Photovoltaic power generation associated physical quantity mining method based on historical time series data analysis, and related apparatus

Abstract

A photovoltaic power generation associated physical quantity mining method based on historical time series data analysis, and a related apparatus. The method comprises: acquiring historical time series data of multiple dimensions during photovoltaic power generation; calculating degrees of mutual information between the historical time series data of the multiple dimensions of photovoltaic power and the photovoltaic power; selecting the historical time series data of which the degree of mutual information satisfies a set requirement to serve as data related to photovoltaic data; and using a linear discriminant analysis (LDA) method to perform feature dimension reduction on the selected historical time series data to obtain data having undergone dimension reduction processing. In the present invention, by using a maximal information coefficient (MIC) feature selection method, data most related to photovoltaic power generation is selected from original feature variables, and then by using an LDA-based feature dimension reduction method, high-dimensional data is mapped to a lower-dimensional space. By means of the MIC feature selection method and the LDA-based feature dimension reduction method, the accuracy of photovoltaic power generation prediction is effectively improved.

Description

Photovoltaic power generation associated physical quantity mining method based on historical time series data analysis and related device

[0001] The present application claims priority from the Chinese patent application No. 202410615822.4 filed on May 17, 2024, and entitled "Photovoltaic power generation associated physical quantity mining method based on historical time series data analysis and related device", the content of which is incorporated herein by reference in its entirety. TECHNICAL FIELD

[0002] The present application belongs to the technical field of power system automation, and particularly relates to a photovoltaic power generation associated physical quantity mining method based on historical time series data analysis and related device. BACKGROUND

[0003] The influence of distributed photovoltaic grid connection on power distribution network safety, power quality and system stability cannot be ignored. In order to better cope with the uncertainty of new energy output and understand the characteristics of distributed new energy power output, photovoltaic power needs to be predicted. The historical time series data of photovoltaic power has large dimensions and large quantities. If the data is not processed, a large amount of redundant data will affect the accuracy of photovoltaic prediction.

[0004] SUMMARY

[0005] Therefore, the present application aims to provide a photovoltaic power generation associated physical quantity mining method based on historical time series data analysis and related device, which effectively improves the accuracy of photovoltaic power prediction through MIC feature selection and LDA feature dimension reduction method.

[0006] In order to solve the above technical problems, the present application provides the following technical solutions:

[0007] In the first aspect, the present application provides a photovoltaic power generation associated physical quantity mining method based on historical time series data analysis, comprising the following steps:

[0008] Obtaining historical time series data of each dimension in the photovoltaic power generation process;

[0009] Calculating the mutual information degree of the historical time series data of each dimension of the photovoltaic power and the photovoltaic power;

[0010] Selecting the historical time series data with mutual information degree meeting the set requirement as the data related to the photovoltaic data;

[0011] Using linear discriminant analysis method to perform feature dimension reduction on the selected historical time series data to obtain the data after dimension reduction processing.

[0012] Further, the calculation formula of the mutual information degree is as follows:

[0013] where MIC(X i ,Y) is the mutual information degree coefficient of the i-th dimension of the historical time series data X i and the photovoltaic power Y, P(x i ,y) is the frequency of the data points falling in the (x i ,y) grid, P(x i ) is the frequency of the data points falling in the x-th row, and P(y) is the frequency of the data points falling in the y-th row.

[0014] Further, the calculation formula of the frequency P(x i ,y) is as follows:

[0015] Further, the selected historical time series data is subjected to feature dimension reduction by using a linear discriminant analysis method, and the method specifically comprises the following steps:

[0016] obtaining original data based on the selected historical time series data;

[0017] calculating an intra-class scatter matrix S w ;

[0018] calculating an inter-class scatter matrix S b ;

[0019] calculating the eigenvalues and eigenvectors of the matrix , selecting the first d eigenvalues and the corresponding d eigenvectors in the order from small to large to obtain a projection matrix W;

[0020] for each sample feature parameter X i in the original data, converting the sample into a new sample by using the projection matrix;

[0021] obtaining data after dimension reduction processing based on the new sample.

