthly weather data of the independent photovoltaic power station;

[0011] Monthly weather data includes the number of sunny days, rainy days, cloudy days, snowy days, monthly maximum temperature, monthly minimum temperature, monthly average high temperature, and monthly average low temperature.

[0012] In a possible implementation of the first aspect, the process of predicting the power generation of the independent photovoltaic power station for the next month based on the predicted values ​​of several neighboring photovoltaic power stations of the independent photovoltaic power station further includes:

[0013] Get the predicted effective working time T of each neighboring PV power station:

[0014]

[0015] Of which: Power pred is the predicted monthly total power generation of each neighboring PV power station, and P is the rated power of each neighboring PV power station;

[0016] Get the average predicted effective working time of nearby photovoltaic power plants

[0017]

[0018] in: is the predicted effective working time of each neighboring PV power station, and k is the number of neighboring power stations;

[0019] Get the predicted power generation y of an independent photovoltaic power station next month t :

[0020]

[0021] Where: P t is the rated power of an independent photovoltaic power station.

[0022] In a possible implementation of the first aspect, the process of predicting the power generation of the independent photovoltaic power station for the next month based on the predicted values ​​of several neighboring photovoltaic power stations of the independent photovoltaic power station further includes:

[0023] Obtaining independent photovoltaic power stations with a number not less than a preset threshold within a preset distance from the independent photovoltaic power station as neighboring photovoltaic power stations;

[0024] When the number of neighboring photovoltaic power stations that meet the above conditions is less than a preset threshold, all independent photovoltaic power stations are regarded as neighboring photovoltaic power stations.

[0025] In a possible implementation of the first aspect, the process of obtaining historical operating condition data of each independent photovoltaic power station further includes:

[0026] Based on the isolation forest algorithm, outlier detection is performed on historical operating condition data;

[0027] When there are outliers in the historical operating condition data, the Lagrange interpolation method is used to correct the outliers based on the normal values ​​in the historical operating condition data to achieve data cleaning of the historical operating condition data.

[0028] In a possible implementation of the first aspect, the multivariate time series model includes:

[0029]

[0030] Where Φ(B) is the autoregressive coefficient polynomial of the residual sequence;

[0031] Θ(B) is the moving average coefficient polynomial of the residual sequence;

[0032] a t is a zero-mean white noise sequence:

[0033] In a possible implementation of the first aspect, the process of predicting the next month's power generation of an independent photovoltaic power station based on historical operating condition data according to a multivariate time series model further includes:

[0034] The historical operating data of independent photovoltaic power plants are used as variable sequences, and a stationarity test is performed on the variable sequences.

[0035] When all variable sequences fail to pass the stationarity test, a cointegration test is performed on all variable sequences to obtain the difference order corresponding to the independent photovoltaic power station;

[0036] The variable sequence is converted into a stationary series according to the difference order.

[0037] A second aspect of the present application provides a medium-term power generation forecasting system for a large-scale distributed photovoltaic power station, which is applied to the medium-term power generation forecasting method for a large-scale distributed photovoltaic power station provided in the first aspect, including:

[0038] An acquisition unit, used to acquire historical operating condition data of each independent photovoltaic power station;

[0039] The judgment unit is used to judge whether the sample quantity of the historical operating condition data meets the preset prediction conditions:

[0040] The first prediction unit is configured to predict the power generation of the independent photovoltaic power station for the next month based on the historical operating condition data according to the pre-trained multivariate time series model when the number of samples meets the prediction conditions;

[0041] The second prediction unit is configured to predict the next month's power generation of the independent photovoltaic power station based on the predicted values ​​of several neighboring photovoltaic power stations if the number of samples does not meet the prediction conditions. A third aspect of the present application provides an electronic device comprising: a memory for storing a processing program; and a processor for implementing the medium-term power generation prediction method for large-scale distributed photovoltaic power stations provided in the first aspect when executing the processing program.

[0042] The fourth aspect of the present application provides a computer-readable storage medium, on which a processing program is stored. When the processing program is executed by a processor, the medium-term power generation prediction method for large-scale distributed photovoltaic power stations provided by the first aspect is implemented.

