Method for constructing typical wind-solar scenarios with meteorological data based on density peak-FCM

Abstract

The present invention discloses a method for constructing typical wind-solar scenarios containing meteorological data based on density peak-FCM. The method for constructing typical wind-solar scenarios includes the following steps: S1: Obtain the data of historical numerical weather forecasts and the historical annual measured output data of wind farms and photovoltaic power stations, and construct an original data set matrix; S2: Perform dimensionality reduction processing on the original data set matrix constructed in S1, and construct a weather feature matrix after dimensionality reduction; S3: According to the density peak optimized fuzzy C-means clustering algorithm, perform clustering processing on the weather feature matrix after dimensionality reduction obtained in S2; S4: Classify the historical output data of wind farms and photovoltaic power stations according to the results of weather pattern recognition in S3, and then fit to obtain typical daily output scenarios of wind and light under different weather patterns. This method can not only retain the original data features to the greatest extent by performing dimensionality reduction processing on various meteorological feature data, but also avoid data redundancy and improve calculation efficiency.

Description

Technical Field

[0001] The present invention relates to a method for generating typical scenarios based on clustering, specifically a method for constructing typical scenarios of wind and light with meteorological data based on density peak - FCM. Background Art

[0002] In the context of the era of achieving the "dual - carbon" goal, the new energy in our country has developed rapidly and achieved remarkable results. Among new energies, wind energy and solar energy have the highest usage ratio and are the most widely studied. However, due to the limitation of natural resources such as wind and light by climate change, wind power generation and photovoltaic power generation have obvious volatility, uncertainty, and reverse peak - shaving characteristics. Therefore, the large - scale grid connection of wind power and photovoltaic power will bring great challenges to the safe and stable operation of the power grid, and the phenomena of curtailment of wind and light are becoming increasingly prominent. Therefore, it is necessary to make full use of the complementary characteristics of wind power generation and photovoltaic power generation output, study the output laws of wind power and photovoltaic power generation in combination with medium - and long - term meteorological data, and consider generating more accurate and representative typical output scenarios of wind and light new energies to cope with their volatility and uncertainty and meet the flexibility requirements of power system dispatching.

[0003] Domestic and foreign scholars have conducted a large number of studies on the construction of typical output scenarios for clean energy. Regarding the classification of output scenario types, it is mainly to directly divide typical scenarios according to the seasons of each month. However, dividing according to seasons does not fully consider climate factors. Especially in summer and winter, the weather types are changeable. For example, in summer, extremely opposite extreme weathers such as high - temperature and sunny days and heavy rains are likely to switch randomly. Therefore, dividing output scenarios according to seasons will result in scenarios that cannot reflect the sensitivity of wind power generation and photovoltaic power generation to many meteorological factors, and the scenarios are not accurate and representative enough.

[0004] Regarding the generation method of typical scenario output curves, there are mainly time - series simulation methods and clustering methods. When calculating and analyzing on an annual or monthly basis, the time - series simulation method will lead to data redundancy, slow calculation speed, and reduced accuracy due to the large input time series and many time sections. And traditional K - means and FCM clustering algorithms are sensitive to the selection of initial clustering centers and the number of clusters. Unreasonable selection will lead to the clustering results falling into local optimal solutions.

[0005] Based on this, it is necessary to propose a reasonable scenario construction method so that the generated typical scenarios of wind and light have high representativeness, can reflect the output laws of new energies and their sensitivity to meteorological factors, and enable the typical scenarios to be well applied to the day - ahead planning and dispatching of the power grid, reducing the negative impacts brought by the high - proportion grid connection of new energies to the power grid. Summary of the Invention

[0006] The object of the present invention is to provide a method for constructing typical wind-solar scenarios with meteorological data based on density peak-FCM. Considering that within a certain geographical range, the output of wind power and photovoltaic power is subject to certain regular characteristics under the influence of the natural environment, it is necessary to construct realistic typical output scenarios of wind and light to meet the requirements of application scenarios such as day-ahead scheduling. This method performs dimensionality reduction on a variety of meteorological characteristic data, which can not only retain the original data characteristics to the greatest extent, but also avoid data redundancy and improve the calculation efficiency. Compared with the traditional method of dividing typical days according to seasons, considering the changeable weather conditions within seasons, the classification of typical day scenarios is more accurate and reasonable. At the same time, this method uses the density peak algorithm to preprocess the data after dimensionality reduction, and determines the number of clustering centers in advance, solving the problem that the FCM algorithm is sensitive to the setting of the number of clustering centers, and effectively improving the clustering accuracy and the representativeness of typical wind-solar output scenarios.

