Novel power distribution network cluster division method, system and device and storage medium

Abstract

The invention relates to the technical field of power distribution network control, in particular to a novel power distribution network cluster division method, system and device and a storage medium, and the method comprises the steps: dividing a time section of a power distribution network into a plurality of time intervals; in each time interval, calculating active power flow data of each node in the power distribution network, and performing weighted summation on the active power flow data of each time interval to obtain an active power flow comprehensive index; according to the voltage data of each node of the power distribution network under the multiple time sections, calculating to obtain a voltage cosine similarity and a voltage Euclidean distance; obtaining a Laplacian matrix according to the active power flow comprehensive index, the voltage cosine similarity and the voltage Euclidean distance; and obtaining a distribution network node cluster division result by using a K-means clustering algorithm according to the cluster division number and the Laplacian matrix. According to the method, the spectral clustering algorithm and the like are adopted for cluster division, dynamic cluster division of the power distribution network is achieved, the accuracy and stability of cluster division are improved, and reasonable resource allocation of the power distribution network is ensured.

Description

Technical Field

[0001] The present invention relates to the field of distribution network control technology, and in particular to a novel distribution network cluster division method, system, device and storage medium. Background Art

[0002] In the power system, the distribution network is the terminal link of power transmission and distribution. Its operating efficiency and stability are directly related to the user's power quality and the overall reliability of the power system. With the acceleration of urbanization and the continuous growth of electricity demand, the scale of the distribution network is expanding and the structure is becoming more complex, which puts higher requirements on the management and optimization of the distribution network. The traditional distribution network management method is often based on experience judgment or simple data statistics, which is difficult to fully and accurately reflect the actual operating status of the distribution network, resulting in unreasonable resource allocation, slow fault response and other problems.

[0003] In the prior art, the clustering of distribution networks mainly relies on static network structure information and limited operating data. This method ignores the dynamic change characteristics of the distribution network in the time dimension, such as the periodic fluctuation of the load, the uncertainty of the power output, etc., resulting in a deviation between the clustering results and the actual operating status, and cannot provide strong support for the optimized operation of the distribution network. In addition, most of the clustering methods in the prior art use relatively simple clustering algorithms. When these algorithms process large-scale, high-dimensional data, the stability and accuracy of the clustering results are not high, resulting in the inability to achieve accurate clustering results.

[0004] In summary, the existing technology has many shortcomings in the distribution network cluster division, and there is an urgent need for a method that can realize the refined and dynamic cluster division of the distribution network to improve the operation efficiency and stability of the distribution network. Summary of the invention

[0005] In order to solve the above technical problems, the present invention provides a novel distribution network cluster division method, system, device and storage medium.

[0006] In a first aspect, the present invention provides a novel method for dividing distribution network clusters, the method comprising the following steps:

[0007] According to the electrical characteristic parameters and operation data of the distribution network, the time section of the distribution network is divided into multiple time intervals;

[0008] In each time interval, the active power flow data of each node in the distribution network is calculated, and the active power flow data of each time interval is weighted and summed to obtain the active power flow comprehensive index;

[0009] According to the voltage data of each node in the distribution network under multiple time sections, the voltage cosine similarity and voltage Euclidean distance between different nodes in the distribution network are calculated;

[0010] According to the comprehensive index of active power flow, voltage cosine similarity and voltage Euclidean distance of distribution network in multiple time sections, the Laplace matrix is ​​constructed.

[0011] According to the predetermined number of cluster divisions and the Laplace matrix, a K-means clustering algorithm is used to perform clustering iterations to obtain a distribution network node cluster division result.

[0012] In a further embodiment, the active power flow data includes an average active power flow expected value and an active power flow variance; the step of calculating the active power flow data of each node in the distribution network in each time interval includes:

[0013] Calculate the ratio between the sum of the time section flow values ​​between different nodes of the distribution network in all time intervals and the number of time sections to obtain the expected value of the average active power flow;

[0014] Calculate the sum of squares of the differences between the power flow values ​​between different nodes of the distribution network and the expected value of the average active power flow in each time interval to obtain the active power flow fluctuation energy;

[0015] According to the number of time sections in each time interval and the active power flow fluctuation energy, the active power flow variance between different nodes of the distribution network in each time interval is obtained.

[0016] In a further implementation scheme, the step of performing weighted summation of active power flow data of each time interval to obtain an active power flow comprehensive index comprises:

[0017] According to the active power flow variance between different nodes of the distribution network in each time interval and the preset time interval weight coefficient, a weighted value of the active power flow variance is obtained;

[0018] Calculating the active power flow segmentation index in each time interval according to the active power flow variance weighted value and the average active power flow expected value;

[0019] The active power flow segment indicators of all time intervals are summed up to obtain the active power flow comprehensive indicator.