[0043] Compared with the prior art, this application has the following beneficial effects:

[0044] The technical solution proposed in this application systematically addresses the issues of low mid-term power generation forecast accuracy, difficulty in assessing expected economic benefits, and inaccurate power station operation and maintenance monitoring following the mass construction of large-scale distributed photovoltaic power stations under the national new energy policy. For financiers providing construction funding for the distributed photovoltaic industry, this technology can be used to manage revenue expectations and strengthen risk early warning capabilities on the asset side. For photovoltaic power station operation and maintenance managers, this technology can be used to accurately identify abnormally operating power stations, improving the management efficiency of large-scale power stations, and has widespread application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Other features, objects and advantages of the present invention will become more apparent upon reading the detailed description of non-limiting embodiments with reference to the following drawings:

[0046] Figure 1 According to an embodiment of the present application, a flow chart of a method for predicting medium-term power generation applied to a large-scale distributed photovoltaic power station is shown.

[0047] Figure 2 According to an embodiment of the present application, a structural schematic diagram of a medium-term power generation prediction system applied to a large-scale distributed photovoltaic power station is shown. DETAILED DESCRIPTION

[0048] The present invention will be described in detail below with reference to specific embodiments. The following examples will help those skilled in the art to further understand the present invention, but are not intended to limit the present invention in any form. It should be noted that, for those skilled in the art, several changes and improvements can be made without departing from the scope of the present invention. These all fall within the scope of protection of the present invention.

[0049] As is understandable, the power generation of photovoltaic power stations is affected by a variety of factors, including season, weather, and temperature. These complex and variable factors pose significant challenges to predicting photovoltaic power generation. This paper, based on numerical weather data, combines the ARIMAX model with a neighboring power station replacement algorithm to achieve medium-term power generation forecasting for distributed photovoltaic power stations. This method is capable of predicting the total power generation of large-scale distributed power stations for the next month, achieving prediction results that meet industrial standards.

[0050] Specifically, Figure 1 According to some embodiments of the present application, a method for predicting medium-term power generation for a large-scale distributed photovoltaic power station is provided, wherein the large-scale distributed photovoltaic power station includes a plurality of independent photovoltaic power stations randomly distributed within a predetermined range. The method for predicting medium-term power generation for a large-scale distributed photovoltaic power station may include:

[0051] Step 101: Obtain historical operating data of each independent photovoltaic power station.

[0052] Step 102: Determine whether the sample quantity of the historical operating condition data meets the preset prediction conditions.

[0053] Step 103: When the number of samples meets the prediction conditions, the power generation of the independent photovoltaic power station in the next month is predicted based on the historical operating condition data according to the pre-trained multivariate time series model.

[0054] Step 104: When the number of samples does not meet the prediction condition, the power generation of the independent photovoltaic power station in the next month is predicted based on the predicted values ​​of several neighboring photovoltaic power stations of the independent photovoltaic power station.

[0055] In some embodiments of the present application, the historical operating condition data includes the monthly power generation of the independent photovoltaic power station and monthly weather data. The monthly weather data includes the number of sunny days, rainy days, cloudy days, overcast days, snowy days, the monthly maximum temperature, the monthly minimum temperature, the monthly average high temperature, and the monthly average low temperature.

[0056] In some embodiments of the present application, the process of predicting the power generation of the independent photovoltaic power station in the next month based on the predicted values ​​of several neighboring photovoltaic power stations of the independent photovoltaic power station further includes:

[0057] Get the predicted effective working time T of each neighboring PV power station:

[0058]

[0059] Of which: Power pred is the predicted monthly total power generation of each neighboring PV power station, and P is the rated power of each neighboring PV power station;

[0060] Get the average predicted effective working time of nearby photovoltaic power plants

[0061]

[0062] in: is the predicted effective working time of each neighboring PV power station, and k is the number of neighboring power stations;

[0063] Get the predicted power generation y of an independent photovoltaic power station next month t :

[0064]

[0065] Where: P t is the rated power of an independent photovoltaic power station.

