A method and system for determining typical operating modes of an electrical power system

Abstract

The present disclosure provides a power system typical operation mode determination method and system, which comprehensively considers the numerical distance and shape fluctuation characteristics of time series, combines the Euclidean distance and dynamic time warping distance into a Euclidean dynamic warping distance, and uses the same as a similarity algorithm to improve the fuzzy C-means algorithm; the initial clustering center of the improved fuzzy C-means clustering algorithm is obtained by combining the peak density idea and the Euclidean dynamic warping distance for data preprocessing, and then the clustering calculation of the typical day of new energy is realized through the improved fuzzy C-means clustering algorithm, and finally the optimal clustering number and the typical day clustering result are selected through the clustering evaluation index contour coefficient. After selecting the typical day, the power system stability margin is calculated, and the operation mode with the minimum and maximum stability margin is selected as the typical operation mode of the power grid.

Description

TECHNICAL FIELD

[0001] The present disclosure belongs to the technical field of power system analysis, and particularly relates to a power system typical operation mode determination method and system. BACKGROUND

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

[0003] In determining the annual operation mode of the power grid, the maximum or minimum operation mode in summer and winter is usually set. When the penetration rate of new energy is low, the adjustable range of the thermal power unit is relatively large, and the adjustment capacity is strong. Therefore, the influence of new energy output fluctuation on the stability of the power grid can be usually ignored to determine the operation mode of the power grid. However, with the continuous increase of the proportion of new energy, constructing a new type of power system mainly composed of new energy has become a development trend. The uncertainty and volatility of new energy bring greater risks to the safe and stable operation of the power system.

[0004] In order to make the selected operation mode more representative and consider the influence brought by the volatility of new energy, it is particularly important to determine the typical fluctuation of load and new energy output. At present, the clustering method is usually used to determine the typical day. The clustering method represented by fuzzy C-means can obtain the clustering center by analyzing and processing the data. The selected result has a certain representativeness and can preserve the time sequence information of the data. However, the inventors find that the clustering algorithm based on the Euclidean distance can only simply reflect the numerical distance of the time sequence, and cannot reflect the similarity degree of the waveform. Moreover, with the development of communication technology, the time interval of the obtained data is continuously reduced, and the dimension of the data is continuously increased. The higher the dimension of the data, the less the significance of the Euclidean distance calculation. At the same time, the fuzzy C-means clustering algorithm is sensitive to the optimal clustering number and the initial clustering center. The selection of the clustering number and the initial clustering center will directly affect the accuracy of the clustering result. SUMMARY

[0005] In order to solve the above problems, the present disclosure provides a power system typical operation mode determination method and system. The scheme improves the fuzzy C-means algorithm by comprehensively considering the numerical distance and morphological fluctuation characteristics of the time sequence and taking the Euclidean dynamic bending distance as the similarity algorithm. Meanwhile, the initial clustering center of the improved fuzzy C-means algorithm is obtained by combining the peak density idea and the Euclidean dynamic bending distance for data preprocessing. Then, the accurate calculation of the new energy typical day is realized based on the improved fuzzy C-means algorithm, the extraction efficiency and quality of the new energy typical day are improved, and the accurate determination of the typical operation mode of the power grid is realized based on the typical day.

[0006] According to a first aspect of an embodiment of the present disclosure, a power system typical operation mode determination method is provided, comprising:

[0007] Step 1: Obtain a sample data set of new energy output and perform normalization processing;

[0008] Step 2: Based on the sample data set after normalization processing, the cluster number and the initial cluster center, the typical day clustering calculation is performed by using the improved fuzzy C-means algorithm; wherein the similarity calculation in the improved fuzzy C-means algorithm adopts the Euclidean distance and the dynamic time warping distance to construct the Euclidean dynamic warping distance, and the initial cluster center is selected based on the peak density idea and the Euclidean dynamic warping distance; the cluster number in the initial state adopts the preset minimum cluster number;

[0009] Step 3: Calculate the silhouette coefficient of the clustering result, if the current cluster number is not greater than the preset maximum cluster number, then the cluster number is increased by 1, and the step 2 is returned; otherwise, the cluster number with the maximum silhouette coefficient and the cluster center are output as the clustering result of the typical day;

[0010] Step 4: Select the operation mode with the minimum and maximum stability margin as the typical operation mode of the power grid through power system stability margin calculation.

