Method and system for checking branch relation of low-voltage transformer area

Abstract

The invention discloses a low-voltage transformer area branch relation checking method, which comprises the following steps of: acquiring current and voltage data of each user in the same transformer area, and recording a matching number between current and voltage effective value cross phases; according to the recorded matching number and the normalized association strength, constructing a voltage and current change association matrix between the users; according to the incidence matrix, constructing a distance matrix based on a Pearson correlation coefficient; using a DBSCAN clustering method to carry out clustering analysis on the user to obtain a preliminary user branch; the DBSCAN clustering result is fused with the minimum spanning tree, and user branch division is optimized; branch relation checking is conducted, a final branch division result is output, and the branch division result comprises normal branches and abnormal users. The invention further discloses a corresponding system. According to the method, the existing voltage and current data are utilized, and the DBSCAN clustering and the minimum spanning tree algorithm are combined, so that accurate identification and optimization of the branch relation are realized.

Description

Technical Field

[0001] The present invention relates to the technical field of low-voltage distribution network management and analysis, and in particular to a low-voltage distribution network branch relationship verification method and system based on DBSCAN clustering and minimum spanning tree. Background Art

[0002] Branch relationship identification technology, also known as branch restoration, is a key technology used in power systems to determine the branch boxes to which each meter box and user meter belongs in the low-voltage area. Currently, low-voltage distribution networks are usually arranged in a tree or ring network structure with a complex structure to ensure the reliability and stability of power supply.

[0003] Branch relationship identification mainly relies on the electricity information collection system, which infers the specific branch relationship of each user by collecting and analyzing the data of the user's electricity meter and combining the distribution of branch boxes and meter boxes. Commonly used technical means include analysis methods based on power parameters such as current and voltage, and technology for spatial data processing combined with geographic information systems (GIS). With the continuous development of smart grids, more and more research has begun to use machine learning and data mining technologies to improve the accuracy and efficiency of branch relationship identification through big data analysis.

[0004] Although branch relationship recognition technology has made some progress in theory and application, there are still many shortcomings. First, the existing electricity consumption information collection system usually lacks the ability to collect detailed electricity consumption information at the branch layer and meter box layer, so it is impossible to achieve accurate branch relationship recognition by directly analyzing the data between the user's meter and the branch box and meter box. This limitation is mainly due to the high installation cost and limited coverage of electricity consumption information collection equipment, as well as technical bottlenecks in data transmission and processing.

[0005] Secondly, when dealing with complex low-voltage distribution network structures, existing methods often face the problems of large data noise and low recognition accuracy. This is mainly because the branch structure of the low-voltage distribution network is diverse and dynamically changing, and traditional rule-based or simple algorithm-based methods are difficult to adapt to this complexity. In addition, some studies ignore the real-time requirements in practical applications, making it difficult to achieve real-time update and maintenance of branch relationships in a dynamically changing power grid environment.

[0006] Finally, current research on branch relationship identification is mostly focused on theoretical models and laboratory environments, lacking data support and verification for large-scale practical applications, which limits the promotion and application of these technologies in actual power systems. Summary of the invention

[0007] The technical problem to be solved by the present invention is to provide a low-voltage area branch relationship verification method and system. The accurate identification and optimization of branch relationships can be achieved by utilizing existing voltage and current data in combination with DBSCAN clustering and minimum spanning tree algorithms.

[0008] In order to solve the above technical problems, as one aspect of the present invention, a low voltage area branch relationship verification method is provided, which is based on DBSCAN clustering and minimum spanning tree technology, and comprises the following steps: Obtain the current and voltage data of each user in the same substation, and record the matching number of the effective values ​​of current and voltage across phases; According to the matching number of records, the correlation strength is normalized and the voltage and current change correlation matrix between users is constructed; According to the association matrix, construct a distance matrix based on the Pearson correlation coefficient; Use DBSCAN clustering method to cluster users and get preliminary user branches; Combine the DBSCAN clustering results with the minimum spanning tree to optimize user branch division; Perform branch relationship verification and output the final branch division result, which includes normal branches and abnormal users.

