Wind power plant output correlation analysis method based on continuous sorting

Abstract

The invention discloses a wind power plant output correlation analysis method based on continuous sorting, and belongs to the technical field of new energy grid-connected scheduling control. Determining a wind speed target correlation matrix, a wind speed sample matrix and a wind speed sample correlation matrix; determining a sorting matrix for replacement, and enabling the wind speed sample correlation matrix to be close to the wind speed target correlation matrix; constructing a relative error matrix based on the wind speed sample correlation matrix, and calculating a total error; and adjusting the wind speed target correlation matrix based on the total error, and outputting a wind power plant output correlation analysis result according to the converged wind speed target correlation matrix. According to the method, the problem that a traditional Cholesky method cannot process a non-positive definite correlation matrix is solved, the method is adaptive to a wind power plant cluster scene with dense geographically adjacent wind fields and complex wind speed correlation, and the effectiveness of correlation analysis in the scene is guaranteed; a relative error matrix is established to evaluate the deviation between the wind speed sample and the original wind speed data, and the blank that the replacement deviation is ignored in a traditional method is filled.

Description

Technical Field

[0001] This invention belongs to the field of new energy grid-connected dispatch and control technology, and relates to a method for analyzing the correlation of wind farm output based on continuous sorting. Background Technology

[0002] The output of renewable energy generation (represented by wind farms) is typically affected by changes in natural conditions. In modern power systems with high wind power penetration, some wind farms are installed in geographically adjacent areas where wind speeds are correlated. To ensure the accuracy of steady-state correlation calculations, such as probabilistic power flow (PPF) calculations and steady-state security analysis, the correlation between wind power outputs needs to be fully considered.

[0003] Currently, the copula method and the permutation method are the most commonly used methods for analyzing wind power correlation. The copula method constructs a multidimensional distribution function (MDF) using a copula function, from which relevant samples can be obtained. Different types of copula functions, such as Gaussian copula and D-vine copula, are used to model the correlation output of wind farms. However, the copula method heavily relies on time-consuming MDF sampling. Furthermore, to accurately describe the correlation of random variables, different copula functions need to be combined, increasing the computational burden. The permutation method generates a sample matrix containing relevant information by permuting samples of each variable. Because the permutation method does not rely on the MDF, it has a wider range of applications, for example, it can be combined with Monte Carlo methods (MCM) and point estimation methods (PEM), which are widely used in power system uncertainty analysis. Moreover, the permutation method does not require time-consuming MDF sampling. The most commonly used permutation method is the Cholesky decomposition method, used to permutate samples and generate a sample matrix containing relevant information. However, the Cholesky decomposition method is not suitable for cases where the correlation matrix of the variables is positive definite. Furthermore, none of the aforementioned studies considered the correlation matrix bias caused by permutations, which could lead to errors in correlation analysis and steady-state calculations. Therefore, fully considering the correlations among wind power outputs is one of the important technical challenges in this field. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a wind farm output correlation analysis method based on continuous sorting, which aims to solve the problems mentioned in the background technology.

[0005] In a first aspect, the present invention provides a method for correlation analysis of wind farm output based on continuous sorting, comprising the following steps: Determine the target wind speed correlation matrix, the wind speed sample matrix, and the wind speed sample correlation matrix; Determine the sorting matrix to be used for the replacement, so that the wind speed sample correlation matrix is ​​close to the wind speed target correlation matrix; Based on the correlation matrix of wind speed samples, a relative error matrix is ​​constructed, and the overall error is calculated. Based on the overall error, the wind speed target correlation matrix is ​​adjusted, and the wind farm output correlation analysis results are output according to the converged wind speed target correlation matrix.