[0011] Further, the Euclidean distance and the dynamic time warping distance are combined to construct the Euclidean dynamic warping distance, which is specifically represented as follows:

[0012]

[0013] Wherein, DTW(L t ,L t’ ) is the dynamic time warping distance; d(L t ,L t’ ) is the Euclidean distance, L t and L t’ represent two samples.

[0014] Further, the initial cluster center is selected based on the peak density idea and the Euclidean dynamic warping distance, which is specifically:

[0015] Step 1: Calculate the Euclidean dynamic warping distance between any two samples in the sample data set;

[0016] Step 2: Calculate the average Euclidean dynamic warping distance of the samples according to the Euclidean dynamic warping distance between the samples;

[0017] Step 3: Calculate the density of each sample based on the average Euclidean dynamic warping distance, and the sample with the maximum density is taken as the first cluster center;

[0018] Step 4: Calculate the relative minimum Euclidean dynamic warping distance of the remaining samples, and calculate the weight product of the samples based on the product of the relative minimum Euclidean dynamic warping distance of the remaining samples and the sample density;

[0019] Step 5: Select the maximum point of the weight value product as the next cluster center;

[0020] Step 6: Repeat steps 4 to 5 until the cluster centers that meet the cluster number requirement are obtained.

[0021] Further, the objective function of the improved fuzzy C-means clustering method is specifically represented as follows:

[0022]

[0023] Wherein, m is a fuzzy weighted index; u tt’ is the membership degree of the sample L t’ calculated; D t is the Euclidean dynamic bending distance between the sample L tt’ and the cluster center L t’ . t

[0024] Further, the membership degree u tt’ satisfies the following constraints:

[0025]

[0026] Wherein, C is the cluster number.

[0027] Further, the calculation of the silhouette coefficient is specifically represented as follows:

[0028]

[0029] Wherein, a i is the average Euclidean dynamic bending distance of sample i to other samples in the class; for the class not containing sample i, the minimum value of the average Euclidean distance of sample i to all samples in the given class is b i ; SE i is the silhouette coefficient of each sample, and the average of all sample SE i is the silhouette coefficient of the clustering result.

[0030] Further, the normalization processing adopts the minimum-maximum standardization method.

[0031] According to a second aspect of the embodiments of the present disclosure, a power system typical operation mode determination system is provided, comprising:

[0032] A data acquisition unit is configured to acquire a sample data set of new energy output and perform normalization processing.

[0033] ​a clustering unit configured to perform typical day clustering calculation by using an improved fuzzy C-means algorithm based on the normalized sample data set, the number of clusters, and the initial cluster centers, wherein similarity calculation in the improved fuzzy C-means algorithm is performed by combining Euclidean distance and dynamic time warping distance to construct Euclidean dynamic warping distance, and the initial cluster centers are selected based on the peak density idea and the Euclidean dynamic warping distance; and the number of clusters in the initial state is a preset minimum number of clusters;

[0034] an optimization solving unit configured to calculate the silhouette coefficient of the clustering result, and if the current number of clusters is not greater than a preset maximum number of clusters, increase the number of clusters by 1 and return to the clustering unit; otherwise, output the number of clusters with the largest silhouette coefficient and the cluster centers as the clustering result of the typical day;

[0035] a typical operation mode acquisition unit configured to select the operation modes with the minimum and maximum stability margins as the typical operation modes of the power grid by power system stability margin calculation.