[0007] The object of the present invention can be achieved by the following technical solutions:

[0008] A method for constructing typical wind-solar scenarios with meteorological data based on density peak-FCM, the method for constructing typical wind-solar scenarios includes the following steps:

[0009] S1: Obtain the data of historical numerical weather forecasts and the historical annual measured output data of wind farms and photovoltaic power plants, and construct an original data set matrix;

[0010] S2: Use the principal component analysis method to perform dimensionality reduction on the original data set matrix constructed in S1, select the climate characteristic index that has the greatest influence on the output of the wind farm and several climate characteristic indexes that have the greatest influence on the output of the photovoltaic power plant respectively, and construct a weather feature matrix after dimensionality reduction;

[0011] S3: According to the density peak optimized fuzzy C-means clustering algorithm, perform clustering on the weather feature matrix after dimensionality reduction obtained in S2, that is, classify the weather events within a year;

[0012] S4: Classify the historical daily output data of the wind farm and the photovoltaic power plant according to the results of weather pattern recognition in S3, and then fit to obtain typical daily output scenarios of wind and light under different weather patterns.

[0013] Further, obtaining the data of historical numerical weather forecasts and the historical annual measured output data of wind farms and photovoltaic power plants in step S1 constitutes a data set matrix;

[0014] S1.1: Assume that there are sampling points, select a total of characteristic indexes such as the output of the wind farm, wind speed, wind direction, air temperature, air pressure, humidity, sunshine hours, precipitation, etc., and the data set matrix of the wind farm is:

[0015] (1)

[0016] Wherein, represents the number of samplings, represents the feature dimension, represents the th sampling point and the characteristic value corresponding to the th wind farm characteristic index;

[0017] S1.2: At the same sampling points, select the photovoltaic power station, as well as a total of characteristic indexes such as wind speed, wind direction, temperature, air pressure, humidity, sunshine hours, precipitation, etc. The data set matrix of the photovoltaic power station is:

[0018] (2)

[0019] Wherein, represents the number of samplings, represents the feature dimension, represents the th sampling point and the characteristic value corresponding to the th photovoltaic power station characteristic index.

[0020] Furthermore, it is characterized in that the unit of the characteristic index in the data set matrix of the wind farm and the photovoltaic power station is different, so standardization processing is performed on it:

[0021] (3)

[0022] Wherein, represents the th sampling data of the feature , represents the mean value of the variable, represents the variance of the variable.

[0023] Furthermore, in step S2, the principal component analysis method is used to perform dimensionality reduction processing on the data set matrices W and Q of the wind farm and the photovoltaic power station in S1, and the weather characteristic indexes that have the most significant influence on the output of the wind farm and the photovoltaic power station are respectively selected, including the following steps;

[0024] S2.1: Calculate the covariance matrix :

[0025] (4)

[0026] In the matrix, , , Reflects the meteorological characteristic indicators And the index The degree of correlation;

[0027] S2.2: Calculate eigenvalues and eigenvectors:

[0028] (5)

[0029] Calculate the covariance matrix according to Equation (5) The eigenvalues of The corresponding eigenvectors are , ,…, And the dataset matrix after dimensionality reduction can be obtained as:

[0030] (6)

[0031] In the formula, Is the first principal component, Is the second principal component, Is the Principal component;

[0032] S2.3: The eigenvalues are , ,…, ( ) The cumulative contribution rate is:

[0033] (7)

[0034] When the cumulative contribution rate Is greater than 0.9, that is, select the first Eigenvectors , ,…, As Principal components, instead of the original Index variables for subsequent calculations.