[0020] In a further implementation scheme, the step of constructing a Laplace matrix according to the comprehensive index of active power flow, voltage cosine similarity and voltage Euclidean distance of the distribution network under multiple time sections includes:

[0021] The comprehensive active power flow index, voltage cosine similarity and voltage Euclidean distance of the distribution network under multiple time sections are indexed respectively to obtain the corresponding indexed comprehensive active power flow index, indexed voltage cosine similarity and indexed voltage Euclidean distance;

[0022] Normalizing the indexed active power flow comprehensive index, the indexed voltage cosine similarity and the indexed voltage Euclidean distance respectively to obtain corresponding normalized active power flow comprehensive index, normalized voltage cosine similarity and normalized voltage Euclidean distance;

[0023] Performing weighted summation on the normalized active power flow comprehensive index, the normalized voltage cosine similarity and the normalized voltage Euclidean distance to construct a spectral clustering similarity matrix;

[0024] A degree matrix is ​​calculated according to the spectral clustering similarity matrix, and a Laplace matrix is ​​calculated by the difference between the degree matrix and the spectral clustering similarity matrix.

[0025] In a further embodiment, the step of performing clustering iteration using a K-means clustering algorithm according to a predetermined number of cluster divisions and the Laplace matrix to obtain a distribution network node cluster division result comprises:

[0026] According to the predetermined number of cluster divisions, several minimum eigenvalues ​​of the Laplacian matrix and their corresponding eigenvectors are calculated;

[0027] A normalized feature matrix is ​​constructed based on the feature vectors, and a preferred clustering center is determined based on the density distribution and distance characteristics of the feature data points in the normalized feature matrix;

[0028] The preferred cluster center is used as the initial cluster center of K-means clustering, and K-means clustering is used to perform clustering iteration on the normalized feature matrix to obtain a distribution network node cluster division result.

[0029] In a further embodiment, the step of constructing a normalized feature matrix according to the feature vector comprises:

[0030] All the eigenvectors are arranged in columns to form a eigenmatrix, and the eigenmatrix is ​​normalized in rows to obtain a normalized eigenmatrix.

[0031] In a further embodiment, the step of determining the preferred cluster center according to the density distribution and distance characteristics of the feature data points in the normalized feature matrix comprises:

[0032] Traversing each feature data point in the normalized feature matrix, and obtaining the density distribution of the feature data point according to the average distance from each feature data point to its nearest plurality of neighboring points;

[0033] Initialize the center point set and calculate the minimum distance from each non-center point feature data point to the center point set;

[0034] According to the density distribution and minimum distance of the feature data points, a comprehensive selection index is obtained;

[0035] The characteristic data point with the largest comprehensive selection index is taken as the next center point until the preferred cluster centers equal to the number of cluster divisions are screened out.

[0036] In a second aspect, the present invention provides a novel distribution network cluster division system, the system comprising:

[0037] An interval division module is used to divide the time section of the distribution network into multiple time intervals according to the electrical characteristic parameters and operation data of the distribution network;

[0038] The power flow analysis module is used to calculate the active power flow data of each node in the distribution network in each time interval, and perform weighted summation of the active power flow data of each time interval to obtain the active power flow comprehensive index;

[0039] The voltage analysis module is used to calculate the voltage cosine similarity and voltage Euclidean distance between different nodes of the distribution network based on the voltage data of each node in the distribution network under multiple time sections;

[0040] A matrix construction module is used to construct a Laplace matrix according to the comprehensive active power flow index, voltage cosine similarity and voltage Euclidean distance of the distribution network under multiple time sections;

[0041] The cluster division module is used to perform clustering iteration using the K-means clustering algorithm according to the predetermined number of cluster divisions and the Laplace matrix to obtain the distribution network node cluster division result.

[0042] In a third aspect, the present invention further provides a computer device, comprising a processor and a memory, wherein the processor is connected to the memory, the memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory so that the computer device performs the steps of implementing the above method.

[0043] In a fourth aspect, the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and the computer program implements the steps of the above method when executed by a processor.

[0044] The present invention provides a novel distribution network cluster division method, system, device and storage medium. The method divides the time section of the distribution network into multiple time intervals according to the electrical characteristic parameters and operation data of the distribution network; in each time interval, the active power flow data of each node in the distribution network is calculated, and the active power flow data of each time interval is weighted and summed to obtain the active power flow comprehensive index; according to the voltage data of each node of the distribution network under multiple time sections, the voltage cosine similarity and the voltage Euclidean distance between different nodes of the distribution network are calculated; according to the active power flow comprehensive index, the voltage cosine similarity and the voltage Euclidean distance of the distribution network under multiple time sections, a Laplace matrix is ​​constructed; according to the predetermined number of cluster divisions and the Laplace matrix, a K-means clustering algorithm is used to perform clustering iteration to obtain the distribution network node cluster division result. Compared with the existing technology, this method makes full use of the inherent connection between distribution network parameters and operation data as well as the influence of time characteristics on cluster division, adopts spectral clustering algorithm for cluster division, realizes refined and dynamic cluster division of distribution network, improves the accuracy and stability of cluster division, ensures the optimized operation of distribution network and reasonable allocation of resources, and improves the operation efficiency and stability of distribution network. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 It is a schematic diagram of the flow of a novel distribution network cluster division method provided by an embodiment of the present invention;