[0022] Further, the calculation formula of the intra-class scatter matrix S w is as follows:

[0023] where K is the number of classes of samples, x is a sample in the data set, N k is the number of the k-th class of samples, D k is the number set of the k-th class of samples, and m k is the mean vector of the k-th class of samples.

[0024] Further, the calculation formula of the inter-class scatter matrix S b is as follows:

[0025] where m i is the mean vector of the i-th class of samples, and m jThe mean vector of the jth sample.

[0026] Further, the calculation formula of the new sample is as follows:

[0027] Z i = W T X i

[0028] In the formula, Z i is the new sample, W is the projection matrix, and X i is the sample before conversion.

[0029] In a second aspect, the present application provides a photovoltaic power generation associated physical quantity mining system based on historical time series data analysis, comprising:

[0030] A data acquisition unit is configured to acquire historical time series data of each dimension in a photovoltaic power generation process.

[0031] A mutual information calculation unit is configured to calculate the mutual information degree of the historical time series data of each dimension of photovoltaic power and photovoltaic power.

[0032] A data selection unit is configured to select historical time series data with a mutual information degree meeting a set requirement as data related to photovoltaic data.

[0033] A feature dimension reduction unit is configured to perform feature dimension reduction on the selected historical time series data by using a linear discriminant analysis method to obtain data after dimension reduction processing.

[0034] Correspondingly, the present application also provides a computer device, which comprises a processor and a memory:

[0035] The memory is configured to store a computer program and send instructions of the computer program to the processor.

[0036] The processor executes a photovoltaic power generation associated physical quantity mining method based on historical time series data analysis according to instructions of the computer program.

[0037] Correspondingly, the present application also provides a computer storage medium, which stores a computer program, and the computer program is executed by a processor to realize the photovoltaic power generation associated physical quantity mining method based on historical time series data analysis.

[0038] In summary, the application provides a photovoltaic power generation associated physical quantity mining method and related device based on historical time series data analysis, comprising obtaining historical time series data of each dimension in a photovoltaic power generation process; calculating mutual information degree of historical time series data of each dimension of photovoltaic power and photovoltaic power; selecting historical time series data meeting set requirements in mutual information degree as data related to photovoltaic data; and performing feature dimension reduction on the selected historical time series data by using a linear discriminant analysis method to obtain data after dimension reduction processing. The application combines the MIC mutual information coefficient feature selection method, filters out the most relevant data of photovoltaic power generation from original feature variables, and then performs feature dimension reduction by using the linear discriminant analysis (LDA) method to map high-dimensional data to a lower-dimensional space. The MIC feature selection and LDA feature dimension reduction method effectively improves the accuracy of photovoltaic power generation prediction. BRIEF DESCRIPTION OF DRAWINGS

[0039] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0040] Fig. 1 is a flowchart of a photovoltaic power generation associated physical quantity mining method based on historical time series data analysis provided by the embodiment of the present application;

[0041] Fig. 2 is a MIC calculation result diagram of 16-dimensional feature parameters in an example provided by the embodiment of the present application;

[0042] Fig. 3 is a scatter plot of feature parameter mapping to LDA in an example provided by the embodiment of the present application;

[0043] Fig. 4 is a comparison prediction diagram of a photovoltaic power prediction method provided by the embodiment of the present application. DETAILED DESCRIPTION

[0044] In order to make the purpose, features and advantages of the present application more obvious and easy to understand, the technical solutions in the embodiments of the present application will be described clearly and completely in combination with the drawings in the embodiments of the present application. Obviously, the following described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0045] Please refer to Fig. 1, the embodiment provides a photovoltaic power generation associated physical quantity mining method based on historical time series data analysis, comprising the following steps:

[0046] S1: Obtain historical time series data of each dimension in the photovoltaic power generation process.