[0035] Furthermore, after the dataset matrices W and Q of the wind farm and the photovoltaic power station are processed by the principal component analysis method, the dimensionality-reduced weather characteristic matrices obtained are respectively:

[0036] (8)

[0037] (9)

[0038] Furthermore, in the step S3: Cluster preprocess the weather events in a year according to the density peak algorithm, including the following steps;

[0039] S3.1: First, for the weather feature matrices of the wind farm and the PV power station, calculate the distance matrix between each sampling data point , and then calculate the local density of each sampling data point and the distance to the high-density points:

[0040] (10)

[0041] (11)

[0042] In the formula, is the local density of the data point, where , is the intercept; is the minimum distance between the data point and the point with a larger density;

[0043] S3.2: Use the local density as the horizontal axis and the distance to the high-density point as the vertical axis to construct a decision graph; select the points where both and are large as the clustering centers. Let the number of clustering centers obtained by the DPCA algorithm be , and the clustering centers be .

[0044] Furthermore, in step S4, according to the number of clustering centers and the positions of the clustering centers obtained by preprocessing the data using the density peak algorithm in step S3, the fuzzy C-means algorithm is used to cluster the historical daily output data of the wind farm and the PV power station respectively, and then different wind and PV typical daily output scenarios under different weather patterns are obtained by permutation and combination;

[0045] S4.1: Initialize the parameters. According to the result of step S3.2, set the initial number of clustering centers to c, set the fuzzy weight index , the number of iterations and the iteration termination condition ;

[0046] S4.2: Initialize the membership matrix, and the membership satisfies the normalization condition:

[0047] (12)

[0048] S4.3: Calculate the new clustering centers:

[0049] (13)

[0050] S4.4: Update the membership matrix:

[0051] (14)

[0052] S4.5: Calculate the objective function:

[0053] (15)

[0054] When the objective function or the number of iterations reaches the maximum, stop the iteration and output the clustering result.

[0055] Furthermore, the clustering result obtained from the wind farm output data is , and the clustering result obtained from the photovoltaic power plant output data is . Based on this permutation and combination, sets of typical scenarios of grid wind power and photovoltaic output under various weather patterns are obtained.

[0056] Advantages of the present invention:

[0057] The method for constructing typical scenarios of wind and light in the present invention takes into account that within a certain geographical range, affected by the natural environment, the output of wind power and photovoltaic power has certain regular characteristics, and it is necessary to construct realistic typical output scenarios of wind and light to meet the requirements of application scenarios such as day-ahead scheduling. This method performs dimensionality reduction processing on various meteorological characteristic data, which can not only retain the original data characteristics to the greatest extent, but also avoid data redundancy and improve the calculation efficiency; compared with the traditional method of dividing typical days according to seasons, considering the changeable weather conditions within seasons, the classification of typical day scenarios is more accurate and reasonable. At the same time, this method uses the density peak algorithm to preprocess the dimensionality-reduced data, determines the number of clustering centers in advance, solves the problem that the FCM algorithm is sensitive to the setting of the number of clustering centers, and effectively improves the clustering accuracy and the representativeness of typical output scenarios of wind and light. Description of the drawings

[0058] The following further describes the present invention with reference to the accompanying drawings.

[0059] Figure 1 is the implementation flowchart of the method for constructing typical scenarios of wind and light in the present invention;

[0060] Figure 2 is the decision diagram obtained by processing the wind farm output data with the density peak algorithm in the present invention;

[0061] Figure 3 is the two-dimensional classification diagram obtained by processing the wind farm output data with the density peak algorithm in the present invention;

[0062] Figure 4 is the decision diagram obtained by processing the photovoltaic power plant output data with the density peak algorithm in the present invention;

[0063] Figure 5It is a two-dimensional classification map obtained by processing the output data of the photovoltaic power station of the present invention through the density peak algorithm;

[0064] Figure 6 It is a three-dimensional classification map obtained by processing the output data of the wind farm of the present invention through the fuzzy C-means algorithm;

[0065] Figure 7 It is a three-dimensional classification map obtained by processing the output data of the photovoltaic power station of the present invention through the fuzzy C-means algorithm;

[0066] Figure 8 It is the typical curve of the high output type of the wind farm of the present invention;

[0067] Figure 9 It is the typical curve of the low output type of the wind farm of the present invention;

[0068] Figure 10 It is the typical curve of the high output type of the photovoltaic power station of the present invention;

[0069] Figure 11 It is the typical curve of the medium output type of the photovoltaic power station of the present invention;

[0070] Figure 12 It is the typical curve of the low output type of the photovoltaic power station of the present invention. Specific embodiments

[0071] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without making creative efforts belong to the scope of protection of the present invention.