[0006] Furthermore, the specific steps are as follows: Obtain raw wind speed samples from multiple wind farms, determine the target wind speed correlation matrix and the raw wind speed sample matrix, and calculate the initial wind speed sample correlation matrix based on the raw wind speed sample matrix. Eigenvalue decomposition is performed on the wind speed target correlation matrix and the wind speed sample correlation matrix respectively; a ranking matrix is ​​constructed, and the current wind speed sample matrix is ​​permuted based on the ranking matrix to obtain the updated wind speed sample matrix, so that the wind speed sample correlation matrix corresponding to the updated wind speed sample matrix approximates the wind speed target correlation matrix. Calculate the corresponding wind speed sample correlation matrix based on the updated wind speed sample matrix, construct the relative error matrix, and calculate the overall error. If the overall error does not meet the preset convergence condition, the wind speed target correlation matrix is ​​adjusted based on the relative error matrix, and the steps of performing eigenvalue decomposition on the wind speed target correlation matrix and the wind speed sample correlation matrix are iteratively executed until the preset convergence condition is met; if the preset convergence condition is met, the optimized wind speed sample matrix and the corresponding wind farm output correlation analysis results are output.

[0007] Further, the target wind speed correlation matrix, the sample wind speed matrix, and the sample wind speed correlation matrix are determined, specifically through the following steps: Wind speed is typically modeled as a correlated random variable, and its relevant information is described by the wind speed target correlation matrix. As shown in formula (1): (1); In the formula: It is the number of random variables; It is the Pearson correlation coefficient between the first random variable and the second random variable; It is the Pearson correlation coefficient between the first random variable and the d-th random variable; It is the Pearson correlation coefficient between the second random variable and the first random variable; It is the Pearson correlation coefficient between the second random variable and the d-th random variable; It is the Pearson correlation coefficient between the d-th random variable and the 1st random variable; It is the Pearson correlation coefficient between the d-th random variable and the 2nd random variable; Suppose that by performing these random variables Second sampling, generate Wind speed sample matrix As shown in formula (2): (2); Where: the result of each sample The wind speed sample matrix consists of wind speed samples of several random variables. A single line; T is the transpose; It is the wind speed sample vector of the first random variable; It is the wind speed sample vector of the second random variable; It is the wind speed sample vector of the kth random variable; Based on wind speed sample matrix Calculate the wind speed sample correlation matrix That is, to calculate the wind speed sample matrix The mean and standard deviation of each column of wind speed samples are calculated, and then the covariance of any two columns of wind speed samples is calculated. Based on the covariance and standard deviation, the wind speed correlation coefficient between any two corresponding wind farms is calculated. The wind speed correlation coefficients between all wind farms are arranged according to the row and column order of the wind farms to obtain a d×d dimensional wind speed sample correlation matrix. .

[0008] Furthermore, feature decomposition is performed on the wind speed target correlation matrix and the wind speed sample correlation matrix, specifically through the following steps: Wind speed target correlation matrix Correlation matrix with wind speed samples The eigenvalues ​​are decomposed separately, as shown in formula (3): (3); In the formula: It is the diagonal matrix of eigenvalues ​​corresponding to the correlation matrix. This represents a matrix composed of eigenvectors; It is the eigenvector matrix of the wind speed sample correlation matrix; It is the diagonal matrix of eigenvalues ​​of the wind speed sample correlation matrix; It is the eigenvector matrix of the wind speed target correlation matrix; It is the diagonal matrix of eigenvalues ​​of the wind speed target correlation matrix.

[0009] Furthermore, a sorting matrix is ​​constructed, and the current wind speed sample matrix is ​​permuted based on the sorting matrix to obtain an updated wind speed sample matrix. The specific steps are as follows: Construct a sorting matrix As shown in formula (4): (4); According to the sorting matrix For the current wind speed sample matrix Arrange the items; set up Represents the wind speed sample matrix The Middle line, number Column entries, Represents the sorting matrix The Middle line, number Column entries; turn up Then the wind speed sample matrix The Middle line, number The column entries were adjusted to the [number]th [item]. OK; In the formula: Indicates the rank of an item in the column; Obtain the updated wind speed sample matrix .