[0009] Preferably, the step of obtaining the current and voltage data of each user in the same substation and recording the number of matches between the effective values ​​of the current and voltage across phases includes: Extract and record the corresponding time when the absolute value of the voltage effective value decreases and increases within the preset time is greater than or equal to the preset voltage threshold; and extract and record the corresponding time when the absolute value of the current effective value decreases and increases within the preset time is greater than or equal to the preset current threshold; Based on the recorded three-phase voltage and current of all users, the records of current increase and voltage increase of each phase between different users and between the same users are screened.

[0010] Preferably, the normalized correlation strength is used to construct a voltage and current change correlation matrix between users, including: According to the number of users who need to check the branch relationship, a matrix is ​​constructed, and all the initial elements of the matrix are 0; Assign different numbers to the users to be verified; Count the intersection of the time points of the cross-phase user current rise and the user voltage rise, and calculate the intersection of the two sets; Assign values ​​to the elements in the current and voltage change correlation matrix based on the records.

[0011] Preferably, constructing a distance matrix based on the Pearson correlation coefficient according to the association matrix further comprises: According to the association matrix, the characteristic vector of each user is defined as the corresponding row of the user in the association matrix, and the characteristic vector representing the distribution of the correlation strength of the voltage and current changes of the user and all other users is obtained; Calculating the Pearson correlation coefficient based on the feature vectors to measure the similarity between feature vectors of different users; Convert the Pearson correlation coefficient to distance and construct a symmetric distance matrix.

[0012] Preferably, the method of clustering the users using the DBSCAN clustering method to obtain preliminary user branches includes: Based on the constructed distance matrix between users, the density-based clustering method DBSCAN is used for clustering analysis; Set DBSCAN clustering parameters, including domain radius and minimum number of cores; the method for selecting the domain radius is: sort the kth nearest neighbor distance of the distance matrix, find the natural jump point of the distance according to the "inflection point method", and set it as the domain radius; the minimum number of cores is adjusted according to the user scale and data distribution of the station area; Execute the DBSCAN clustering process, mark all user points as unvisited, define core points, boundary points, and noise points, check the neighborhood, and output the clustering results.

[0013] Preferably, the step of fusing the DBSCAN clustering result with the minimum spanning tree to optimize the user branch division includes: Constructing a minimum spanning tree based on a distance matrix includes: creating an empty edge set and vertex set, sorting all edges in the edge set according to weights from small to large, traversing the sorted edge set, and for each edge, checking whether the vertex belongs to the same connected component. If not, adding the edge to the edge set and merging the connected components where the vertex is located, stopping the algorithm when the number of edges reaches the number of user points minus one, thereby generating a minimum spanning tree; The clustering results are integrated with the minimum spanning tree, including: clustering consistency check, cross-clustering connection optimization, and noise point merging to identify abnormal users.

[0014] Accordingly, as another aspect of the present invention, a low voltage area branch relationship verification system based on DBSCAN clustering and minimum spanning tree is also provided, which comprises: The data acquisition module is used to obtain the current and voltage data of each user in the same substation area, and record the matching number between the effective values ​​of current and voltage across phases; The correlation matrix construction module is used to normalize the correlation strength according to the matching number of records and construct the voltage and current change correlation matrix between users; A distance matrix construction module, used to construct a distance matrix based on the Pearson correlation coefficient according to the association matrix; Cluster analysis module, used to perform cluster analysis on users using DBSCAN clustering method to obtain preliminary user branches; Clustering fusion module, used to fuse DBSCAN clustering results with the minimum spanning tree to optimize user branch division; The output module is used to perform branch relationship verification and output the final branch division result, which includes normal branches and abnormal users.

[0015] Preferably, the data acquisition module further comprises: A high-frequency wave recording unit, used to extract and record the corresponding moments when the absolute value of the voltage effective value decreases and increases within a preset time is greater than or equal to a preset voltage threshold; and to extract and record the corresponding moments when the absolute value of the current effective value decreases and increases within a preset time is greater than or equal to a preset current threshold; The screening unit is used to screen the records of current rise and voltage rise of each phase between all different users and between the same user according to the recorded three-phase voltage and current of all users.