[0047] It should be noted that the factors affecting photovoltaic power generation mainly include environmental temperature, solar elevation angle, photovoltaic components, etc. In order to mine the associated physical quantities of photovoltaic power generation, first, the historical time series data of each dimension in the photovoltaic power generation process is obtained.

[0048] S2: Calculate the mutual information degree of the historical time series data of each dimension of the photovoltaic power and the photovoltaic power.

[0049] It should be noted that mutual information refers to the degree of association between two random variables, that is, the degree of weakening of the uncertainty of another random variable given a random variable. Therefore, the minimum value of mutual information is 0, which means that given a random variable, there is no relationship between determining another random variable, and the maximum value is the entropy of the random variable, which means that given a random variable, it can completely eliminate the uncertainty of another random variable.

[0050] In this embodiment, step S1 obtains data of each dimension in the photovoltaic power generation process. In order to mine the relationship between each dimension data and the photovoltaic power generation power, the mutual information degree between each dimension data and the photovoltaic power generation power is calculated.

[0051] S3: Select historical time series data with mutual information degree meeting the set requirement as data related to photovoltaic data.

[0052] It should be noted that the foregoing step has analyzed the correlation degree between the multiple dimension data and the photovoltaic power generation power. On this basis, the embodiment sets a selection standard so as to select one or several dimension data more related to the photovoltaic power generation power from the numerous dimension data, so as to reduce the data redundancy degree on the basis of not affecting the photovoltaic prediction accuracy.

[0053] S4: Perform feature dimension reduction on the selected historical time series data by using a linear discriminant analysis method to obtain data after dimension reduction processing.

[0054] It should be noted that linear discriminant analysis (Linear Discriminant Analysis, LDA for short) is a commonly used multivariate statistical analysis method, which is usually used for classification and feature extraction. Its purpose is to find a linear transformation to project the data into a low-dimensional space under the condition of a set of labeled data, so that the data points of different categories can be more easily distinguished in the low-dimensional space. In short, the purpose of LDA is to project high-dimensional data into a low-dimensional space while maximizing the difference between categories and minimizing the difference within categories.

[0055] The selected feature data in the embodiment is mapped to a lower dimensional space by using a linear discriminant analysis feature dimension reduction method.

[0056] The application provides a photovoltaic power generation related physical quantity mining method based on historical time series data analysis. The original feature variables are screened out to obtain the most relevant data of photovoltaic power generation, and then a linear discriminant analysis (LDA) feature dimension reduction method is used to map high-dimensional data to a lower dimensional space. The MIC feature selection and LDA feature dimension reduction method effectively improve the prediction accuracy of photovoltaic power generation.

[0057] In one embodiment of the application, the mutual information degree of the feature parameters of each dimension of photovoltaic power and photovoltaic power is calculated according to the following formula:

[0058] In the formula, MIC (X i ,Y) is the mutual information degree coefficient of the i-th dimension of historical time series data X i and photovoltaic power Y, P (x i ,y) is the frequency of data points falling in the (x i ,y) grid, P (x i ) is the frequency of data points falling in the x row, and P (y) is the frequency of data points falling in the y row.

[0059] The data points of one dimension of feature parameters and photovoltaic power are distributed in a two-dimensional space, an m*n grid is used to divide the data space, the frequency of data points falling in the (x,y) grid is used as the estimation of P (x,y), the frequency of data points falling in the x row is used as the estimation of P (x), and the estimation of P (y) is obtained in the same way.

[0060] In a further embodiment of the application, the calculation formula of P (x i ,y) is as follows:

[0061] In one embodiment of the application, a linear discriminant analysis method is used for feature dimension reduction, and the specific process is as follows:

[0062] (1) The feature parameters of the MIC coefficient set condition (for example, greater than 0.2) are used to do the original data D = {(x1,y1),(x2,y2),…(xn,yn)};

[0063] (2) Assuming that the data set of the K-th class sample is D k , the mean vector of the changed sample is The variance of the Kth class sample is Wherein

[0064] The variance of the Kth class sample is

[0065] The within-class scatter matrix can be obtained by summing the sample variances of each class

[0066] (3) Calculate the between-class scatter matrix

[0067] The center distance between different classes i, j is

[0068] The between-class scatter matrix can be obtained by summing the distances between all classes

[0069] (4) Calculate the matrix S w -1 S b The eigenvalues and eigenvectors of S are selected in ascending order, and the first d eigenvalues and corresponding d eigenvectors are obtained to obtain the projection matrix W.