[0072] The present invention discloses a method for constructing typical scenarios of wind and light containing meteorological data based on density peak - FCM. In this specific implementation, a wind farm and a photovoltaic power station in a certain area are selected as examples, and the hourly power generation data of the wind farm and the photovoltaic power station in one year are obtained. And the meteorological data of this area is obtained, and the categories include: small evaporation (mm), large evaporation (mm), average surface temperature (°C), daily maximum surface temperature (°C), daily minimum surface temperature (°C), cumulative precipitation from 20 to 20 (mm), average station pressure (hPa), average relative humidity (1%), sunshine hours (h), average temperature (°C), average wind speed (m / s), maximum wind speed (m / s), wind direction of the maximum wind speed (16 directions).

[0073] The method for constructing typical output scenarios of wind and light containing meteorological information based on density peak - FCM proposed by the present invention, as Figure 1 shown, includes the following steps:

[0074] S1: Obtain the data of historical numerical weather forecasts and the historical annual measured output data of wind farms and photovoltaic power stations, and construct the original dataset matrix, including the following steps:

[0075] S1.1: Assume that there are sampling points, and select a total of characteristic indicators such as wind farm output, wind speed, wind direction, temperature, air pressure, humidity, sunshine hours, precipitation, etc. The dataset matrix of the wind farm is:

[0076] (1)

[0077] In the formula, represents the number of samplings, represents the feature dimension, represents the characteristic value corresponding to the th sampling point and the th wind farm characteristic indicator.

[0078] S1.2: At the same sampling points, select a total of characteristic indicators such as photovoltaic power station output, wind speed, wind direction, temperature, air pressure, humidity, sunshine hours, precipitation, etc. The dataset matrix of the photovoltaic power station is:

[0079] (2)

[0080] In the formula, represents the number of samplings, represents the feature dimension, represents the characteristic value corresponding to the th sampling point and the th photovoltaic power station characteristic indicator.

[0081] S2: Use the principal component analysis method to perform dimensionality reduction on the original dataset matrix constructed in S1, select the climate characteristic indicators that have the greatest impact on wind farm output and several climate characteristic indicators that have the greatest impact on photovoltaic power station output respectively, and construct the weather feature matrix after dimensionality reduction;

[0082] The dimensionality reduction process includes the following steps:

[0083] S2.1: Since the units of the characteristic indicators in the dataset matrices of wind farms and photovoltaic power stations are different, standardize them:

[0084] (3)

[0085] In the formula, The th sampling data, represents the mean of the variable, represents the variance of the variable.

[0086] S2.2: Calculate the covariance matrix :

[0087] (4)

[0088] In the matrix, , , reflects the correlation degree between the meteorological characteristic index and the index .

[0089] S2.3: Calculate the eigenvalues and eigenvectors:

[0090] (5)

[0091] Calculate the covariance matrix obtained according to formula (5), the eigenvalues are , , …, , and the reduced-dimension dataset matrix can be obtained as:

[0092] (6)

[0093] In the formula, is the first principal component, is the second principal component, is the th principal component.

[0094] S2.4: The eigenvalues are , , …, ( ) and the cumulative contribution rate is:

[0095] (7)

[0096] When the cumulative contribution rate is greater than 0.9, that is, select the first eigenvectors , , …, as the principal components to replace the original Subsequent calculations are performed on the indicator variables. Let the weather feature matrices obtained by the wind farm and the PV power station through principal component analysis be and .

[0097] S3: According to the density peak-optimized fuzzy C-means clustering algorithm, perform clustering on the dimensionality-reduced weather feature matrix obtained in S2, that is, classify the weather events within one year;

[0098] Perform clustering on the dimensionality-reduced weather feature matrix: The steps to obtain the number of cluster centers using the density peak algorithm include the following:

[0099] S3.1: First, for the weather feature matrices of the wind farm and the PV power station, calculate the distance matrix between each sampling data point , and then calculate the local density sum and the distance to the high-density points of each sampling data point :

[0100] (8)

[0101] (9)

[0102] In the formula, is the local density of the data point, where , is the intercept. is the minimum distance between the data point and the point with a larger density.