[0010] Further, a relative error matrix is ​​constructed, and the overall error is calculated, specifically through the following steps: To represent the arrangement deviation, a relative error matrix is ​​established. Relative error matrix Entries As shown in formula (5): (5); In the formula: It is the wind speed sample correlation matrix The Middle Line number Column entries; Correlation matrices with wind speed The Middle Line number Column entries; Calculate the total error As shown in formula (6): (6). Furthermore, the process of adjusting the wind speed target correlation matrix based on the overall error is a continuous permutation process achieved through an iterative algorithm, i.e., a continuous arrangement process; Based on the overall error, the wind speed target correlation matrix is ​​adjusted. Then, based on the converged wind speed target correlation matrix, the wind farm output correlation analysis results are output. The specific steps are as follows: Based on total error Adjusting the wind speed target correlation matrix Then the adjusted wind speed target correlation matrix Rearrange and set the updated wind speed sample matrix. Through the first This is generated by permutation, i.e. , It is the wind speed sample matrix generated by the (k+1)th permutation; By solving formulas (5) and (6), the relative error matrix of the (k+1)th permutation is determined. The total error of the (k+1)th permutation ; The wind speed correlation matrix for the (k+1)th permutation is updated by adjusting the entries. The update method is shown in formula (7): (7); In the formula: It is the Pearson correlation coefficient between the i-th random variable and the j-th variable in the (k+1)-th permutation; It is the Pearson correlation coefficient between the i-th random variable and the j-th variable in the k-th permutation; These are the entries of the relative error matrix for the (k+1)th permutation; It is the weight, representing the step size, which affects the speed and accuracy of iterative convergence.

[0011] Secondly, the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements a wind farm output correlation analysis method based on continuous sorting.

[0012] Thirdly, the present invention provides an electronic device including at least one processor and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to execute a wind farm output correlation analysis method based on continuous sorting.

[0013] The present invention has the following beneficial effects: (1) By using the eigenvalue decomposition method in CPED to permutate the sample matrix, the problem that the traditional Cholesky method cannot handle non-positive definite correlation matrices is solved. This method is adapted to wind farm cluster scenarios with densely adjacent geographical wind fields and complex wind speed correlations (which are prone to non-positive definite matrices), ensuring the effectiveness of correlation analysis in such scenarios. A relative error matrix is ​​established to evaluate the deviation between the wind speed sample and the original wind speed data, filling the gap in traditional methods that ignore permutation deviation. Continuous permutation is achieved through an iterative algorithm, and permutation deviation is reduced by adjusting the target correlation matrix. This reduces the overall error of wind farm output correlation analysis by more than 80%, significantly improving the analysis. Accuracy; From the perspective of wind farm application, the above-mentioned technical advantages can be transformed into tangible value: On the one hand, high-precision correlation analysis can accurately characterize the output characteristics of wind farms, providing reliable correlation basis data for wind farm output prediction, helping wind farms optimize power generation scheduling strategies, reduce wind curtailment losses caused by output fluctuations, and maximize power generation benefits; on the other hand, the output high-precision correlation results can directly support power system probabilistic power flow calculation and steady-state security analysis, providing a strong basis for the grid to scientifically accept wind farm output, effectively improving the wind power absorption capacity in areas with high wind power penetration, and ensuring the safety and stability of the coordinated operation of the grid and wind farms.

[0014] (2) The computational efficiency of this method is comparable to that of the traditional Cholesky method and significantly better than that of the Copula method. It can well adapt to the timeliness requirements of analysis for daily operation scheduling and medium- and long-term planning of wind farms, and reduce the computational cost of wind farm and grid coordinated operation. In addition, this method has strong scalability and can flexibly adapt to the output correlation analysis of various scenarios such as single wind farm multi-unit group, wind farm cluster and wind-solar complementary power station. It provides general technical support for the large-scale development of wind farms and multi-energy coordinated operation, and helps wind farms to integrate more efficiently into high-proportion renewable energy power systems. Attached Figure Description

[0015] Exemplary embodiments of the present invention can be more fully understood by referring to the following figures: Figure 1 The specific flow of the continuous arrangement process provided in the embodiments of the present invention.