[0070] (5) For each sample feature X i in the sample set, convert it to a new sample Z i =W T X i ;

[0071] (6) Obtain the output sample D'={(z1,y1),(z2,y2),…(z m ,y m )}.

[0072] A photovoltaic power generation related physical quantity mining method based on historical time series data analysis is introduced below with an example.

[0073] In this example, there are 16 feature parameters, including air temperature, azimuth angle, cloud opacity, dew point temperature, DHI (solar scattered radiation index), DNI (solar direct radiation index), GHI (total solar horizontal radiation), GTI (fixed tilt angle radiation), RTI (tracking tilt angle radiation), atmospheric precipitable water, relative humidity, snow depth, ground pressure, 10m height wind direction, 10m height wind speed, and zenith angle. Each dimension has 7073 sample points. MIC feature selection is performed on the 16-dimensional parameters, and 9-dimensional feature parameters are retained. Figure 2 shows the MIC calculation results.

[0074] The characteristic parameters with the MIC coefficient greater than 0.2 are reserved, and the reserved parameters are air temperature, azimuth angle, cloud layer opacity, dew point temperature, DHI (solar scattered radiation index), DNI (solar direct radiation index), GHI (total horizontal solar radiation), GTI (fixed tilt angle radiation), RTI (tracking tilt angle radiation), atmospheric precipitable water, relative humidity, snow depth, ground pressure, 10m height wind direction, 10m height wind speed, and zenith angle. The sample data does not need to be classified, and therefore, the following is the non-classification LDA feature dimension reduction, and the dimension reduction dimension is set to 5 dimensions.

[0075] FIG. 3 is a scatter plot of 7073 sample points mapped to LDA, and 5-dimensional data is obtained after dimension reduction. The following is photovoltaic power prediction to verify whether the feature selection and feature dimension reduction affect the prediction performance of the data. The prediction simulation is performed using LSTM (Long Short Term Neural Network), the data sampling interval is 15 minutes, 1000 sample points are selected, the training set: verification set: test set is divided into 7:2:1. MAE (Mean Absolute Error), MAPE (Mean Absolute Percentage Error), and R2 (Determination Coefficient) are used as evaluation indexes. The LSTM setting parameters are as follows.

[0076] The prediction results are shown in FIG. 4 and the following table:

[0077] As shown in the error statistics table of the prediction results, the MIC feature rotation and LDA feature dimension reduction process does not affect the accuracy of the photovoltaic power prediction. After mapping the high-latitude data to low dimensions, the prediction accuracy is increased. This is because, under the given hyperparameters, the lower the feature dimension, the less computing power is needed. Therefore, the process of feature selection and feature dimension reduction of high-latitude data, that is, the mining of photovoltaic power-related physical quantities, is beneficial to the subsequent photovoltaic power prediction.

[0078] The method first screens the data most related to photovoltaic power generation through the MIC maximum mutual information coefficient feature selection method. The components with strong correlation with the original power sequence are screened out through the maximum information coefficient (MIC), and the application of linear discriminant analysis (LDA) in feature dimension reduction is combined with the MIC feature selection and LDA feature dimension reduction, which can reduce the calculation cost and improve the prediction accuracy, and provide strong support for the scheduling operation of distributed photovoltaic power generation. Future research can further optimize the feature selection and dimension reduction method to realize more efficient and accurate photovoltaic power prediction.

[0079] The above is a detailed introduction to an embodiment of the photovoltaic power-related physical quantity mining method based on historical time series data analysis of the present application. The following will introduce an embodiment of the photovoltaic power-related physical quantity mining system based on historical time series data analysis of the present application in detail.