[0036] According to a third aspect of the embodiments of the present disclosure, an electronic device is provided, which includes a memory, a processor, and a computer program stored in the memory and run on the memory, and the processor implements the power system typical operation mode determination method when executing the program.

[0037] According to a fourth aspect of the embodiments of the present disclosure, a non-transitory computer readable storage medium is provided, which stores a computer program, and the program is executed by a processor to implement the power system typical operation mode determination method.

[0038] Compared with the prior art, the present disclosure has the following beneficial effects:

[0039] (1) The present disclosure provides a power system typical operation mode determination method and system, which improves the extraction efficiency and quality of the typical day of new energy by comprehensively considering the numerical distance and morphological fluctuation characteristics of the time series, proposing Euclidean dynamic warping distance as a similarity algorithm to improve the fuzzy C-means algorithm, combining the peak density idea and the Euclidean dynamic warping distance for data preprocessing to obtain the initial cluster centers of the improved fuzzy C-means algorithm, and then realizing the accurate calculation of the typical day of new energy based on the improved fuzzy C-means algorithm, and realizing the accurate determination of the typical operation mode of the power grid based on the typical day.

[0040] (2) The scheme combines the Euclidean distance and the dynamic time warping distance (DTW) into a Euclidean dynamic warping distance, which is used as a similarity algorithm to improve the fuzzy C-means algorithm, solves the problem that the traditional clustering algorithm based on the Euclidean distance cannot consider the morphological fluctuation characteristics of the time sequence when clustering, directly calculates the initial clustering center by combining the peak density idea and the Euclidean dynamic warping distance, solves the problem that the traditional fuzzy C-means algorithm is sensitive to the selection of the initial clustering center, and finally selects the optimal clustering number K through the clustering evaluation index profile coefficient, solves the problem that the traditional fuzzy C-means algorithm is sensitive to the selection of the clustering number K value. The improved algorithm improves the accuracy of clustering, greatly improves the efficiency and quality of extracting the typical day of new energy. The final clustering result can provide strong data support for researchers of power grid companies and power systems when studying the static voltage stability of the power system, and has good application prospect.

[0041] Advantages of the additional aspects of the present disclosure will be given in part in the following description, some will become apparent from the following description, or will be learned by practice of the present disclosure. BRIEF DESCRIPTION OF DRAWINGS

[0042] The accompanying drawings, which form a part of the present disclosure, are used to provide further understanding of the present disclosure, and the illustrative embodiments of the present disclosure and their description serve to explain the present disclosure. They do not constitute improper limitations on the present disclosure.

[0043] Figure 1 A flow chart of a power system typical operation mode determination method is described in embodiment one of the present disclosure. DETAILED DESCRIPTION

[0044] The present disclosure will be further described below in conjunction with the drawings and embodiments.

[0045] It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the present disclosure. Unless otherwise indicated, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present disclosure belongs.

[0046] It should be noted that the terms used herein are only for the purpose of describing specific embodiments, and are not intended to limit the exemplary embodiments according to the present disclosure. As used herein, the singular form is intended to include the plural form, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprise" and / or "include" are used in the specification, they refer to the presence of a feature, step, operation, device, component, and / or combination thereof.

[0047] The embodiments in the present disclosure and the features in the embodiments can be combined with each other without conflict, if necessary.

[0048] Embodiment one:

[0049] The purpose of this embodiment is to provide a power system typical operation mode determination method.

[0050] As shown in Figure 1 A power system typical operation mode determination method, comprising:

[0051] Step 1: Obtain the sample data set of new energy output and perform normalization processing;

[0052] Step 2: Based on the normalized sample data set, the number of clusters and the initial cluster center, the improved fuzzy C-means algorithm is used to calculate the typical day clustering; wherein the similarity calculation in the improved fuzzy C-means algorithm adopts the combination of Euclidean distance and dynamic time warping distance to construct the Euclidean dynamic warping distance, and selects the initial cluster center based on the peak density idea and the Euclidean dynamic warping distance; the number of clusters in the initial state adopts the preset minimum cluster number;

[0053] Step 3: Calculate the silhouette coefficient of the clustering result, if the current number of clusters is not greater than the preset maximum number of clusters, then add 1 to the number of clusters, and go to step 2; otherwise, output the cluster number and cluster center with the maximum silhouette coefficient as the clustering result of the typical day;

[0054] Step 4: Select the operation mode with the minimum and maximum stability margin as the typical operation mode of the power grid through power system stability margin calculation.