[0046] Figure 2 is a schematic diagram of the topology of distribution network nodes provided by an embodiment of the present invention;

[0047] Figure 3 is a schematic diagram of the distribution network node cluster division result provided by an embodiment of the present invention;

[0048] Figure 4 It is a block diagram of a novel distribution network cluster division system provided by an embodiment of the present invention;

[0049] Figure 5 It is a schematic diagram of the structure of a computer device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0050] The following specifically illustrates the implementation mode of the present invention in conjunction with the accompanying drawings. The embodiments are provided for illustrative purposes only and cannot be understood as limiting the present invention. The accompanying drawings are provided for reference and illustration only and do not constitute a limitation on the scope of patent protection of the present invention, because many changes can be made to the present invention without departing from the spirit and scope of the present invention.

[0051] refer to Figure 1 , the embodiment of the present invention provides a new distribution network cluster division method, such as Figure 1 As shown, the method comprises the following steps:

[0052] S1. According to the electrical characteristic parameters and operation data of the distribution network, the time section of the distribution network is divided into multiple time intervals.

[0053] S2. In each time interval, the active power flow data of each node in the distribution network is calculated, and the active power flow data of each time interval is weighted and summed to obtain a comprehensive active power flow index.

[0054] In this embodiment, the active power flow data includes an average active power flow expected value and an active power flow variance; the step of calculating the active power flow data of each node in the distribution network in each time interval includes:

[0055] Calculate the ratio between the sum of the time section flow values ​​between different nodes of the distribution network in all time intervals and the number of time sections to obtain the expected value of the average active power flow;

[0056] Calculate the sum of squares of the differences between the power flow values ​​between different nodes of the distribution network and the expected value of the average active power flow in each time interval to obtain the active power flow fluctuation energy;

[0057] According to the number of time sections in each time interval and the active power flow fluctuation energy, the active power flow variance between different nodes of the distribution network in each time interval is obtained.

[0058] Specifically, this embodiment divides the time section of the distribution network into multiple time intervals according to the actual operation status and load characteristics of the distribution network to reflect the load characteristics and active power flow changes in different time intervals. For example, the time section is divided into four time periods: morning peak, flat peak, evening peak and night valley. Then, in each time interval, the expected value (mean) and variance (volatility) of the active power flow between each node pair in the distribution network are calculated. This embodiment assumes that the distribution network includes N nodes, numbered 0 to N-1, where node 0 is a balancing node; the branches are numbered 1 to L, with a total of L branches. For the branch l(i, j) between node i and node j, in each time interval g, the average active power flow expected value between node i and node j is calculated according to the real-time operation data of the distribution network. The calculation formula of the average active power flow expected value is:

[0059]

[0060] In the formula, p ij (t) represents the active power flow between nodes i and j at time t; T g Indicates the number of time sections in the time interval g; S g Represents the set of all time sections corresponding to the time interval g; It represents the expected value of the average active power flow between nodes i and j, where the expected value of the average active power flow describes the average power flow between nodes i and j in a certain time interval, reflecting the overall level of the power flow in this period, where nodes i and j are located at the two ends of line l respectively.

[0061] At the same time, in each time interval g, the active power flow variance between node i and node j is calculated. The calculation formula of the active power flow variance is as follows:

[0062]

[0063] In the formula, It represents the variance of the active power flow between nodes i and j in time interval g. It reflects the fluctuation amplitude of the power flow between nodes i and j in time interval g. A time interval with large fluctuation will have a higher variance value, indicating that the change amplitude of the power flow in this time interval is large.