[0080] The embodiment provides a photovoltaic power generation associated physical quantity mining system based on historical time series data analysis, comprising:

[0081] a data acquisition unit configured to acquire historical time series data of each dimension in a photovoltaic power generation process;

[0082] a mutual information calculation unit configured to calculate mutual information degrees of the historical time series data of each dimension of photovoltaic power and the photovoltaic power;

[0083] a data selection unit configured to select historical time series data with the mutual information degrees meeting a set requirement as data associated with photovoltaic data;

[0084] a feature dimension reduction unit configured to perform feature dimension reduction on the selected historical time series data by using a linear discriminant analysis method to obtain data processed by dimension reduction.

[0085] It should be noted that the photovoltaic power generation associated physical quantity mining system based on historical time series data analysis provided in the embodiment is based on the same inventive concept, and is proposed based on the photovoltaic power generation associated physical quantity mining method based on historical time series data analysis. The implementation scheme for solving problems provided by the system is similar to the implementation scheme described in the above method, and therefore will not be described here.

[0086] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the division of the above functional units and modules is exemplified, and in actual application, the above functions can be completed by different functional units and modules according to needs, that is, the internal structure of the system is divided into different functional units or modules to complete all or part of the functions described above. Each functional unit or module in the embodiment can be integrated in one processing unit, or each unit can exist physically, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or software. In addition, the specific names of each functional unit or module are only used for convenient distinction, and do not limit the protection scope of the application. The specific working process of the units and modules in the system can refer to the corresponding process in the foregoing method embodiments, and will not be described here.

[0087] The embodiment of the application further provides a computer device, comprising a memory, a processor and a computer program stored in the memory, and when the computer program is executed on the processor, the photovoltaic power generation associated physical quantity mining method based on historical time series data analysis is realized.

[0088] The computer device can be a desktop computer, a notebook computer, a palm computer, a cloud server, and the like. The computer device can include, but is not limited to, a processor and a memory. Those skilled in the art can understand that the embodiments are only examples of the computer device, and do not limit the computer device, which can include more or fewer components, or combine certain components, or different components, for example, can also include an input / output device, a network access device, and the like.

[0089] The processor can be a central processing unit (CPU), and can also be other general-purpose processors, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, and the like. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor.

[0090] The memory can be an internal storage unit of the computer device in some embodiments, for example, a hard disk or a memory of the computer device. The memory can also be an external storage device of the computer device in other embodiments, for example, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, and the like. Further, the memory can include both the internal storage unit and the external storage device of the computer device. The memory is used to store an operating system, an application program, a boot loader, data, and other programs, for example, program codes of the computer program, and the like. The memory can also be used to temporarily store data that has been output or will be output.

[0091] The embodiments of the present application also provide a computer readable storage medium, which stores a computer program. When the computer program is run by a processor, the method for mining a photovoltaic power generation related physical quantity based on historical time series data analysis is implemented.

[0092] In this embodiment, the integrated unit, if implemented in the form of a software functional unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on such understanding, the computer program for instructing the relevant hardware to complete all or part of the processes in the above-described embodiment methods can be stored in a computer readable storage medium. The computer program can be executed by a processor to implement the steps of the above-mentioned various method embodiments. The computer program includes computer program code, which can be in the form of source code, object code, executable file, or some intermediate form. The computer readable medium at least includes any entity or device capable of carrying the computer program code to the photographing device / terminal equipment, recording medium, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium. For example, U disk, mobile hard disk, magnetic disk or optical disk, etc. In some jurisdictions, according to legislation and patent practice, the computer readable medium can not be an electrical carrier signal and a telecommunication signal.

[0093] In the above embodiments, the description of each embodiment has its own focus, and the parts not described or recorded in detail in a certain embodiment can be referred to the relevant description of other embodiments.