[0055] Further, the combination of Euclidean distance and dynamic time warping distance to construct Euclidean dynamic warping distance is specifically represented as follows:

[0056]

[0057] Where, DTW(L t ,L t’ ) is the dynamic time warping distance; d(L t ,L t’ ) is the Euclidean distance, L t and L t’ represent two samples.

[0058] Further, the selection of the initial cluster center based on the peak density idea and the Euclidean dynamic warping distance is specifically:

[0059] Step 1: Calculate the Euclidean dynamic warping distance between any two samples in the sample data set;

[0060] Step 2: Calculate the average Euclidean dynamic warping distance of the samples according to the Euclidean dynamic warping distance between the samples;

[0061] Step 3: Calculate the density of each sample based on the average Euclidean dynamic bending distance, and take the sample with the highest density as the first cluster center;

[0062] Step 4: Calculate the relative minimum Euclidean dynamic bending distance of the remaining samples, and calculate the weighted product of the samples based on the product of the relative minimum Euclidean dynamic bending distance of the remaining samples and the sample density.

[0063] Step 5: Select the point with the largest weighted product as the next cluster center;

[0064] Step 6: Repeat steps 4 to 5 until cluster centers that meet the cluster number requirement are obtained.

[0065] Furthermore, the objective function of the improved fuzzy C-means clustering method is specifically expressed as follows:

[0066]

[0067] Where m is the fuzzy weighted index; u tt’ It is sample L t’ The calculated cluster centers L t Membership degree; D tt’ For sample L t’ With cluster center L t The European-style dynamic bending distance.

[0068] Furthermore, the membership degree u tt’ The following constraints must be satisfied:

[0069]

[0070] Where C is the number of clusters.

[0071] Furthermore, the calculation of the contour coefficient is specifically expressed as follows:

[0072]

[0073] Among them, a i Let b be the average Euclidean dynamic bending distance from sample i to all other samples within the class; for a class that does not contain sample i, the minimum average Euclidean distance from sample i to all samples in the given class is b. i SE i The silhouette coefficient for each sample, and the SE for all samples. i The mean of the values ​​is the silhouette coefficient of the clustering result.

[0074] Furthermore, the normalization process employs a min-max normalization method.

[0075] Specifically, for ease of understanding, the following is described in detail from the perspective of specific implementation in conjunction with the drawings:

[0076] To solve the problems in the prior art, the embodiment provides a power system typical operation mode determination method. The scheme comprehensively considers the numerical distance and the shape fluctuation characteristics of the time sequence, combines the Euclidean distance and the dynamic time warping distance (DTW) into a Euclidean dynamic warping distance, and uses the Euclidean dynamic warping distance as a similarity algorithm to improve the fuzzy C-means algorithm. The peak density idea and the Euclidean dynamic warping distance are combined to pre-process data, so that the initial clustering center of the improved fuzzy C-means clustering algorithm is obtained, and then the clustering calculation of the typical day of new energy is realized through the improved fuzzy C-means clustering algorithm. Finally, the optimal clustering number and the typical day clustering result are selected through the clustering evaluation index profile coefficient. After the typical day is selected, the power system stability margin is calculated, and the operation mode with the minimum and maximum stability margin is selected as the typical operation mode of the power grid.

[0077] Specifically, as shown in the figure, a power system typical operation mode determination method comprises the following steps: Figure 1

[0078] Step 1: Obtain a sample data set of new energy output, normalize the new energy output data set by using a minimum-maximum standardization method to obtain an input matrix, and set a minimum clustering number K min and a maximum clustering number K max .