[0016] Figure 2 An algorithm for a continuous permutation process is provided in an embodiment of the present invention. Detailed Implementation

[0017] To make the technical problems to be solved, the technical solutions, and the beneficial effects of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present invention and are not intended to limit the present invention.

[0018] 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 invention pertains; the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the invention.

[0019] This invention provides a method for correlation analysis of wind farm output based on continuous sorting, comprising the following steps: Determine the target wind speed correlation matrix, the wind speed sample matrix, and the wind speed sample correlation matrix; Determine the sorting matrix to be used for the replacement, so that the wind speed sample correlation matrix is ​​close to the wind speed target correlation matrix; Based on the correlation matrix of wind speed samples, a relative error matrix is ​​constructed, and the overall error is calculated. Based on the overall error, the wind speed target correlation matrix is ​​adjusted, and the wind farm output correlation analysis results are output according to the converged wind speed target correlation matrix.

[0020] In some embodiments, the specific steps are as follows: Obtain raw wind speed samples from multiple wind farms, determine the target wind speed correlation matrix and the raw wind speed sample matrix, and calculate the initial wind speed sample correlation matrix based on the raw wind speed sample matrix. Eigenvalue decomposition is performed on the wind speed target correlation matrix and the wind speed sample correlation matrix respectively; a ranking matrix is ​​constructed, and the current wind speed sample matrix is ​​permuted based on the ranking matrix to obtain the updated wind speed sample matrix, so that the wind speed sample correlation matrix corresponding to the updated wind speed sample matrix approximates the wind speed target correlation matrix. Calculate the corresponding wind speed sample correlation matrix based on the updated wind speed sample matrix, construct the relative error matrix, and calculate the overall error. If the overall error does not meet the preset convergence condition, the wind speed target correlation matrix is ​​adjusted based on the relative error matrix, and the steps of performing eigenvalue decomposition on the wind speed target correlation matrix and the wind speed sample correlation matrix are iteratively executed until the preset convergence condition is met; if the preset convergence condition is met, the optimized wind speed sample matrix and the corresponding wind farm output correlation analysis results are output.

[0021] In some embodiments, determining the wind speed target correlation matrix, the wind speed sample matrix, and the wind speed sample correlation matrix specifically involves the following steps: Wind speed is typically modeled as a correlated random variable, and its relevant information is described by the wind speed target correlation matrix. As shown in formula (1): (1); In the formula: It is the number of random variables; It is the Pearson correlation coefficient between the first random variable and the second random variable; It is the Pearson correlation coefficient between the first random variable and the d-th random variable; It is the Pearson correlation coefficient between the second random variable and the first random variable; It is the Pearson correlation coefficient between the second random variable and the d-th random variable; It is the Pearson correlation coefficient between the d-th random variable and the 1st random variable; It is the Pearson correlation coefficient between the d-th random variable and the 2nd random variable; Suppose that by performing these random variables Second sampling, generate Wind speed sample matrix As shown in formula (2): (2); Where: the result of each sample The wind speed sample matrix consists of wind speed samples of several random variables. A single line; T is the transpose; It is the wind speed sample vector of the first random variable; It is the wind speed sample vector of the second random variable; It is the wind speed sample vector of the kth random variable; In some embodiments, based on the wind speed sample matrix Calculate the wind speed sample correlation matrix That is, to calculate the wind speed sample matrix The mean and standard deviation of each column of wind speed samples are calculated, and then the covariance of any two columns of wind speed samples is calculated. Based on the covariance and standard deviation, the wind speed correlation coefficient between any two corresponding wind farms is calculated. The wind speed correlation coefficients between all wind farms are arranged according to the row and column order of the wind farms to obtain a d×d dimensional wind speed sample correlation matrix. .