[0064] After obtaining the average active power flow expected value and the active power flow variance, this embodiment combines the active power flow expected value and variance of each time interval, as well as the basic weight coefficient and the adjustment coefficient of the time interval to calculate the comprehensive index of the active power flow. In this embodiment, the step of performing weighted summation on the active power flow data of each time interval to obtain the active power flow comprehensive index includes:

[0065] According to the active power flow variance between different nodes of the distribution network in each time interval and the preset time interval weight coefficient, a weighted value of the active power flow variance is obtained;

[0066] Calculating the active power flow segmentation index in each time interval according to the active power flow variance weighted value and the average active power flow expected value;

[0067] The active power flow segment indexes of all time intervals are summed to obtain the active power flow comprehensive index; the calculation formula of the active power flow comprehensive index is:

[0068]

[0069] In the formula, represents the comprehensive index of active power flow; G represents the total number of time intervals; α g represents the basic weight coefficient of the time interval g, which represents the impact of this period on the final tidal flow value, where, This implementation can assign a larger weight to the time periods with larger loads (such as morning peak and evening peak); β represents the adjustment coefficient, which is used to control the influence of the power flow variance on the weight. In this embodiment, the adjustment coefficient is usually set to a positive value, indicating that the time period with large fluctuations has less influence on the overall power flow; represents the variance of active power flow between node i and node j in time interval g; Represents the expected value of the average active power flow between node i and node j; it should be noted that those skilled in the art can set and adjust the basic weight coefficient and adjustment coefficient of each time interval according to the importance of the time interval and the load characteristics.

[0070] In summary, this embodiment realizes a comprehensive analysis and evaluation of the active power flow characteristics of the distribution network through steps such as time section division and active power flow characteristic calculation, which can reflect the changes in the active power flow of the distribution network in different time periods and provide strong technical support for the optimized operation of the distribution network.

[0071] S3. According to the voltage data of each node in the distribution network under multiple time sections, the voltage cosine similarity and voltage Euclidean distance between different nodes in the distribution network are calculated.

[0072] In this embodiment, a voltage curve is formed according to the voltage data of each node in the distribution network at multiple time sections, and then the cosine similarity and Euclidean distance between the voltage curves of each node are calculated to evaluate the similarity and difference of the voltage curves. For node i and node j, the voltage cosine similarity calculation formula of their voltage curves is:

[0073]

[0074] In the formula, δ ij represents the voltage cosine similarity between distribution network nodes i and j; V i (t) represents the voltage of distribution network node i at time section t; V j (t) represents the voltage of distribution network node j at time section t; T represents the total number of time sections; B represents the node set consisting of nodes other than the balancing node; S represents the set of time sections, S = {1, 2, ..., T}.

[0075] At the same time, for the distribution network nodes i and j, according to the definition of Euclidean distance, this embodiment calculates the voltage Euclidean distance of their voltage curves as follows:

[0076]

[0077] In the formula, e ij Represents the voltage Euclidean distance.

[0078] S4. According to the comprehensive index of active power flow, voltage cosine similarity and voltage Euclidean distance of the distribution network in multiple time sections, a Laplace matrix is ​​constructed.

[0079] In this embodiment, the step of constructing a Laplace matrix according to the comprehensive active power flow index, voltage cosine similarity and voltage Euclidean distance of the distribution network at multiple time sections includes:

[0080] The comprehensive active power flow index, voltage cosine similarity and voltage Euclidean distance of the distribution network under multiple time sections are indexed respectively to obtain the corresponding indexed comprehensive active power flow index, indexed voltage cosine similarity and indexed voltage Euclidean distance;

[0081] Normalizing the indexed active power flow comprehensive index, the indexed voltage cosine similarity and the indexed voltage Euclidean distance respectively to obtain corresponding normalized active power flow comprehensive index, normalized voltage cosine similarity and normalized voltage Euclidean distance;

[0082] Performing weighted summation on the normalized active power flow comprehensive index, the normalized voltage cosine similarity and the normalized voltage Euclidean distance to construct a spectral clustering similarity matrix;

[0083] A degree matrix is ​​calculated according to the spectral clustering similarity matrix, and a Laplace matrix is ​​calculated by the difference between the degree matrix and the spectral clustering similarity matrix.

[0084] Specifically, this embodiment uses an exponential function to transform the active power flow comprehensive index to obtain an indexed active power flow comprehensive index, and normalizes the indexed active power flow comprehensive index so that the values ​​of all elements are between 0 and 1 to obtain a normalized active power flow comprehensive index, wherein the calculation formula of the indexed active power flow comprehensive index is:

[0085]

[0086] Where P ij It represents the comprehensive index of the indexed active power flow between node i and node j.

[0087] At the same time, for each pair of nodes (i, j), this embodiment uses an exponential function to transform the voltage cosine similarity according to the voltage cosine similarity between the voltage curves to obtain an exponential voltage cosine similarity, and normalizes the exponential voltage cosine similarity to obtain a normalized voltage cosine similarity, wherein the calculation formula of the exponential voltage cosine similarity is:

[0088]

[0089] In the formula, Δ ij represents the exponential voltage cosine similarity between node i and node j; ξ and ζ represent adjustment coefficients, wherein the adjustment coefficient here is used to adjust the degree of indexation.