[0079] The step 1 is specifically:

[0080] In step 1, the initial sample of new energy is obtained by selecting the new energy output data of a certain region in one year with a granularity of 15 minutes as the initial sample data set.

[0081] In step 1, the minimum-maximum standardization method is used for normalization processing, and the specific formula is as follows:

[0082]

[0083] In the formula, x * , x respectively represent the values before and after normalization, x max , x min represent the maximum and minimum values of the new energy output before normalization; and the obtained input matrix is L t =[L1,L2,…,L n ] T , wherein t=1, 2, 3, …, n, n is the total number of days, and T is a transpose matrix symbol. The data vector L t of each day is composed of 96 new energy output data.

[0084] ​Step 2: combine the Euclidean distance and DTW to build the Euclidean dynamic bending distance as a similarity algorithm to improve the fuzzy C-means clustering algorithm;

[0085] The step 2 is specifically:

[0086] The Euclidean dynamic bending distance in step 2 is calculated by the following steps:

[0087] Step 2-1, input matrix L t in step 1 as a data set D, two sample elements L t =(x1,x2,…,x n ) and L t’ =(y1,y2,…,y n ) are known, and a distance matrix G n*n of n*n is constructed, wherein each element value in the matrix is as follows:

[0088]

[0089] Step 2-2, the set composed of adjacent elements in the matrix is a bending path, and DTW is calculated by the optimal bending path P=(P1,P2,…,P k ) from point (1,1) to point (n,n) in the matrix, so that the cumulative distance of sequences L t and L t’ is minimized, that is,

[0090]

[0091] Step 2-3, the cumulative cost matrix is constructed by dynamic programming method to solve the above formula:

[0092]

[0093] In the formula, ω(0,0)=0; ω(i,0)=ω(0,j)=+∞; the dynamic time bending distance DTW(L t ,L t’ ) of time series L t and L t’ is ω(i,j).

[0094] Step 2-4, combine the Euclidean distance and DTW to build the Euclidean dynamic bending distance, which is calculated by the following formula:

[0095]

[0096] In the formula, DTW(L t ,L t’ ) is the dynamic time bending distance; d(L t ,L t’) is the Euclidean distance, calculated by the following formula:

[0097]

[0098] Step 3: selecting initial clustering centers by combining the peak density idea and the Euclidean dynamic bending distance;

[0099] The step 3 is specifically:

[0100] The step 3 of selecting initial clustering centers by combining the peak density idea and the Euclidean dynamic bending distance, the specific steps are:

[0101] Step 3-1, according to step 2, calculating the Euclidean dynamic bending distance D(L t , L n ) between any two sample elements L t’ = (x1, x2, …, x n ) and L t = (y1, y2, …, y t’ );

[0102] Step 3-2, calculating the average Euclidean dynamic bending distance MeanDis(D) of samples according to the Euclidean dynamic bending distance between samples:

[0103]

[0104] Step 3-3, calculating the density p(t) of each sample according to MeanDis(D):

[0105]

[0106] In the formula p(t) represents the number of samples whose Euclidean dynamic bending distance with sample t is less than MeanDis(D).

[0107] Step 3-4, taking the point with the maximum density as the first clustering center;

[0108] Step 3-5, calculating the relative minimum Euclidean dynamic bending distance S(t) of the remaining samples: if p(t) is the maximum, take max{D(L t , L t' )}, otherwise take min{D(L t , L t' )} (t': p(t') > p(t)):

[0109]

[0110] Defining the product of its and the sample density p(t) of the data set as the weight product:

[0111] w = p(t) * S(t)

[0112] Step 3-6, select other class cluster centers. Select the maximum point of the weight product as the next cluster center.

[0113] Step 3-7, continue to repeat step 3-5, 3-6 until K cluster centers are selected, end the iteration, and get the initial cluster center V = {v1, v2, …, vK}. k}.