[0016] Preferably, the association matrix construction module comprises: A matrix construction unit is used to construct a matrix according to the number of users who need to perform branch relationship verification, and all initial elements of the matrix are 0; A numbering unit is used to assign different numbers to the users to be verified; A statistical unit, used to count the intersection of the time points of the cross-phase user current rise and the user voltage rise, and calculate the intersection of the two sets; The assignment unit is used to measure the similarity between user feature vectors using the Pearson correlation coefficient, convert the Pearson correlation coefficient into a distance, and assign values ​​to the elements in the current and voltage change association matrix according to the records.

[0017] Preferably, the distance matrix construction module further comprises: A feature vector obtaining unit, used to define the feature vector of each user as the corresponding row of the user in the association matrix according to the association matrix, and obtain a feature vector representing the distribution of the association strength of the voltage and current changes of the user with all other users; A coefficient calculation unit, used to calculate the Pearson correlation coefficient according to the feature vector to measure the similarity between the feature vectors of different users; The distance calculation unit is used to convert the Pearson correlation coefficient into a distance and construct a symmetric distance matrix.

[0018] Preferably, the cluster analysis module includes: A clustering unit, used for performing clustering analysis using a density-based clustering method DBSCAN based on the constructed distance matrix between users; The parameter setting unit is used to set the DBSCAN clustering parameters, including the domain radius and the minimum number of cores; the method for selecting the domain radius is: sorting the kth nearest neighbor distance of the distance matrix, finding the natural jump point of the distance according to the "inflection point method", and setting it as the domain radius; the minimum number of cores is adjusted according to the user scale and data distribution of the station area; The clustering execution unit is used to execute the DBSCAN clustering process, mark all user points as unvisited, define core points, boundary points and noise points, check the neighborhood, and output the clustering results.

[0019] Preferably, the clustering fusion module includes: A minimum spanning tree construction unit is used to construct a minimum spanning tree based on a distance matrix, including: creating an empty edge set and vertex set, sorting all edges in the edge set in ascending order according to weight, traversing the sorted edge set, and for each edge, checking whether the vertex belongs to the same connected component, if not, adding the edge to the edge set and merging the connected components where the vertex is located, and stopping the algorithm when the number of edges reaches the number of user points minus one, thereby generating a minimum spanning tree; The checking unit is used to fuse the clustering results with the minimum spanning tree, including: clustering consistency check, cross-clustering connection optimization, and noise point merging to determine abnormal users.

[0020] The implementation of the embodiments of the present invention has the following beneficial effects: The present invention provides a low-voltage distribution area branch relationship verification method and system, which can achieve high-precision identification by relying on existing data without additional power signal injection through an analysis method based on user voltage change time records. The algorithm is designed to be efficient and has a short calculation cycle, and can quickly complete the branch relationship identification task to meet the real-time requirements of dynamic changes in the low-voltage distribution network.

[0021] This method is based on the voltage data analysis in the power consumption information collection system, and does not require the installation of additional hardware equipment, such as sensors or monitoring units. Compared with traditional methods, the present invention significantly reduces the equipment installation and operation and maintenance costs required for distribution network transparency, providing an economical and efficient solution for large-scale engineering applications.

[0022] The present invention optimizes the operation and management mode of the low-voltage distribution network through accurate branch relationship identification, and improves the level of digitalization and transparency of the distribution network. At the same time, it provides important technical support for the intelligent operation and maintenance of the low-voltage distribution network, and helps to improve the power supply reliability and management efficiency of the distribution network.

[0023] The present invention can handle complex tree or ring network structures in low-voltage distribution networks, and has strong robustness against data noise and dynamic changes in branch relationships. In practical applications, the recognition effect is stable and adaptable, and can meet the branch relationship recognition needs in different scenarios.

[0024] The present invention has shown good performance in practical applications and is suitable for large-scale promotion. Its core algorithm is simple and efficient, has both theoretical feasibility and engineering practicality, and does not require high equipment investment. It has broad commercial application prospects and can significantly improve the intelligence level of low-voltage distribution networks. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For a person skilled in the art, other drawings obtained based on these drawings still belong to the scope of the present invention without creative labor. Figure 1 A schematic diagram of the main process of an embodiment of a low-voltage station area branch relationship verification method provided by the present invention; Figure 2 A schematic structural diagram of an embodiment of a low voltage station area branch relationship verification system provided by the present invention; Figure 3 for Figure 2 The structural diagram of the data acquisition module in FIG. Figure 4 for Figure 2 A schematic diagram of the structure of the association matrix construction module in FIG. Figure 5 for Figure 2 Schematic diagram of the distance matrix construction module in ; Figure 6 for Figure 2 The schematic diagram of the cluster analysis module; Figure 7 for Figure 2 Schematic diagram of the structure of the clustering fusion module in . DETAILED DESCRIPTION

[0026] In order to make the objectives, technical solutions and advantages of the present invention more clear, the present invention will be further described in detail below with reference to the accompanying drawings.