[0090] For each pair of nodes (i, j), this embodiment calculates the voltage Euclidean distance between their voltage curves, and uses an exponential function to transform the voltage Euclidean distance to obtain an exponential voltage Euclidean distance. The exponential process here helps to convert the distance difference into a similarity metric, and the exponential voltage Euclidean distance is normalized to obtain a normalized voltage Euclidean distance, wherein the calculation formula of the exponential voltage Euclidean distance is:

[0091]

[0092] In the formula, E ij represents the exponential voltage Euclidean distance between node i and node j.

[0093] According to actual needs, this embodiment determines the weight coefficients of the normalized active power flow comprehensive index, the normalized voltage cosine similarity, and the normalized voltage Euclidean distance respectively, and according to the weight coefficient, the normalized active power flow comprehensive index, the normalized voltage cosine similarity, and the normalized voltage Euclidean distance are weighted and summed to obtain a spectral clustering similarity matrix. The elements of the spectral clustering similarity matrix are expressed as:

[0094]

[0095] Where α, β and γ represent the weight coefficients of the normalized active power flow comprehensive index, the normalized voltage cosine similarity and the normalized voltage Euclidean distance, respectively, and satisfy α+β+γ=1.

[0096] Then, this embodiment calculates the degree matrix according to the spectral clustering similarity matrix. The degree matrix is ​​a diagonal matrix whose elements are equal to the sum of the elements of each row in the spectral clustering similarity matrix. The elements of the degree matrix are expressed as follows:

[0097]

[0098] Where D ii Represents the elements of the degree matrix; W ij Represents the elements of the spectral clustering similarity matrix.

[0099] Finally, the Laplace matrix is ​​calculated based on the spectral clustering similarity matrix and degree matrix. The calculation formula of the Laplace matrix is:

[0100] L=DW

[0101] Where L represents the Laplace matrix; D represents the degree matrix; W represents the spectral clustering similarity matrix.

[0102] S5. According to the predetermined number of cluster divisions and the Laplace matrix, a K-means clustering algorithm is used to perform clustering iterations to obtain a distribution network node cluster division result.

[0103] In this embodiment, the step of performing clustering iteration using a K-means clustering algorithm according to a predetermined number of cluster divisions and the Laplace matrix to obtain a distribution network node cluster division result includes:

[0104] According to the predetermined number of cluster divisions, several minimum eigenvalues ​​of the Laplacian matrix and their corresponding eigenvectors are calculated;

[0105] A normalized feature matrix is ​​constructed based on the feature vectors, and a preferred clustering center is determined based on the density distribution and distance characteristics of the feature data points in the normalized feature matrix;

[0106] The preferred cluster center is used as the initial cluster center of K-means clustering, and K-means clustering is used to perform clustering iteration on the normalized feature matrix to obtain a distribution network node cluster division result.

[0107] Wherein, the step of determining the preferred cluster center according to the density distribution and distance characteristics of the feature data points in the normalized feature matrix includes:

[0108] Traversing each feature data point in the normalized feature matrix, and obtaining the density distribution of the feature data point according to the average distance from each feature data point to its nearest plurality of neighboring points;

[0109] Initialize the center point set and calculate the minimum distance from each non-center point feature data point to the center point set;

[0110] According to the density distribution and minimum distance of the feature data points, a comprehensive selection index is obtained;

[0111] The characteristic data point with the largest comprehensive selection index is taken as the next center point until the preferred cluster centers equal to the number of cluster divisions are screened out.

[0112] Specifically, this embodiment calculates the smallest k eigenvalues ​​of the Laplace matrix L and the eigenvectors corresponding to these eigenvalues ​​according to the predetermined number of cluster divisions k. The calculation formula of the eigenvector is:

[0113] Lf i =λ i f i i=1,2,...,k

[0114] In the formula, λ i represents the i-th eigenvalue of the Laplace matrix L; f i Represents the eigenvector corresponding to the i-th eigenvalue.

[0115] Then, this embodiment arranges the calculated k eigenvectors in columns to form a feature matrix F. The formula of the feature matrix F is as follows:

[0116] F=[f1 f2 f3...f k ]

[0117] In the formula, each column of the feature matrix F is a feature vector f i , which correspond to the smallest k eigenvalues ​​of the Laplacian matrix L.

[0118] Next, this embodiment normalizes the feature matrix F by row to eliminate the dimensional differences between different feature vectors, making the clustering algorithm more stable, and obtaining a normalized feature matrix. For each feature data point (i.e., each row) in the normalized feature matrix, the inverse of the average distance to the nearest p neighbor points is calculated as the density distribution of the feature data point, where p is a preset neighbor number parameter, and the density distribution calculation formula of the feature data point is:

[0119]

[0120] in,

[0121]

[0122] In the formula, ρ i Represents the density distribution of the i-th feature data point; d ij represents the Euclidean distance between feature data point i and feature data point j; top-p represents the set of p neighboring points closest to feature data point i; Represents the value corresponding to the characteristic data point i in the normalized characteristic matrix for the mth eigenvalue; Represents the value corresponding to the feature data point j in the normalized feature matrix for the mth eigenvalue.