[0114] Step 4: Take the input matrix, the number of clusters and the initial cluster center as the input information of the improved fuzzy C-means algorithm to perform typical day clustering calculation;

[0115] The step 4 is specifically:

[0116] The clustering process in step 4 includes the following sub-steps:

[0117] Step 4-1, the input matrix L obtained in step 1, the optimal number of clusters C obtained in step 3 and the initial cluster center V = {v1, v2, …, vK} are input into the improved fuzzy C-means clustering algorithm as initial parameters; t k Step 4-2, set the fuzzy C-means clustering algorithm model objective function as:

[0118] Step 4-2, set the fuzzy C-means clustering algorithm model objective function as:

[0119]

[0120] In the formula, m is the fuzzy weighted index, generally taken as 2; u tt’ is the membership degree of the sample L t’ calculated to the cluster center L t ; D tt’ is the Euclidean dynamic bending distance of the sample L t’ to the cluster center L t calculated according to step 2. Wherein u tt’ satisfies the following constraint condition:

[0121]

[0122] Step 4-3, calculate the membership function according to the following formula with the current cluster center:

[0123]

[0124] In the formula, D kt’ is the Euclidean dynamic bending distance of the sample L k to the cluster center L t calculated according to step 2.

[0125] ​Step 4-4, using the current membership function, calculate the new cluster center according to the following formula:

[0126]

[0127] Step 4-5, repeat step 4-3, 4-4 to update the cluster center and the membership of each sample, so that the objective function in step 4-2 tends to be stable, and when the following convergence criterion is met, the typical day clustering ends:

[0128]

[0129] In the formula, k is the iteration step number, is the sample L t’ The calculated membership of the cluster center L t , the meaning of the above formula is that when the membership function value of each sample is gradually stable and does not change, the clustering is ended.

[0130] Step 5: Calculate the clustering validity index: the silhouette coefficient, if K max , then K=K+1, and go to step 4), otherwise, output the clustering number K with the largest silhouette coefficient and output the cluster center as the clustering result of the typical day.

[0131] The step 5 is specifically:

[0132] The clustering validity index silhouette coefficient in step 5 is calculated as follows:

[0133]

[0134] In the formula, a i is the average Euclidean dynamic bending distance of sample i to other samples in the class; for a class not containing sample i, the minimum value of the average Euclidean distance of sample i to all samples in the given class is b i ; SE i is the silhouette coefficient of each sample, whose value is between [-1, 1], and the closer to 1, the more optimal the intra-class cohesion and inter-class separation; the mean value of all sample SE i is the silhouette coefficient of the clustering result.

[0135] Step 6: Select the operating mode with the smallest and largest stability margin as the typical operating mode of the power grid through power system stability margin calculation.

[0136] Embodiment two:

[0137] The purpose of this embodiment is to provide a power system typical operating mode determination system.

[0138] A power system typical operating mode determination system, comprising:

[0139] a data acquisition unit configured to acquire a sample data set of new energy output and perform normalization processing;

[0140] a clustering unit configured to perform typical day clustering calculation by using an improved fuzzy C-means algorithm based on the normalized sample data set, a clustering number, and an initial clustering center; wherein similarity calculation in the improved fuzzy C-means algorithm is combined to construct an Euclidean dynamic warping distance by using Euclidean distance and dynamic time warping distance, and the initial clustering center is selected based on a peak density idea and the Euclidean dynamic warping distance; the clustering number in an initial state is a preset minimum clustering number;

[0141] an optimization solving unit configured to calculate a silhouette coefficient of a clustering result, if the current clustering number is not greater than a preset maximum clustering number, then the clustering number is increased by 1, and the clustering unit is turned to; otherwise, the clustering number and the clustering center with the largest silhouette coefficient are output as the clustering result of the typical day;

[0142] a typical operation mode acquisition unit configured to select an operation mode with the minimum and maximum stability margin as the typical operation mode of the power grid by power system stability margin calculation.