[0022] In some embodiments, feature decomposition is performed on the wind speed target correlation matrix and the wind speed sample correlation matrix, specifically through the following steps: Wind speed target correlation matrix Correlation matrix with wind speed samples The eigenvalues ​​are decomposed separately, as shown in formula (3): (3); In the formula: It is the diagonal matrix of eigenvalues ​​corresponding to the correlation matrix. This represents a matrix composed of eigenvectors; It is the eigenvector matrix of the wind speed sample correlation matrix; It is the diagonal matrix of eigenvalues ​​of the wind speed sample correlation matrix; It is the eigenvector matrix of the wind speed target correlation matrix; It is the diagonal matrix of eigenvalues ​​of the wind speed target correlation matrix.

[0023] In some embodiments, a sorting matrix is ​​constructed, and the current wind speed sample matrix is ​​permuted based on the sorting matrix to obtain an updated wind speed sample matrix. Specifically, the steps are as follows: Construct a sorting matrix As shown in formula (4): (4); According to the sorting matrix For the current wind speed sample matrix Arrange the items; set up Represents the wind speed sample matrix The Middle line, number Column entries, Represents the sorting matrix The Middle line, number Column entries; turn up Then the wind speed sample matrix The Middle line, number The column entries were adjusted to the [number]th [item]. OK; In the formula: Indicates the rank of the item in the column; Obtain the updated wind speed sample matrix .

[0024] In some embodiments, the relative error matrix is ​​constructed and the overall error is calculated, specifically through the following steps: To represent the arrangement deviation, a relative error matrix is ​​established. Relative error matrix Entries As shown in formula (5): (5); In the formula: It is the wind speed sample correlation matrix The Middle Line number Column entries; Correlation matrices with wind speed The Middle Line number Column entries; Calculate the total error As shown in formula (6): (6).

[0025] In some embodiments, the process of adjusting the wind speed target correlation matrix based on the overall error is a process of continuous permutation through an iterative algorithm, i.e., a continuous arrangement process. Based on the overall error, the wind speed target correlation matrix is ​​adjusted. Then, based on the converged wind speed target correlation matrix, the wind farm output correlation analysis results are output. The specific steps are as follows: Based on total error Adjusting the wind speed target correlation matrix Then the adjusted wind speed target correlation matrix Rearrange and set the updated wind speed sample matrix. Through the first This is generated by permutation, i.e. , It is the wind speed sample matrix generated by the (k+1)th permutation; By solving formulas (5) and (6), the relative error matrix of the (k+1)th permutation is determined. The total error of the (k+1)th permutation ; The wind speed correlation matrix for the (k+1)th permutation is updated by adjusting the entries. The update method is shown in formula (7): (7); In the formula: It is the Pearson correlation coefficient between the i-th random variable and the j-th variable in the (k+1)-th permutation; It is the Pearson correlation coefficient between the i-th random variable and the j-th variable in the k-th permutation; These are the entries of the relative error matrix for the (k+1)th permutation; It is the weight, representing the step size, which affects the speed and accuracy of iterative convergence.

[0026] In some embodiments, the specific flow of the continuous arrangement process is as follows: Figure 1 As shown, the algorithm for the continuous permutation process is as follows: Figure 2 As shown; The continuous permutation process includes: (1) Initialization: First, determine the wind speed target correlation matrix, wind speed sample matrix and wind speed sample correlation matrix for the kth iteration, and set the initial wind speed target correlation matrix (i.e. the true correlation benchmark between wind fields). (2) Sample replacement step: Based on the wind speed target correlation matrix and wind speed sample correlation matrix of the kth iteration, the wind speed sample matrix is ​​arranged (in fact, the sorting matrix is ​​constructed by feature decomposition and the element order of the sample column is adjusted) to obtain the updated wind speed sample matrix, and it is assigned the value of the wind speed sample matrix of the k+1th iteration. (3) Error assessment: Calculate the wind speed sample correlation matrix corresponding to the updated wind speed sample matrix (i.e.), and then compare the wind speed target correlation matrix of the (k+1)th iteration with the initial wind speed target correlation matrix using formula (5) to calculate the relative error matrix; (4) Convergence judgment: If the overall error meets the tolerance requirement or the number of iterations reaches the threshold, convergence is determined, and the wind farm output correlation analysis results are directly output. The wind farm output correlation analysis results include the relative error matrix, the final wind speed sample correlation matrix and the optimized wind speed sample matrix, which can be used for wind farm output correlation analysis and power system steady-state calculation. If the overall error does not meet the preset convergence condition, the wind speed target correlation matrix is ​​adjusted based on the relative error matrix, and the wind speed target correlation matrix is ​​updated by formula (7). Then the iteration counter is incremented by 1, and the step of performing feature decomposition on the wind speed target correlation matrix and the wind speed sample correlation matrix is ​​executed iteratively (starting the next round of loop) until the convergence condition is met.