[0027] like Figure 1As shown, a main process diagram of an embodiment of a low-voltage substation branch relationship verification method provided by the present invention is shown; it can be understood that according to the network topology and circuit principles of the low-voltage distribution network, when a user uses electricity, its current increases, and its own voltage effective value will decrease as a result. At the same time, the large voltage drop caused by the electricity consumption of this phase will cause a small increase in the voltage of other phases, that is, when the current of one phase increases when the user uses electricity, it will cause the voltage of other phases to increase, and this cross-phase transmission occurs more obviously in the same branch box, so that the branch relationship verification of the low-voltage substation can be realized based on this.

[0028] Therefore, in this embodiment, the method provided by the present invention includes at least the following steps: Step S10, obtaining the current and voltage data of each user in the same substation, and recording the number of matches between the effective values ​​of current and voltage across phases.

[0029] Specifically, it includes: step S101, according to the high-frequency recording data of the actual three-phase current and voltage values ​​of each user in the same substation, extracting and recording the corresponding moments when the absolute value of the voltage effective value decreases and increases within a preset time (such as 0.1 second) is greater than or equal to a preset voltage threshold (such as 0.1 volt), extracting and recording the corresponding moments when the absolute value of the current effective value decreases and increases within a preset time (0.1 second) is greater than or equal to a preset current threshold (5 amperes), so as to form a voltage change time record of the user, and the time record data format is year-month-day-hour-minute-second, and the time is in 24-hour format.

[0030] Step S102, filter the records of current increase and voltage increase of each phase between different users and the same user. Specifically, filter all records of current increase of phase A and voltage increase of phase B between different users according to the recorded three-phase voltage and current of all users, and filter the records of current increase of phase A and voltage increase of phase B for the same user. A, B, C are the default phase order.

[0031] Step S11, normalizing the correlation strength according to the matching number of the records, and constructing the voltage and current change correlation matrix between users.

[0032] Step S110: Check the number of users for branch relationship verification as needed , build a The matrix . All initial elements of are 0.

[0033] Step S111: assign different numbers to the users to be verified. .

[0034] Step S112: Counting cross-phase users Current rise and user The intersection of voltage rise time points, assuming that the user The current rise time set is ,user The voltage rise time set is . Calculate the intersection of two sets , that is, user and users At the same time point, the voltage and current change moments are collected and used Indicates the number of elements in the intersection; Step S113: Calculate the current and voltage change correlation matrix based on the records in step 2.3 Assign values ​​to the elements in .

[0035] Specifically , that is, user Current rise and user The voltage rise matches the number of user The ratio of the number of current rise events to the total number of current rise events. Indicates user The total number of current rising events is used to eliminate the impact of the number of events between different users and ensure the relative fairness of the association strength.

[0036] when >0, indicating that the user Current rise and user There is a correlation between the current and voltage changes across phases at the same time point of voltage rise, and the greater the strength, the closer the correlation.

[0037] Step S12: constructing a distance matrix based on the Pearson correlation coefficient (i.e., the Pearson correlation coefficient) according to the association matrix; Step S121: Construction of feature vector. According to the correlation matrix of step 2 , the feature vector of each user Defined as User In the corresponding row of the incidence matrix: , the eigenvector Characterizing Users The distribution of correlation strength of voltage and current changes with all other users is , the feature vector dimension of each user is .

[0038] Step S122: Pearson correlation coefficient calculation To measure users and users The similarity between the feature vectors of is measured using the Person correlation coefficient: ; in: and For users and users The eigenvector of The values ​​in the dimensions; and For users and users The mean of the eigenvector.

[0039] Step S12: Pearson distance conversion. To more intuitively reflect the "distance" between users, the Pearson correlation coefficient is converted into distance: . distance The smaller the size, the more users and users The closer the voltage and current changes are, the closer the structure The symmetric distance matrix ,in .