[0143] Further, the system described in the embodiment corresponds to the method described in Embodiment One, and the technical details have been described in detail in Embodiment One, so here is not repeated.

[0144] In more embodiments, there are also provided:

[0145] An electronic device includes a memory and a processor, and computer instructions stored in the memory and run on the processor, when the computer instructions are run by the processor, the method described in Embodiment One is completed. For the sake of brevity, it is not repeated here.

[0146] It should be understood that in the embodiment, the processor can be a central processing unit CPU, and the processor can also be other general-purpose processors, digital signal processors DSP, application-specific integrated circuits ASIC, ready programmable gate arrays FPGA or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc.

[0147] The memory can include read-only memory and random access memory, and provide instructions and data to the processor, a part of the memory can also include non-volatile random access memory. For example, the memory can also store device type information.

[0148] A computer readable storage medium for storing computer instructions, the computer instructions being executed by a processor to complete the method described in embodiment one.

[0149] The method in embodiment one can be directly embodied as a hardware processor to complete, or be completed by a combination of hardware and software modules in the processor. The software modules can be located in a storage medium in the art, such as random access memory, flash memory, read-only memory, programmable read-only memory, or electrically erasable programmable memory, register, etc. The storage medium is located in the memory, and the processor reads the information in the memory to complete the steps of the above method in combination with the hardware. To avoid repetition, it will not be described in detail here.

[0150] Those skilled in the art can realize that the units and algorithm steps of the examples described in combination with the embodiments can be realized in electronic hardware or in combination of computer software and electronic hardware. Whether the functions are realized in hardware or software mode depends on the specific application and design constraints of the technical solutions. The skilled person can use different methods to realize the described functions for each specific application, but such implementation should not be considered beyond the scope of the present disclosure.

[0151] The power system typical operation mode determination method and system provided by the above embodiments can be realized, and has broad application prospects.

[0152] The above only describes the preferred embodiments of the present disclosure and is not intended to limit the present disclosure. Those skilled in the art can make various modifications and changes to the present disclosure. Any modification, equivalent replacement, improvement, etc. within the spirit and principles of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A method for determining typical operating modes of a power system, characterized in that, include: Step 1: Obtain a sample dataset of new energy output and perform normalization processing; Step 2: Based on the normalized sample dataset, the number of clusters, and the initial cluster centers, perform typical daily clustering calculations using the improved fuzzy C-means algorithm. Specifically, the similarity calculation in the improved fuzzy C-means algorithm combines Euclidean distance and dynamic time curvature distance to construct the Euclidean dynamic curvature distance, and the initial cluster centers are selected based on the peak density concept and the Euclidean dynamic curvature distance. The initial number of clusters is set to a preset minimum number. The Euclidean dynamic bending distance is constructed by combining Euclidean distance and dynamic time bending distance, as specifically represented below: Among them, DTW ( L t , L t’ () represents the dynamic time-bending distance; d ( L t , L t’ () represents the Euclidean distance. L t and L t’ This represents two samples; The selection of initial cluster centers based on the peak density concept and the Euclidean dynamic bending distance is as follows: (1) Calculate the Euclidean dynamic bending distance between any two samples in the sample dataset; (2) Calculate the average Euclidean dynamic bending distance of the samples based on the Euclidean dynamic bending distance between the samples; (3) Calculate the density of each sample based on the average Euclidean dynamic bending distance, and take the sample with the highest density as the first cluster center; (4) Calculate the relative minimum Euclidean dynamic bending distance of the remaining samples, and calculate the weight product of the samples based on the product of the relative minimum Euclidean dynamic bending distance of the remaining samples and the sample density. (5) Select the point with the largest weight product as the next cluster center; (6) Repeat (4) to (5) until cluster centers that meet the cluster number requirement are obtained; Step 3: Calculate the silhouette coefficient of the clustering results. If the current number of clusters is not greater than the preset maximum number of clusters, increment the number of clusters by 1 and go to step 2; otherwise, output the number of clusters with the largest silhouette coefficient and the cluster centers as the clustering results for a typical day. Step 4: Through power system stability margin calculation, select the operating modes with the minimum and maximum stability margins as typical power grid operating modes.