[0027] The continuous permutation process continuously optimizes the correlation of the sample matrix and adjusts the benchmark of the target matrix through iterative cycles, ultimately achieving the analytical goal of minimizing the deviation and closely approximating the real situation in sample correlation. This solves the problems of traditional methods being unable to handle non-positive definite matrices and having large permutation deviations.

[0028] In some embodiments, the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements a wind farm output correlation analysis method based on continuous sorting.

[0029] In some embodiments, the present invention provides an electronic device including at least one processor and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform a wind farm output correlation analysis method based on continuous sorting.

[0030] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0031] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for correlation analysis of wind farm output based on continuous sorting, characterized in that: Includes the following steps: Determine the target wind speed correlation matrix, the wind speed sample matrix, and the wind speed sample correlation matrix; Determine the sorting matrix to be used for the replacement, so that the wind speed sample correlation matrix is ​​close to the wind speed target correlation matrix; Based on the correlation matrix of wind speed samples, a relative error matrix is ​​constructed, and the overall error is calculated. Based on the overall error, the wind speed target correlation matrix is ​​adjusted, and the wind farm output correlation analysis results are output according to the converged wind speed target correlation matrix.

2. The wind farm output correlation analysis method based on continuous sorting as described in claim 1, characterized in that: The specific steps are as follows: Obtain raw wind speed samples from multiple wind farms, determine the target wind speed correlation matrix and the raw wind speed sample matrix, and calculate the initial wind speed sample correlation matrix based on the raw wind speed sample matrix. Eigenvalue decomposition is performed on the wind speed target correlation matrix and the wind speed sample correlation matrix, respectively. Construct a sorting matrix, and perform permutation processing on the current wind speed sample matrix based on the sorting matrix to obtain an updated wind speed sample matrix, so that the wind speed sample correlation matrix corresponding to the updated wind speed sample matrix approximates the wind speed target correlation matrix. Calculate the corresponding wind speed sample correlation matrix based on the updated wind speed sample matrix, construct the relative error matrix, and calculate the overall error. If the overall error does not meet the preset convergence condition, the wind speed target correlation matrix is ​​adjusted based on the relative error matrix, and the steps of performing eigenvalue decomposition on the wind speed target correlation matrix and the wind speed sample correlation matrix are iteratively executed until the preset convergence condition is met; if the preset convergence condition is met, the optimized wind speed sample matrix and the corresponding wind farm output correlation analysis results are output.

3. The wind farm output correlation analysis method based on continuous sorting as described in claim 1, characterized in that: The following steps are used to determine the target wind speed correlation matrix, the sample wind speed matrix, and the sample wind speed correlation matrix: Constructing a wind speed target correlation matrix As shown in formula (1): （1）； In the formula: It is the number of random variables; It is the Pearson correlation coefficient between the first random variable and the second random variable; It is the Pearson correlation coefficient between the first random variable and the d-th random variable; It is the Pearson correlation coefficient between the second random variable and the first random variable; It is the Pearson correlation coefficient between the second random variable and the d-th random variable; It is the Pearson correlation coefficient between the d-th random variable and the 1st random variable; It is the Pearson correlation coefficient between the d-th random variable and the 2nd random variable; Perform random variable Second sampling, generate Wind speed sample matrix As shown in formula (2): （2）； Where: the result of each sample The wind speed sample matrix consists of wind speed samples of several random variables. A single line; T is the transpose; It is the wind speed sample vector of the first random variable; It is the wind speed sample vector of the second random variable; It is the wind speed sample vector of the kth random variable.