[0066] It is understandable that in the specific implementation process of the above embodiment, when it comes to distance calculation between power stations, the distance between power stations can be calculated in batches using matrix calculation. The longitude and latitude coordinates of the power stations are constructed as a k×2 matrix C k×2 , respectively extract the longitude coordinate matrix Lng k×1 and the latitude coordinate matrix Lat k×1 ,The latitude and longitude difference matrix of each power station is:

[0067]

[0068]

[0069] The distance matrix between all power stations can be calculated based on the formula for calculating distance based on longitude and latitude:

[0070]

[0071] Where R is the radius of the earth and D is the distance matrix between power stations.

[0072] In some embodiments of the present application, the process of predicting the power generation of the independent photovoltaic power station in the next month based on the predicted values ​​of several neighboring photovoltaic power stations of the independent photovoltaic power station further includes:

[0073] Obtain independent photovoltaic power stations with a quantity not less than a preset threshold within a preset distance of an independent photovoltaic power station as neighboring photovoltaic power stations;

[0074] In the case where the quantity of neighboring photovoltaic power stations satisfying the above conditions is less than the preset threshold, all independent photovoltaic power stations are taken as neighboring photovoltaic power stations.

[0075] In some embodiments of the present application, in the process of obtaining the historical operating conditions data of each independent photovoltaic power station, it further includes:

[0076] Based on the isolation forest algorithm, perform outlier detection on the historical operating conditions data.

[0077] It can be understood that in the implementation of the specific steps of the above embodiments, the monthly average equivalent duration is calculated for all power stations in each city each month, and the isolation forest algorithm is used to perform outlier detection on the power stations according to the average equivalent duration. Specifically, the isolation forest outlier detection algorithm model is constructed using data dimensions such as the equipment attribute information, operation information, and fault log information of the photovoltaic power station. The algorithm is introduced as follows.

[0078] First, give the definitions of the isolation tree (Isolation Tree) and the path length h((x) of the sample point x in the isolation tree:

[0079] Isolation tree: If it is a node of the T isolation tree, there are two cases for T: an external node without child nodes, and an internal node with two child nodes (T l ,T r ) and a test. The test in T consists of an attribute q and a splitting point p. Points where q < p belong to T l , and vice versa belong to T r .

[0080] h(x): The number of edges passed by the sample point x from the root node of the iTree to the leaf node.

[0081] [[ID=三十二]]The first stage of the algorithm:

[0082] Step1: X = x1,…,x n is a given data set, Randomly select ψ sample points from X to form a subset X′ of X and place them in the root node.

[0083] Step2: Randomly specify a dimension q from d dimensions, and randomly generate a cut point p in the current data, min(x ij ,j = q,x ij ∈X′) < p < max(x ij ,j = q,x ij ∈X′).

[0084] Step 3: This cutting point p generates a hyperplane that divides the current data space into two subspaces: sample points with a specified dimension less than p are placed in the left child node, and those greater than or equal to p are placed in the right child node.

[0085] Step 4: Recursively repeat Step 2 and Step 3 until all leaf nodes have only one sample point or the isolated tree iTree has reached the specified height.

[0086] Step 5: Loop Step 1 to Step 4 until t isolated trees iTree are generated.

[0087] The second stage of the algorithm:

[0088] Step 1: For each data point x i , let it traverse each isolated tree (iTree) and calculate the point x i The average height h(x) in the forest is normalized by the average height of all points. The formula for calculating the outlier score is as follows:

[0089]

[0090] in,

[0091]

[0092] When there are outliers in the historical operating condition data, the Lagrange interpolation method is used to correct the outliers based on the normal values ​​in the historical operating condition data to achieve data cleaning of the historical operating condition data.