[0103] S3.2: Use the local density as the horizontal axis and the distance to the high-density points as the vertical axis to construct a decision graph. Select points where both and are large as the cluster centers according to the decision graph. Let the number of cluster centers obtained by the DPCA algorithm be , and the cluster centers be .

[0104] First, standardize the historical output data and each meteorological data, and then perform dimensionality reduction on the data using the principal component analysis method. The dimensionality-reduced data is pre-clustered using the density peak algorithm, and the results are as shown in Figure 2 , Figure 3 , Figure 4 and Figure 5 . In the density peak algorithm, points with relatively large local density and far distance from other high-density points are defined as cluster centers. According to the definition and the decision graph, it can be clearly seen that the output data of this wind farm should be divided into two categories, and the output data of the PV power station should be divided into three categories most reasonably.

[0105] S4: Classify the historical output data of wind farms and PV power plants according to the results of weather pattern recognition in S3, and then fit to obtain the typical daily output scenarios of wind and PV under different weather patterns.

[0106] Use the fuzzy C-means algorithm to perform clustering processing on the historical output data of wind farms and PV power plants respectively.

[0107] S4.1: Initialize the parameters. According to the results of step S3.2, set the initial number of clustering centers to c, set the fuzzy weight index , the number of iterations and the iteration termination condition .

[0108] S4.2: Initialize the membership matrix, and the membership satisfies the normalization condition:

[0109] (10)

[0110] S4.3: Calculate the new clustering centers:

[0111] (11)

[0112] S4.4: Update the membership matrix:

[0113] (12)

[0114] S4.5: Calculate the objective function:

[0115] (13)

[0116] When the objective function or the number of iterations reaches the maximum, stop the iteration and output the clustering results.

[0117] S4.5: Let the clustering result obtained from the output data of the wind farm be , and the clustering result obtained from the output data of the PV power plant be , and successively obtain sets of typical scenarios of grid-connected wind power and PV power output under different weather patterns.

[0118] After obtaining the number of clustering centers according to the density peak algorithm, then use the FCM clustering method to process the output data. When clustering the output data of the wind farm, set the number of clustering centers to 2; when clustering the output data of the PV power plant, set the number of clustering centers to 3. The three-dimensional classification diagrams obtained by processing the output data of the wind farm and PV power plant through the fuzzy C-means algorithm are shown in Figure 6 and Figure 7As shown. The pentagram-shaped points indicate the locations of the clustering centers. Based on this, the typical output curves of the wind farm and the PV power station can be obtained. The typical daily output curve types of the wind farm can be divided into two categories, namely the high wind power output type and the low wind power output type, as Figure 8 , Figure 9 shown. The typical daily output curve of the PV power station can be divided into three categories, namely the high PV power output type, the low PV power output type, and the medium PV power output type, as Figure 10 , Figure 11 and Figure 12 shown.

[0119] In the description of this specification, the descriptions referring to terms such as "one embodiment", "example", "specific example", etc. mean that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in a suitable manner in any one or more embodiments or examples.

[0120] The above shows and describes the basic principles, main features, and advantages of the present invention. Those skilled in the art of this industry should understand that the present invention is not limited by the above embodiments. The above embodiments and the descriptions in the specification only illustrate the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed.

Claims

1. A method for constructing typical wind-solar scenarios with meteorological data based on density peak-FCM, characterized in that, The construction method of the typical wind-solar scene includes the following steps: S1: Obtain the data of historical numerical weather forecasts and the historical annual measured output data of the wind farm and the photovoltaic power station, and construct the original dataset matrix; S2: Use the principal component analysis method to perform dimensionality reduction on the original dataset matrix constructed in S1, respectively select the climate characteristic indicators that have the greatest impact on the output of the wind farm and several climate characteristic indicators that have the greatest impact on the output of the photovoltaic power station, and construct the weather feature matrix after dimensionality reduction; S3: According to the fuzzy C-means clustering algorithm optimized by density peaks, perform clustering on the weather feature matrix after dimensionality reduction obtained in S2, specifically as follows: S3.1: First, for the weather feature matrix after dimensionality reduction of the wind farm and the PV power station, calculate the distance matrix between each sampling data point , and then calculate the local density of each sampling data point and the distance to the high-density points: （10） （11） In the formula, is the local density of the data point, where , is the intercept; is the minimum distance between the data point and the point with a larger density; S3.2: Using the local density as the horizontal axis and the distance to the high-density points as the vertical axis, construct a decision graph; select points where both and are large as the clustering centers. Let the number of clustering centers obtained by the DPCA algorithm be , and the clustering centers be ; S4: Classify the historical output data of the wind farm and the photovoltaic power station according to the results of weather pattern recognition in S3, and then fit to obtain the typical daily output scenarios of wind and light under different weather patterns.