[0094] Those skilled in the art can appreciate that the units and algorithm steps of the examples described in combination with the embodiments disclosed herein can be implemented in electronic hardware or a combination of computer software and electronic hardware. Whether the functions are implemented in hardware or software depends on the specific application and design constraints of the technical solutions. The skilled person can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0095] In the embodiments disclosed in the present application, it should be understood that the disclosed apparatus / terminal equipment and methods can be implemented in other ways. For example, the apparatus / terminal equipment embodiments described above are only schematic, for example, the division of the modules or units is only a logical function division, and actual implementation can have another division manner, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the shown or discussed mutual units can be indirect coupling or communication connection through some interfaces, devices or units, and can be electrical, mechanical or other forms.

[0096] The above examples are only used to illustrate the technical solutions of the present application, and are not intended to limit the present application; although the present application has been described in detail with reference to the foregoing examples, those skilled in the art should understand that the technical solutions recorded in the foregoing examples can be modified, or some technical features can be replaced by equivalent features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for mining associated physical quantities of photovoltaic power generation based on historical time series data analysis, characterized in that, The method comprises the following steps: acquiring historical time series data of each dimension in a photovoltaic power generation process; calculating mutual information degrees of the historical time series data of each dimension of photovoltaic power and photovoltaic power; selecting the historical time series data satisfying a set requirement as data related to photovoltaic data according to the mutual information degrees; performing feature dimension reduction on the selected historical time series data by using a linear discriminant analysis method to obtain data after dimension reduction. 2.The method of claim 1, wherein, The calculation formula of the mutual information degree is specifically as follows: where MIC(X i ,Y) is the mutual information coefficient of the i-th dimension of the historical time series X i and the photovoltaic power Y, P(x i ,y) is the frequency of the data points falling in the (x i ,y) grid, P(x i ) is the frequency of the data points falling in the x-th row, and P(y) is the frequency of the data points falling in the y-th row. 3.The method of claim 2, wherein, The calculation of the frequency P(x i ,y) is as follows: 4.The method of claim 1, wherein, The feature dimension reduction on the selected historical time series data by using the linear discriminant analysis method specifically comprises: obtaining original data based on the selected historical time series data; Computing the within-class scatter matrix S w ; Computing the inter-class scatter matrix S b ; Computing matrix selecting the first d eigenvalues and the corresponding d eigenvectors in ascending order to obtain a projection matrix W; for each sample feature parameter X in the original data i converts the new sample using the projection matrix; obtaining data after dimension reduction based on the new sample. 5.The method of claim 4, wherein, the intra-class scatter matrix S w the formula for the calculation of S where K is the number of classes of samples, x is a sample in the dataset, N k is the number of samples of the kth class, D k is the number set of the kth class, m k is the mean vector of the kth class. 6.The method of claim 4, wherein, the inter-class scatter matrix S b the formula for the calculation of S where m i is the mean vector of the ith class of samples, m j is the mean vector of the jth class of samples. 7.The method of claim 4, wherein, The calculation formula of the new sample is as follows: Z i = W T X i In the formula, Z i is the new sample, W is the projection matrix, X i is the sample before conversion.

8. A photovoltaic power generation related physical quantity mining system based on historical time series data analysis, characterized in that, The method comprises the following steps: a data acquisition unit configured to acquire historical time series data of each dimension in a photovoltaic power generation process; a mutual information calculation unit configured to calculate mutual information degrees of the historical time series data of each dimension of photovoltaic power and photovoltaic power; a data selection unit configured to select the historical time series data satisfying a set requirement as data related to photovoltaic data according to the mutual information degrees; a feature dimension reduction unit configured to perform feature dimension reduction on the selected historical time series data by using a linear discriminant analysis method to obtain data after dimension reduction.

9. A computer device, comprising: The device comprises a processor and a memory: the memory is configured to store a computer program and send instructions of the computer program to the processor; the processor is configured to execute the method according to the instructions of the computer program.

10. A computer storage medium, characterized in that, The computer readable storage medium stores a computer program, and the computer program is executed by the processor to realize the method according to any one of claims 1-7. The computer readable storage medium stores a computer program, and the computer program is executed by the processor to realize the method according to any one of claims 1-7.

Images