[0093] It is understood that in the above embodiment, for abnormal power stations, the Lagrangian interpolation method is used to correct the abnormal months based on the power generation of the power station in the past months. The Lagrangian interpolation method is defined as follows:

[0094] For a polynomial function, it is known that there are given k+1 value points:

[0095] (x0,y0),…,(x k ,y k )

[0096] where x j Corresponding to the position of the independent variable, y j The value of the corresponding function at this position. Assume that any two different x j are all different from each other, then the Lagrange difference polynomial obtained by applying the Lagrange interpolation formula is:

[0097]

[0098] Each of the lj (x) is the Lagrange basic polynomial, and its expression is:

[0099]

[0100] In a possible implementation of the first aspect, the multivariate time series model includes:

[0101]

[0102] Where Φ(B) is the autoregressive coefficient polynomial of the residual sequence;

[0103] Θ(B) is the moving average coefficient polynomial of the residual sequence;

[0104] a t is a zero-mean white noise sequence:

[0105] In some embodiments of the present application, further, in the process of predicting the power generation of the independent photovoltaic power station next month based on the historical operating condition data according to the multivariate time series model, the following steps are further included:

[0106] The historical operating data of independent photovoltaic power plants are used as variable sequences, and a stationarity test is performed on the variable sequences.

[0107] When all variable sequences fail to pass the stationarity test, a cointegration test is performed on all variable sequences to obtain the difference order corresponding to the independent photovoltaic power station;

[0108] The variable sequence is converted into a stationary series according to the difference order.

[0109] It can be understood that in a possible implementation of the above embodiment, for each power station, it is necessary to first perform a unit root test on the response sequence and the independent variable sequence to be substituted into the ARIMAX model. If all the original sequences do not have unit roots, then all the original sequences are stationary and the stationarity test passes. If some of the original sequences are stationary and some are not, it is necessary to perform a cointegration test on all variable sequences. Cointegration means that two or more variable sequences are non-stationary, but they have a long-term equilibrium relationship. If the regression residual sequences of two or more sequences are stationary, then the two or more sequences have a cointegration relationship. Through the cointegration test, it can be effectively avoided that the ARIMAX established for each power station is a pseudo-regression model. After performing unit integration and cointegration tests on all power stations using the ARIMAX model, the difference order d that makes each power station pass the test is recorded. After the variable sequence of each power station is converted into a stationary sequence through the d-order difference, the ARIMAX model is built for the power station data.

[0110] In the above embodiment, further, during the ARIMAX model establishment process, the orders of p and q are first determined: During the model fitting process, the orders of p and q in the ARIMAX model need to be determined. Here, relative optimal model identification is used to determine p and q. A range of p and q values ​​is taken, typically from 0 to 0.1 times the sequence length. The BIC information content of all p and q combinations is observed, and the p and q parameter combination with the lowest information content is selected.

[0111] The function of BIC information is defined as:

[0112] BIC=-2ln(L)+ln(n)*k

[0113] Where L is the maximum likelihood under the model, n is the number of samples, and k is the number of model variables. According to the relative optimal model, the value range of p and q is Where N is the sequence length, then The values ​​are:

[0114]

[0115] When BIC(p,q) reaches its minimum, determine the corresponding is the order of the model.

[0116] In the above embodiment, after the p and q orders are determined, the least squares method can be used to solve the model parameters and predict the power generation of the power station.

[0117] In some embodiments of the present application, Figure 2 A medium-term power generation forecasting system for a large-scale distributed photovoltaic power station is shown, which is applied to the medium-term power generation forecasting method for a large-scale distributed photovoltaic power station provided in the above embodiment, and specifically may include:

[0118] The acquisition unit 001 is used to acquire the historical operating data of each independent photovoltaic power station.

[0119] The judgment unit 002 is used to judge whether the sample quantity of the historical operating condition data meets the preset prediction conditions.

[0120] The first prediction unit 003 is used to predict the power generation of the independent photovoltaic power station in the next month based on the historical operating condition data according to the pre-trained multivariate time series model when the number of samples meets the prediction conditions.

[0121] The second prediction unit 004 is configured to predict the power generation of the independent photovoltaic power station in the next month based on the predicted values ​​of several neighboring photovoltaic power stations of the independent photovoltaic power station when the number of samples does not meet the prediction condition.

[0122] It is understandable that the functions implemented by the various functional modules in the above-mentioned medium-term power generation forecasting system are consistent with the implementation of the various steps in the power generation forecasting method in the above-mentioned embodiment, and will not be described in detail here.

[0123] Some embodiments of the present application also provide an electronic device. This electronic device includes a memory and a processor, wherein the memory is used to store a processing program, and the processor executes the processing program according to instructions. When the processor executes the processing program, the medium-term power generation forecasting method for a large-scale distributed photovoltaic power station described in the aforementioned embodiments is implemented.