2. The method for constructing typical wind-solar scenarios with meteorological data based on density peak-FCM according to claim 1, characterized in that In step S1, obtaining the data of historical numerical weather forecasts and the historical annual measured output data of the wind farm and the photovoltaic power station constitutes the dataset matrix; S1.1: Assume that there are a total of sampling points, and characteristic indicators of the wind farm are selected. The dataset matrix of the wind farm is: （1） In the formula, represents the number of samplings, represents the feature dimension, denotes the th sampling point and the characteristic value corresponding to the th wind farm characteristic index; S1.2: At the same sampling points, select characteristic indicators of the PV power station. The dataset matrix of the PV power station is as follows: （2） Wherein, represents the number of samplings, represents the feature dimension, denotes the th sampling point and the characteristic value corresponding to the th characteristic index of the PV power station.

3. The method for constructing typical wind-solar scenarios with meteorological data based on density peak-FCM according to claim 2, characterized in that, The unit of the characteristic indicators in the dataset matrix of the wind farm and the photovoltaic power station is different, so standardize it: （3） In the formula, represents the th sampling data of the feature represents the mean value of the variable, represents the variance of the variable.

4. The method for constructing typical wind-solar scenarios with meteorological data based on density peak-FCM according to claim 3, characterized in that In step S2, use the principal component analysis method to perform dimensionality reduction on the dataset matrices W and Q of the wind farm and the photovoltaic power station in S1, and respectively select the weather characteristic indicators that have the most significant impact on the output of the wind farm and the photovoltaic power station, including the following steps; S2.1: Calculate the covariance matrix : （4） In the matrix, , , reflects the correlation degree of the meteorological characteristic index and the index ; S2.2: Calculate the eigenvalues and eigenvectors: （5） The covariance matrix is calculated according to Equation (5). The eigenvalues are such that the corresponding eigenvectors are , , …, , and the dataset matrix after dimensionality reduction is obtained as follows: （6） In the formula, is the first principal component, is the second principal component, is the principal component; S2.3: The cumulative contribution rate of eigenvalues , , …, ( ) is: （7） When the cumulative contribution rate is greater than 0.9, the first eigenvectors , , …, are selected as principal components to replace the original indicator variables for subsequent calculations.

5. The method for constructing typical wind-solar scenarios with meteorological data based on density peak-FCM according to claim 4, characterized in that After the dataset matrices W and Q of the wind farm and the photovoltaic power station are processed by the principal component analysis method, the obtained weather feature matrices after dimensionality reduction are respectively: （8） （9） 6. The method for constructing typical wind-solar scenarios with meteorological data based on density peak-FCM according to claim 5, characterized in that In step S4, according to the number and positions of the clustering centers obtained by preprocessing the data using the density peak algorithm in step S3, then use the fuzzy C-means algorithm to perform clustering on the historical output data of the wind farm and the photovoltaic power station respectively, and then arrange and combine to obtain the typical daily output scenarios of wind and light under different weather patterns; S4.1: Initialize parameters. According to the result of step S3.2, the initial number of cluster centers is set to c, and the fuzzy weight index , the number of iterations and the iteration termination condition are set; S4.2: Initialize the membership matrix, and the membership satisfies the normalization condition: （12） S4.3: Calculate the new clustering centers: （13） S4.4: Update the membership matrix: （14） S4.5: Calculate the objective function: （15） When the objective function or the number of iterations reaches the maximum, stop the iteration and output the clustering result.

7. The method for constructing typical wind-solar scenarios with meteorological data based on density peak-FCM according to claim 6, wherein Let the clustering result obtained from the wind farm output data be , and the clustering result obtained from the photovoltaic power plant output data be . Based on this permutation and combination, sets of typical scenarios of grid wind power and photovoltaic output under weather patterns are obtained.

Images