[0040] Step S13: Perform DBSCAN cluster analysis.

[0041] Step S131: The user distance matrix constructed in step 3 Based on this, the density-based clustering method DBSCAN (Density-Based Spatial Clustering of Applications with Noise) is used for cluster analysis. This method dynamically detects user groups and automatically identifies noise points by defining the neighborhood density relationship between users, which is suitable for the identification of branch structures in low-voltage areas.

[0042] Step S132: Setting parameters, including the area radius and minimum number of cores ; (Area radius): It is used to define whether two users can be considered as neighbors in the distance matrix. The selection method is as follows: No. Nearest neighbor distance ( ) are sorted, and the natural jump point of the distance is found according to the "inflection point method", which is set as .

[0043] (Minimum number of cores): indicates that each core point is The minimum number of users that must be included in the neighborhood is adjusted according to the user scale and data distribution of the station area.

[0044] Step S133: Execute DBSCAN clustering. Specifically, it includes: 1) Mark all user points as unvisited. Define three types of user points: core point, border point, and noise point.

[0045] 2) For each unvisited user point Do the following to check the neighborhood: calculate of The point set in the neighborhood , that is, satisfy All points .

[0046] like Points in ,Will Mark as core point. ,but Mark as boundary points or noise points.

[0047] For each core point ,like Not assigned to any cluster yet: Create a new cluster and will Join the cluster. All neighboring points Join a queue , check each point in the queue in turn: if the queue point is the core point, and its neighboring points Also join the queue , if the queue point If it is a boundary point, it will be added to the cluster directly. .queue After processing, we get the clustering The complete set of users.

[0048] 4) If a user clicks It is neither a core point nor belongs to any core point. neighborhood, it is marked as a noise point.

[0049] Step S134: output the clustering results.

[0050] DBSCAN outputs several clusters ,Each cluster represents a preliminarily identified user branch. Each user point is uniquely assigned to a cluster or marked as a noise point.

[0051] Step S14: Fusion of clustering results and minimum spanning tree (MST) to optimize user branching.

[0052] After completing DBSCAN clustering, in order to further verify and optimize the user branch division results, the minimum spanning tree (MST) is combined to check and adjust the clustering results.

[0053] The method includes: Step S141: constructing a minimum spanning tree (MST) based on the distance matrix. The method includes: creating an empty edge set and vertex set, sorting all edges in the edge set according to weight from small to large, traversing the sorted edge set, and for each edge, checking whether the vertex belongs to the same connected component. If not, adding the edge to the edge set and merging the connected components where the vertex is located, stopping the algorithm when the number of edges reaches the number of user points minus one, thereby generating a minimum spanning tree. Specifically, the user point is regarded as a vertex in the graph, and the user point and The edge weights between OK, use Kruskal's algorithm (a classic graph theory algorithm) to generate MST , ensuring the shortest connectivity structure between user points. Kruskal algorithm is a commonly used greedy algorithm for generating a minimum spanning tree (MST) in a weighted undirected graph. The specific steps are as follows.

[0054] The set of vertices of the graph ,correspond User points. The edge set of the graph , where each edge The weight of the user and users The distance between . Create an empty edge set Used to store edges in MST. Create an empty vertex set Used to store vertices in a minimum spanning tree.

[0055] The edge collection All edges in are sorted from small to large according to weight:

[0056] Traverse the sorted edge set , for each edge , check the vertex and Whether they belong to the same connected component (i.e., whether they are already connected); if and If they are not in the same connected component, the edge Add to , merge vertices and The connected component where . If and If they are already in the same connected component, skip the edge.

[0057] when The number of edges in When (that is, the tree contains Vertices and edges), stop the algorithm.

[0058] Step S142: Fusion strategy of clustering results and minimum spanning tree (MST).

[0059] Cluster consistency check: For each DBSCAN cluster , check its connectivity on MST, if If all user points of are connected in the MST, they are directly retained. If there are disconnected subsets in the memory, further split them.

[0060] Optimization of cross-cluster connections: Check the cross-cluster edges in the MST. If the weight of a cross-cluster edge is , indicating that the two clusters connected by this edge should be merged into one branch. , the clusters remain independent.