[0124] In some embodiments of the present application, a readable storage medium is further provided. The readable storage medium may be a non-volatile readable storage medium or a volatile readable storage medium. The readable storage medium stores instructions that, when executed on a computer, cause an electronic device containing the readable storage medium to execute the aforementioned method for medium-term power generation forecasting for a large-scale distributed photovoltaic power station.

[0125] It is understandable that, for each functional module in the aforementioned medium-term power generation forecasting system, if they are all implemented in the form of software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the various embodiments of the method of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, and other media that can store program code.

[0126] Computer-readable storage media may include a data signal propagated in baseband or as part of a carrier wave, which carries readable program code. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. The readable storage medium may also be any readable medium other than a readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on the readable storage medium may be transmitted using any appropriate medium, including but not limited to wireless, wired, optical cable, RF radio frequency, etc., or any suitable combination thereof.

[0127] The program code used to execute the technical solutions disclosed in this application can be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java, C++, etc., and conventional procedural programming languages ​​such as C or similar programming languages. The program code can be executed entirely on the user computing device, partially on the user device, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (for example, using an Internet service provider to connect via the Internet).

[0128] In summary, through the technical solution proposed in this application, in the process of model establishment, for the screening of abnormal power stations, the isolation forest algorithm is used to detect abnormal power stations through data dimensions such as equipment attribute information, operation information, and fault log information, and the power generation data of abnormal power stations are interpolated and replaced by the Lagrange interpolation method to ensure the standardization of model training data and the training effect of the model. In the training and prediction of the model, relatively stable data dimensions such as the monthly average and extreme values ​​of temperature, and the number of days of various weather types are used to avoid the complex and changeable medium-term climate conditions. At the same time, for power stations with less data, the algorithm of equivalent time replacement of neighboring power stations is used to make up for the defect of less historical data of power stations, and the algorithm has higher time efficiency after optimization.

[0129] The application effect of this technical solution is verified by a specific embodiment below:

[0130] During the verification process, the deviation is used as an evaluation indicator for the application effect of a single power station, where:

[0131]

[0132] Furthermore, by calculating the deviation distribution description of all power stations in the test set, including the mean, median, quantile, etc., the corresponding verification and evaluation test set is obtained:

[0133]

[0134] Furthermore, the monthly deviation distribution of all power plants in the test set, the monthly deviation of total power generation, and the overall deviation between the actual and predicted values ​​of total power generation of all power plants are evaluated, where:

[0135]

[0136]

[0137] It can be seen that the overall prediction deviation percentage is controlled within the range of about 1%, and the prediction accuracy is high.

[0138] The embodiments of the present disclosure are described in detail above with reference to the accompanying drawings, but the present disclosure is not limited to the above embodiments. Even if various changes are made to the present disclosure, if these changes fall within the scope of the claims of the present disclosure and their equivalents, they will still fall within the scope of protection of the present disclosure.

[0139] The above describes specific embodiments of the present invention. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art may make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. The embodiments of this application and the features in the embodiments may be combined with each other in any manner unless there is a conflict.

Claims

1. A method for predicting medium-term power generation for a large-scale distributed photovoltaic power station, wherein the large-scale distributed photovoltaic power station comprises a plurality of independent photovoltaic power stations randomly distributed within a predetermined range; It is characterized in that The medium-term power generation forecasting method includes: Obtaining historical operating data of each independent photovoltaic power station; Determine whether the sample size of the historical operating condition data meets the preset prediction conditions: When the number of samples meets the prediction condition, predicting the power generation of the independent photovoltaic power station next month based on the historical operating condition data according to the pre-trained multivariate time series model; When the number of samples does not meet the prediction condition, the power generation of the independent photovoltaic power station in the next month is predicted based on the predicted values ​​of several neighboring photovoltaic power stations of the independent photovoltaic power station.

2. The method for predicting mid-term power generation for a large-scale distributed photovoltaic power station according to claim 1, wherein: The historical operating condition data includes the historical monthly power generation and monthly weather data of the independent photovoltaic power station; The monthly weather data includes the number of sunny days, rainy days, cloudy days, snowy days, the highest temperature, the lowest temperature, the average high temperature and the average low temperature.