4. The wind farm output correlation analysis method based on continuous sorting as described in claim 3, characterized in that: The feature decomposition of the wind speed target correlation matrix and the wind speed sample correlation matrix is ​​performed separately, specifically in the following steps: Wind speed target correlation matrix Correlation matrix with wind speed samples The eigenvalues ​​are decomposed separately, as shown in formula (3): （3）； In the formula: It is the diagonal matrix of eigenvalues ​​corresponding to the correlation matrix. This represents a matrix composed of eigenvectors; It is the eigenvector matrix of the wind speed sample correlation matrix; It is the diagonal matrix of eigenvalues ​​of the wind speed sample correlation matrix; It is the eigenvector matrix of the wind speed target correlation matrix; It is the diagonal matrix of eigenvalues ​​of the wind speed target correlation matrix.

5. The wind farm output correlation analysis method based on continuous sorting as described in claim 4, characterized in that: Construct a sorting matrix, and then permutate the current wind speed sample matrix based on the sorting matrix to obtain an updated wind speed sample matrix. The specific steps are as follows: Construct a sorting matrix As shown in formula (4): （4）； According to the sorting matrix For the current wind speed sample matrix Arrange the items; set up Represents the wind speed sample matrix The Middle line, number Column entries, Represents the sorting matrix The Middle line, number Column entries; turn up Then the wind speed sample matrix The Middle line, number The column entries were adjusted to the [number]th [item]. OK; In the formula: Indicates the rank of an item in the column; Obtain the updated wind speed sample matrix .

6. The wind farm output correlation analysis method based on continuous sorting as described in claim 5, characterized in that: The relative error matrix is ​​constructed, and the overall error is calculated, specifically through the following steps: Establish the relative error matrix Relative error matrix Entries As shown in formula (5): （5）； In the formula: It is the wind speed sample correlation matrix The Middle Line number Column entries; Correlation matrices with wind speed The Middle Line number Column entries; Calculate the total error As shown in formula (6): （6）。 7. The wind farm output correlation analysis method based on continuous sorting as described in claim 6, characterized in that: Based on the overall error, the wind speed target correlation matrix is ​​adjusted. Then, based on the converged wind speed target correlation matrix, the wind farm output correlation analysis results are output. The specific steps are as follows: Based on total error Adjusting the wind speed target correlation matrix Then the adjusted wind speed target correlation matrix Rearrange and set the updated wind speed sample matrix. Through the first This is generated by permutation, i.e. , It is the wind speed sample matrix generated by the (k+1)th permutation; By solving formulas (5) and (6), the relative error matrix of the (k+1)th permutation is determined. The total error of the (k+1)th permutation ; The wind speed correlation matrix for the (k+1)th permutation is updated by adjusting the entries. The update method is shown in formula (7): （7）； In the formula: It is the Pearson correlation coefficient between the i-th random variable and the j-th variable in the (k+1)-th permutation; It is the Pearson correlation coefficient between the i-th random variable and the j-th variable in the k-th permutation; These are the entries of the relative error matrix for the (k+1)th permutation; It is the weight, representing the step size, which affects the speed and accuracy of iterative convergence.

8. A computer-readable storage medium, characterized in that: The device contains a computer program that, when executed by a processor, implements the wind farm output correlation analysis method based on continuous sorting as described in any one of claims 1 to 7.

9. An electronic device, characterized in that: The method includes at least one processor and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the wind farm output correlation analysis method based on continuous sorting as described in any one of claims 1 to 7.

Images