[0061] Merging of noise points: For noise points in DBSCAN, check their connections in MST. If the MST edge weight of the noise point and a branch is , merge it into this branch. Otherwise, keep it as an independent point and output it as an abnormal user.

[0062] Step S15: perform branch relationship verification and output the final branch division result, which includes normal branches and abnormal users.

[0063] Specifically, through DBSCAN clustering and MST fusion, the final branch division results are output, and each branch represents a group of associated users on the physical topology of the substation area. Noise points or independent user points are regarded as abnormal users for subsequent on-site verification or data correction.

[0064] In order to verify the advantages of the method provided by the present invention, a specific example will be described below.

[0065] In a community in a selected city (such as Nanjing), the power consumption data of 17 branches in 5 substations were selected as experimental data to verify this method. The experimental data is the voltage change time record and 96-point voltage data of 17 branches within 15 consecutive days of a certain month. The collection method of 96-point voltage data is: the initial sampling time of each user is the same and the sampling is performed every 15 minutes. Each user can obtain a total of 96 time points of voltage effective value data every day.

[0066] Table 1 Test results of branch relationship recognition algorithm

[0067] The branch relationship recognition was performed on the 17 branches in the 5 stations included in the test set, and compared with the prior topological structure, the branch relationship recognition accuracy was 96.77%. As can be seen from Table 1, this method can accurately realize the recognition of branch relationships, and has good results over multiple days with high stability.

[0068] like Figure 2 FIG. 1 is a schematic diagram showing a structure of an embodiment of a low voltage area branch relationship verification system based on DBSCAN clustering and minimum spanning tree provided by the present invention. Figures 3 to 7 As shown, in this embodiment, the low-voltage area branch relationship verification system 1 based on DBSCAN clustering and minimum spanning tree at least includes: The data acquisition module 10 is used to obtain the current and voltage data of each user in the same substation area, and record the matching number between the effective values ​​of current and voltage across phases; The association matrix construction module 11 is used to normalize the association strength according to the matching number of records and construct the voltage and current change association matrix between users; A distance matrix construction module 12, used to construct a distance matrix based on the Pearson correlation coefficient according to the association matrix; The cluster analysis module 13 is used to perform cluster analysis on users using the DBSCAN clustering method to obtain preliminary user branches; The clustering fusion module 14 is used to fuse the DBSCAN clustering results with the minimum spanning tree to optimize the user branch division; The output module 15 is used to check the branch relationship and output the final branch division result, which includes normal branches and abnormal users.

[0069] More specifically, as shown in 3, in a specific example, the data acquisition module 10 further includes: The high-frequency recording unit 100 is used to extract and record the corresponding time when the absolute value of the voltage effective value decreases and increases within a preset time is greater than or equal to a preset voltage threshold; and to extract and record the corresponding time when the absolute value of the current effective value decreases and increases within a preset time is greater than or equal to a preset current threshold; The screening unit 101 is used to screen the records of current increase and voltage increase of each phase between all different users and between the same user according to the recorded three-phase voltage and current of all users.

[0070] More specifically, as shown in 4, in a specific example, the association matrix construction module 11 includes: A matrix construction unit 110, configured to construct a matrix according to the number of users who need to perform branch relationship verification, wherein all initial elements of the matrix are 0; A numbering unit 111 is used to assign different numbers to the users to be checked; A statistical unit 112, used to count the intersection of the time points of the cross-phase user current rise and the user voltage rise, and calculate the intersection of the two sets; The assignment unit 113 is used to measure the similarity between user feature vectors using the Pearson correlation coefficient, convert the Pearson correlation coefficient into a distance, and assign values ​​to elements in the current and voltage change association matrix according to the records.

[0071] More specifically, as shown in 5, in a specific example, the distance matrix construction module 12 further includes: A feature vector obtaining unit 120 is used to define the feature vector of each user as the corresponding row of the user in the association matrix according to the association matrix, and obtain a feature vector representing the distribution of the association strength between the user and all other users; A coefficient calculation unit 121, used to calculate the Pearson correlation coefficient according to the feature vector to measure the similarity between the feature vectors of different users; The distance calculation unit 122 is used to convert the Pearson correlation coefficient into a distance and construct a symmetric distance matrix.