3. The method for predicting mid-term power generation for a large-scale distributed photovoltaic power station according to claim 1, wherein: The process of predicting the power generation of the independent photovoltaic power station in the next month based on the predicted values ​​of several neighboring photovoltaic power stations of the independent photovoltaic power station also includes: Obtain the predicted effective working time T of each of the adjacent photovoltaic power stations: in: is the predicted monthly total power generation of each of the adjacent photovoltaic power plants, The rated power of each of the adjacent photovoltaic power plants; Get the average predicted effective working time of the neighboring photovoltaic power station ; in: The predicted effective working time of each of the adjacent photovoltaic power plants, is the number of neighboring power stations; Get the predicted power generation of the independent photovoltaic power station next month : in: is the rated power of the independent photovoltaic power station.

4. The method for predicting mid-term power generation for a large-scale distributed photovoltaic power station according to claim 1, wherein: The process of predicting the power generation of the independent photovoltaic power station in the next month based on the predicted values ​​of several neighboring photovoltaic power stations of the independent photovoltaic power station also includes: Acquire the independent photovoltaic power stations whose number is not less than a preset threshold within a preset distance from the independent photovoltaic power station as the neighboring photovoltaic power stations; When the number of the adjacent photovoltaic power stations that meet the above conditions is less than the preset threshold, all the independent photovoltaic power stations are regarded as the adjacent photovoltaic power stations.

5. The method for predicting mid-term power generation for a large-scale distributed photovoltaic power station according to claim 1, wherein: The process of obtaining the historical operating condition data of each independent photovoltaic power station further includes: Performing outlier detection on the historical operating condition data based on the isolation forest algorithm; In the case where the abnormal value exists in the historical operating condition data, the abnormal value is corrected based on the normal value in the historical operating condition data using Lagrange interpolation method to achieve data cleaning of the historical operating condition data.

6. The method for predicting mid-term power generation for a large-scale distributed photovoltaic power station according to claim 1, wherein: The multivariate time series model includes: in, is the autoregressive coefficient polynomial of the residual sequence; is the moving average coefficient polynomial of the residual sequence; is a zero-mean white noise sequence: .

7. The method for predicting mid-term power generation for a large-scale distributed photovoltaic power station according to claim 6, wherein: In the process of predicting the power generation of the independent photovoltaic power station next month based on the historical operating condition data according to the multivariate time series model, the method further includes: Taking the historical operating condition data of the independent photovoltaic power station as a variable sequence, and performing a stationarity test on the variable sequence; In the case that all the variable sequences fail to pass the stationarity test, performing a cointegration test on all the variable sequences to obtain the difference order corresponding to the independent photovoltaic power station; The variable sequence is converted into a stationary sequence according to the difference order.

8. A medium-term power generation forecasting system for large-scale distributed photovoltaic power stations, characterized in that: The method for predicting medium-term power generation of a large-scale distributed photovoltaic power station as claimed in any one of claims 1 to 7 comprises: an acquisition unit, configured to acquire historical operating condition data of each of the independent photovoltaic power stations; A judgment unit is used to judge whether the sample quantity of the historical operating condition data meets the preset prediction conditions: A first prediction unit is configured to predict the power generation of the independent photovoltaic power station next month based on the historical operating condition data according to a pre-trained multivariate time series model when the number of samples meets the prediction condition; The second prediction unit is configured to predict the power generation of the independent photovoltaic power station in the next month based on the predicted values ​​of several neighboring photovoltaic power stations of the independent photovoltaic power station when the number of samples does not meet the prediction condition.

9. An electronic device, characterized in that: include: a memory for storing a processing program; A processor, wherein when executing the processing program, the processor implements the medium-term power generation prediction method applied to a large-scale distributed photovoltaic power station as described in any one of claims 1 to 7.

10. A readable storage medium, characterized in that: The readable storage medium stores a processing program, and when the processing program is executed by the processor, the method for predicting medium-term power generation applied to a large-scale distributed photovoltaic power station according to any one of claims 1 to 7 is implemented.

Images