[0123] At the same time, for each data point that is not selected as a center point, this embodiment calculates its minimum distance to the set of selected center points. The specific formula is as follows:

[0124]

[0125] Where c represents the center point in the selected center point set C; if no center point has been selected (i.e. ), then let δ i is a large constant (such as infinity).

[0126] This embodiment defines a comprehensive selection index Z i It is the product of the density distribution of feature data points and the minimum distance. The specific calculation formula of the comprehensive selection index is:

[0127] Z i =ρ i ·δ i

[0128] In this embodiment, the point with the largest comprehensive selection index is preferentially selected as the next center point each time until k center points are selected. These center points will be used as the initial clustering centers of K-means clustering. According to the screened k initial clustering centers, K-means clustering is performed on the normalized feature matrix. The K-means clustering algorithm is an iterative algorithm. By continuously adjusting the position of the clustering center, the sum of the distances from each data point to its clustering center is minimized until the clustering center no longer changes or the preset number of iterations is reached, and the cluster label to which each data point belongs, that is, the distribution network node cluster division result, is obtained.

[0129] For ease of understanding, this embodiment uses Figure 2 Taking the IEEE69 node system shown in the figure as an example, its system baseline capacity is set to Sbase=10MVA, and the baseline voltage is Ubase=12.66kV. The system contains 69 nodes in total, among which nodes 15, 20, 25, 34, 45, 49, 57, 61, 65, and 67 are selected as photovoltaic access points, and the connected photovoltaic capacities are 0.8MVA, 1.2MVA, 0.8MVA, 0.53MVA, 0.46MVA, 0.32MVA, 0.8MVA, 0.53MVA, 1.2MVA, and 0.8MVA, respectively, and the power factor of all photovoltaic access points is set to 0.95. In terms of data collection, the system continuously samples the load and photovoltaic output with a sampling period of 15 minutes. The sampling period covers 365 days a year. Based on these sampled data, the system generates a distribution network operation data sample containing 365244 time sections.

[0130] In order to effectively analyze the data of these time sections, this embodiment divides a day into four time periods and assigns a basic weight coefficient to each time period. The weight coefficient of the morning peak period is 0.3; the weight coefficient of the flat peak period is 0.2; the weight coefficient of the evening peak period is 0.4; the weight coefficient of the night valley period is 0.1. In addition, this embodiment sets the variance adjustment coefficient to 1, and the cosine similarity index adjustment coefficient of the node voltage curve is 100 and 99.9 respectively. In terms of cluster division, this embodiment sets the number of clusters to 4, and assigns weight coefficients to the three normalized index indicators, specifically 0.1, 0.2 and 0.7. In the process of optimizing the initial cluster center, the number of neighbor points considered is set to 5. Finally, this embodiment uses MATLAB software for calculation and obtains the result of cluster division, such as Figure 3 As shown, Figure 3 Nodes with the same color in the figure indicate that they belong to the same cluster, while triangle marks represent PV access nodes.

[0131] The embodiment of the present invention provides a novel distribution network clustering method, which divides the time section of the distribution network into multiple time intervals according to the electrical characteristic parameters and operation data of the distribution network; in each time interval, the active power flow data of each node in the distribution network is calculated, and the active power flow data of each time interval is weighted and summed to obtain the active power flow comprehensive index; according to the voltage data of each node of the distribution network under multiple time sections, the voltage cosine similarity and voltage Euclidean distance between different nodes of the distribution network are calculated; according to the active power flow comprehensive index, voltage cosine similarity and voltage Euclidean distance of the distribution network under multiple time sections, a Laplace matrix is ​​constructed; according to the predetermined number of cluster divisions and the Laplace matrix, a K-means clustering algorithm is used to perform clustering iteration to obtain the distribution network node clustering result. Compared with the prior art, the method makes full use of the intrinsic connection between the network parameters and operation data of the distribution network and the influence of time characteristics on cluster division, adopts the spectral clustering algorithm for cluster division, realizes the refined and dynamic cluster division of the distribution network, improves the accuracy and stability of cluster division, and ensures the optimized operation of the distribution network and the reasonable allocation of resources.