[0072] More specifically, as shown in FIG6 , in a specific example, the cluster analysis module 13 includes: A clustering unit 130 is used to perform clustering analysis using a density-based clustering method DBSCAN based on the constructed distance matrix between users; The parameter setting unit 131 is used to set the DBSCAN clustering parameters, including the domain radius and the minimum number of cores; wherein the method for selecting the domain radius is: sorting the kth nearest neighbor distance of the distance matrix, finding the natural jump point of the distance according to the "inflection point method", and setting it as the domain radius; the minimum number of cores is adjusted according to the user scale and data distribution of the station area; The clustering execution unit 132 is used to execute the DBSCAN clustering process, mark all user points as unvisited, define core points, boundary points and noise points, check the neighborhood, and output the clustering results.

[0073] More specifically, as shown in FIG7 , in a specific example, the clustering fusion module 14 includes: The minimum spanning tree construction unit 140 is used to construct a minimum spanning tree based on the distance matrix, including: creating an empty edge set and vertex set, sorting all edges in the edge set in ascending order according to weight, traversing the sorted edge set, and for each edge, checking whether the vertex belongs to the same connected component, if not, adding the edge to the edge set and merging the connected components where the vertex is located, and stopping the algorithm when the number of edges reaches the number of user points minus one, thereby generating a minimum spanning tree; The checking unit 141 is used to merge the clustering result with the minimum spanning tree, including: performing clustering consistency check, optimizing cross-clustering connections, and merging noise points to determine abnormal users.

[0074] For more details, please refer to and combine the above Figure 1 The description is not repeated here.

[0075] As another aspect of the present invention, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following is implemented: Figure 1 For more details, please refer to and combine the above Figure 1 The description is not repeated here.

[0076] The implementation of the embodiments of the present invention has the following beneficial effects: The present invention provides a low-voltage distribution area branch relationship verification method and system, which can achieve high-precision identification by relying on existing data without additional power signal injection through an analysis method based on user voltage change time records. The algorithm is designed to be efficient and has a short calculation cycle, and can quickly complete the branch relationship identification task to meet the real-time requirements of dynamic changes in the low-voltage distribution network.

[0077] This method is based on the voltage data analysis in the power consumption information collection system, and does not require the installation of additional hardware equipment, such as sensors or monitoring units. Compared with traditional methods, the present invention significantly reduces the equipment installation and operation and maintenance costs required for distribution network transparency, providing an economical and efficient solution for large-scale engineering applications.

[0078] The present invention optimizes the operation and management mode of the low-voltage distribution network through accurate branch relationship identification, and improves the level of digitalization and transparency of the distribution network. At the same time, it provides important technical support for the intelligent operation and maintenance of the low-voltage distribution network, and helps to improve the power supply reliability and management efficiency of the distribution network.

[0079] The present invention can handle complex tree or ring network structures in low-voltage distribution networks, and has strong robustness against data noise and dynamic changes in branch relationships. In practical applications, the recognition effect is stable and adaptable, and can meet the branch relationship recognition needs in different scenarios.

[0080] The present invention has shown good performance in practical applications and is suitable for large-scale promotion. Its core algorithm is simple and efficient, has both theoretical feasibility and engineering practicality, and does not require high equipment investment. It has broad commercial application prospects and can significantly improve the intelligence level of low-voltage distribution networks.

[0081] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, units, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0082] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A unit of functionality specified in a box or multiple boxes.

[0083] The above disclosure is only a preferred embodiment of the present invention, which certainly cannot be used to limit the scope of the present invention. Therefore, equivalent changes made according to the claims of the present invention are still within the scope of the present invention.

Claims

1. A low voltage area branch relationship verification method based on DBSCAN clustering and minimum spanning tree technology, characterized in that: The following steps are involved: Obtain the current and voltage data of each user in the same substation, and record the matching number of the effective values ​​of current and voltage across phases; According to the matching number of records, the correlation strength is normalized and the voltage and current change correlation matrix between users is constructed; According to the association matrix, construct a distance matrix based on the Pearson correlation coefficient; Use DBSCAN clustering method to cluster users and get preliminary user branches; Combine the DBSCAN clustering results with the minimum spanning tree to optimize user branch division; Perform branch relationship verification and output the final branch division result, which includes normal branches and abnormal users.