2. The method for determining typical operating modes of a power system as described in claim 1, characterized in that, The objective function of the improved fuzzy C-means clustering method is specifically expressed as follows: Where m is the fuzzy weighted index; u tt’ It is a sample L t’ Calculated cluster centers L t Membership degree; D tt’ For the sample L t’ With cluster center L t The European-style dynamic bending distance.

3. The method for determining typical operating modes of a power system as described in claim 2, characterized in that, membership degree u tt’ The following constraints must be satisfied: in, C This represents the number of clusters.

4. The method for determining typical operating modes of a power system as described in claim 1, characterized in that, The calculation of the profile coefficient is specifically expressed as follows: in, a i For the sample i The average Euclidean dynamic bending distance to other samples within the class; for a class that does not contain sample i, the minimum average Euclidean distance from sample i to all samples in the given class is... b i ; SE i The silhouette coefficient for each sample, and for all samples. SE i The mean of the values ​​is the silhouette coefficient of the clustering result.

5. The method for determining typical operating modes of a power system as described in claim 1, characterized in that, The normalization process employs a min-max normalization method.

6. A system for determining typical operating modes of a power system, characterized in that, include: The data acquisition unit is used to acquire a sample dataset of new energy output and perform normalization processing; A clustering unit is used to perform typical daily clustering calculations based on a normalized sample dataset, the number of clusters, and initial cluster centers using an improved fuzzy C-means algorithm. The similarity calculation in the improved fuzzy C-means algorithm combines Euclidean distance and dynamic time curvature distance to construct a Euclidean dynamic curvature distance, and the initial cluster centers are selected based on the peak density concept and the Euclidean dynamic curvature distance. The initial number of clusters is a preset minimum number. The Euclidean dynamic bending distance is constructed by combining Euclidean distance and dynamic time bending distance, as specifically represented below: Among them, DTW ( L t , L t’ () represents the dynamic time-bending distance; d ( L t , L t’ () represents the Euclidean distance. L t and L t’ This represents two samples; The selection of initial cluster centers based on the peak density concept and the Euclidean dynamic bending distance is as follows: (1) Calculate the Euclidean dynamic bending distance between any two samples in the sample dataset; (2) Calculate the average Euclidean dynamic bending distance of the samples based on the Euclidean dynamic bending distance between the samples; (3) Calculate the density of each sample based on the average Euclidean dynamic bending distance, and take the sample with the highest density as the first cluster center; (4) Calculate the relative minimum Euclidean dynamic bending distance of the remaining samples, and calculate the weight product of the samples based on the product of the relative minimum Euclidean dynamic bending distance of the remaining samples and the sample density. (5) Select the point with the largest weight product as the next cluster center; (6) Repeat (4) to (5) until cluster centers that meet the cluster number requirement are obtained; The optimization solution unit is used to calculate the silhouette coefficient of the clustering result. If the current number of clusters is not greater than the preset maximum number of clusters, the number of clusters is incremented by 1 and the process is transferred to the clustering unit; otherwise, the number of clusters with the largest silhouette coefficient and the cluster center are output as the clustering result of a typical day. The typical operating mode acquisition unit is used to select the operating modes with the minimum and maximum stability margins as the typical operating modes of the power grid through power system stability margin calculation.

7. An electronic device, comprising a memory, a processor, and a computer program stored in the memory and running thereon, characterized in that, When the processor executes the program, it implements a method for determining a typical operating mode of a power system as described in any one of claims 1-5.

8. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements a method for determining a typical operating mode of a power system as described in any one of claims 1-5.

Images