[0132] It should be noted that the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0133] In one embodiment, Figure 4 As shown, an embodiment of the present invention provides a novel distribution network cluster division system, the system comprising:

[0134] The interval division module 101 is used to divide the time section of the distribution network into multiple time intervals according to the electrical characteristic parameters and operation data of the distribution network;

[0135] The power flow analysis module 102 is used to calculate the active power flow data of each node in the distribution network in each time interval, and perform weighted summation on the active power flow data of each time interval to obtain a comprehensive active power flow index;

[0136] The voltage analysis module 103 is used to calculate the voltage cosine similarity and voltage Euclidean distance between different nodes of the distribution network according to the voltage data of each node of the distribution network under multiple time sections;

[0137] A matrix construction module 104 is used to construct a Laplace matrix according to the comprehensive index of active power flow, voltage cosine similarity and voltage Euclidean distance of the distribution network under multiple time sections;

[0138] The cluster division module 105 is used to perform clustering iteration using a K-means clustering algorithm according to a predetermined number of cluster divisions and the Laplace matrix to obtain a distribution network node cluster division result.

[0139] For the specific definition of a new distribution network cluster partitioning system, please refer to the above-mentioned definition of a new distribution network cluster partitioning method, which will not be repeated here. A person of ordinary skill in the art can appreciate that the various modules and steps described in conjunction with the embodiments disclosed in this application can be implemented in hardware, software, or a combination of both. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0140] The embodiment of the present invention provides a novel distribution network cluster division system, wherein the interval division module of the system divides the time section of the distribution network into multiple time sections according to the electrical characteristic parameters and operation data of the distribution network; the power flow analysis module calculates the active power flow data of each node in the distribution network in each time section, and performs weighted summation on the active power flow data of each time section to obtain an active power flow comprehensive index; the voltage analysis module calculates the voltage cosine similarity and voltage Euclidean distance between different nodes of the distribution network according to the voltage data of each node of the distribution network under multiple time sections; the matrix construction module constructs a Laplace matrix according to the active power flow comprehensive index, voltage cosine similarity and voltage Euclidean distance of the distribution network under multiple time sections; the cluster division module performs clustering iteration using the K-means clustering algorithm according to a predetermined number of cluster divisions and the Laplace matrix to obtain a distribution network node cluster division result. Compared with the existing technology, this system makes full use of the intrinsic connection between distribution network parameters and operation data as well as the influence of time characteristics on cluster division, adopts spectral clustering algorithm for cluster division, realizes refined and dynamic cluster division of distribution network, improves the accuracy and stability of cluster division, and ensures the optimized operation of distribution network and rational allocation of resources.

[0141] Figure 5 A computer device provided by an embodiment of the present invention includes a memory, a processor and a transceiver, which are connected via a bus; the memory is used to store a set of computer program instructions and data, and can transmit the stored data to the processor, and the processor can execute the program instructions stored in the memory to perform the steps of the above method.

[0142] The memory may include a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories; the processor may be a central processing unit, a microprocessor, an application-specific integrated circuit, a programmable logic device, or a combination thereof. By way of example but not limitation, the programmable logic device may be a complex programmable logic device, a field programmable gate array, a general purpose array logic, or any combination thereof.

[0143] Additionally, the memory may be a physically separate unit or may be integrated with the processor.

[0144] It can be understood by those skilled in the art that Figure 5 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have the same component arrangement.

[0145] In one embodiment, the present invention provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of the above method are implemented.

[0146] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented by software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the embodiment of the present invention is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from a website site, computer, server or data center to another website site, computer, server or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line) or wireless (e.g., infrared, wireless, microwave, etc.) mode. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more available media integrated. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., an SSD), etc.

[0147] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing related hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods.

[0148] The above-mentioned embodiments only express several preferred implementation modes of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for ordinary technicians in the technical field, several improvements and substitutions can be made without departing from the technical principles of the present invention, and these improvements and substitutions should also be regarded as the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be based on the protection scope of the claims.

Claims

1. A new distribution network cluster division method, characterized in that: The following steps are involved: According to the electrical characteristic parameters and operation data of the distribution network, the time section of the distribution network is divided into multiple time intervals; In each time interval, the active power flow data of each node in the distribution network is calculated, and the active power flow data of each time interval is weighted and summed to obtain the active power flow comprehensive index; According to the voltage data of each node in the distribution network under multiple time sections, the voltage cosine similarity and voltage Euclidean distance between different nodes in the distribution network are calculated; According to the comprehensive index of active power flow, voltage cosine similarity and voltage Euclidean distance of distribution network in multiple time sections, the Laplace matrix is ​​constructed. According to the predetermined number of cluster divisions and the Laplace matrix, a K-means clustering algorithm is used to perform clustering iterations to obtain a distribution network node cluster division result.

2. A new distribution network cluster division method as claimed in claim 1, characterized in that: The active power flow data includes an average active power flow expected value and an active power flow variance; The step of calculating the active power flow data of each node in the distribution network in each time interval includes: Calculate the ratio between the sum of the time section flow values ​​between different nodes of the distribution network in all time intervals and the number of time sections to obtain the expected value of the average active power flow; Calculate the sum of squares of the differences between the power flow values ​​between different nodes of the distribution network and the expected value of the average active power flow in each time interval to obtain the active power flow fluctuation energy; According to the number of time sections in each time interval and the active power flow fluctuation energy, the active power flow variance between different nodes of the distribution network in each time interval is obtained.

3. A new distribution network cluster division method as claimed in claim 2, characterized in that: The step of performing weighted summation of the active power flow data of each time interval to obtain the active power flow comprehensive index comprises: According to the active power flow variance between different nodes of the distribution network in each time interval and the preset time interval weight coefficient, a weighted value of the active power flow variance is obtained; Calculating the active power flow segmentation index in each time interval according to the active power flow variance weighted value and the average active power flow expected value; The active power flow segment indicators of all time intervals are summed up to obtain the active power flow comprehensive indicator.

4. A new distribution network cluster division method as claimed in claim 1, characterized in that: The step of constructing a Laplace matrix according to the comprehensive active power flow index, voltage cosine similarity and voltage Euclidean distance of the distribution network under multiple time sections includes: The comprehensive active power flow index, voltage cosine similarity and voltage Euclidean distance of the distribution network under multiple time sections are indexed respectively to obtain the corresponding indexed comprehensive active power flow index, indexed voltage cosine similarity and indexed voltage Euclidean distance; Normalizing the indexed active power flow comprehensive index, the indexed voltage cosine similarity and the indexed voltage Euclidean distance respectively to obtain corresponding normalized active power flow comprehensive index, normalized voltage cosine similarity and normalized voltage Euclidean distance; Performing weighted summation on the normalized active power flow comprehensive index, the normalized voltage cosine similarity and the normalized voltage Euclidean distance to construct a spectral clustering similarity matrix; A degree matrix is ​​calculated according to the spectral clustering similarity matrix, and a Laplace matrix is ​​calculated by the difference between the degree matrix and the spectral clustering similarity matrix.

5. A new distribution network cluster division method as claimed in claim 1, characterized in that: The step of performing clustering iteration using a K-means clustering algorithm according to a predetermined number of cluster divisions and the Laplace matrix to obtain a distribution network node cluster division result comprises: According to the predetermined number of cluster divisions, several minimum eigenvalues ​​of the Laplacian matrix and their corresponding eigenvectors are calculated; A normalized feature matrix is ​​constructed based on the feature vectors, and a preferred clustering center is determined based on the density distribution and distance characteristics of the feature data points in the normalized feature matrix; The preferred cluster center is used as the initial cluster center of K-means clustering, and K-means clustering is used to perform clustering iteration on the normalized feature matrix to obtain a distribution network node cluster division result.

6. A new distribution network cluster division method as claimed in claim 5, characterized in that: The step of constructing a normalized feature matrix according to the feature vector comprises: All the eigenvectors are arranged in columns to form a eigenmatrix, and the eigenmatrix is ​​normalized in rows to obtain a normalized eigenmatrix.

7. A new distribution network cluster division method as claimed in claim 5, characterized in that: The step of determining the preferred cluster center according to the density distribution and distance characteristics of the feature data points in the normalized feature matrix comprises: Traversing each feature data point in the normalized feature matrix, and obtaining the density distribution of the feature data point according to the average distance from each feature data point to its nearest plurality of neighboring points; Initialize the center point set and calculate the minimum distance from each non-center point feature data point to the center point set; According to the density distribution and minimum distance of the feature data points, a comprehensive selection index is obtained; The characteristic data point with the largest comprehensive selection index is taken as the next center point until the preferred cluster centers equal to the number of cluster divisions are screened out.

8. A new distribution network cluster division system, characterized in that: The system comprises: An interval division module is used to divide the time section of the distribution network into multiple time intervals according to the electrical characteristic parameters and operation data of the distribution network; The power flow analysis module is used to calculate the active power flow data of each node in the distribution network in each time interval, and perform weighted summation of the active power flow data of each time interval to obtain the active power flow comprehensive index; The voltage analysis module is used to calculate the voltage cosine similarity and voltage Euclidean distance between different nodes of the distribution network based on the voltage data of each node in the distribution network under multiple time sections; A matrix construction module is used to construct a Laplace matrix according to the comprehensive active power flow index, voltage cosine similarity and voltage Euclidean distance of the distribution network under multiple time sections; The cluster division module is used to perform clustering iteration using the K-means clustering algorithm according to the predetermined number of cluster divisions and the Laplace matrix to obtain the distribution network node cluster division result.

9. A computer device, characterized in that: The computer device comprises a processor and a memory, wherein the processor is connected to the memory, the memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory, so that the computer device executes the method according to any one of claims 1 to 7.

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

Images