2. The method according to claim 1, characterized in that The step of obtaining the current and voltage data of each user in the same substation and recording the number of cross-phase matches between the effective values ​​of the current and voltage includes: Extract and record the corresponding time when the absolute value of the voltage effective value decreases and increases within the preset time is greater than or equal to the preset voltage threshold; and extract and record the corresponding time when the absolute value of the current effective value decreases and increases within the preset time is greater than or equal to the preset current threshold; Based on the recorded three-phase voltage and current of all users, the records of current increase and voltage increase of each phase between different users and between the same users are screened.

3. The method according to claim 2, characterized in that The normalized correlation strength is used to construct a voltage and current change correlation matrix between users, including: According to the number of users who need to check the branch relationship, a matrix is ​​constructed, and all the initial elements of the matrix are 0; Assign different numbers to the users to be verified; Count the intersection of the time points of the cross-phase user current rise and the user voltage rise, and calculate the intersection of the two sets; Assign values ​​to the elements in the current and voltage change correlation matrix based on the records.

4. The method according to claim 3, characterized in that The step of constructing a distance matrix based on the Pearson correlation coefficient according to the association matrix further comprises: According to the association matrix, the characteristic vector of each user is defined as the corresponding row of the user in the association matrix, and the characteristic vector representing the distribution of the correlation strength of the voltage and current changes of the user and all other users is obtained; Calculating the Pearson correlation coefficient based on the feature vectors to measure the similarity between feature vectors of different users; Convert the Pearson correlation coefficient to distance and construct a symmetric distance matrix.

5. The method according to claim 4, characterized in that The DBSCAN clustering method is used to perform cluster analysis on users to obtain preliminary user branches, including: Based on the constructed distance matrix between users, the density-based clustering method DBSCAN is used for clustering analysis; Set DBSCAN clustering parameters, including domain radius and minimum number of cores. The method for selecting the domain radius is as follows: sort the kth nearest neighbor distance of the distance matrix, find the natural jump point of the distance according to the "inflection point method", and set it as the domain radius. The minimum number of cores is adjusted according to the user scale and data distribution of the station area. Execute the DBSCAN clustering process, mark all user points as unvisited, define core points, boundary points, and noise points, check the neighborhood, and output the clustering results.

6. The method according to claim 4, characterized in that The method of fusing the DBSCAN clustering result with the minimum spanning tree to optimize the user branch division includes: Constructing a minimum spanning tree based on a distance matrix includes: creating an empty edge set and vertex set, sorting all edges in the edge set according to weights from small to large, traversing the sorted edge set, and for each edge, checking whether the vertex belongs to the same connected component, if not, adding the edge to the edge set and merging the connected components where the vertex is located, stopping the algorithm when the number of edges reaches the number of user points minus one, thereby generating a minimum spanning tree; The clustering results are integrated with the minimum spanning tree, including: clustering consistency check, cross-clustering connection optimization, and noise point merging to identify abnormal users.

7. A low voltage area branch relationship verification system, characterized in that: include: The data acquisition module is used to obtain the current and voltage data of each user in the same substation area, and record the matching number between the effective values ​​of current and voltage across phases; The correlation matrix construction module is used to normalize the correlation strength according to the matching number of records and construct the voltage and current change correlation matrix between users; A distance matrix construction module, used to construct a distance matrix based on the Pearson correlation coefficient according to the association matrix; Cluster analysis module, used to perform cluster analysis on users using DBSCAN clustering method to obtain preliminary user branches; Clustering fusion module, used to fuse DBSCAN clustering results with the minimum spanning tree to optimize user branch division; The output module is used to perform branch relationship verification and output the final branch division result, which includes normal branches and abnormal users.

8. The system according to claim 7, characterized in that The data acquisition module further comprises: A high-frequency wave recording unit, used to extract and record the corresponding moments when the absolute value of the voltage effective value decreases and increases within a preset time is greater than or equal to a preset voltage threshold; and to extract and record the corresponding moments when the absolute value of the current effective value decreases and increases within a preset time is greater than or equal to a preset current threshold; The screening unit is used to screen the records of current rise and voltage rise of each phase between all different users and between the same user according to the recorded three-phase voltage and current of all users.

9. The system according to claim 8, characterized in that The association matrix construction module further comprises: A matrix construction unit is used to construct a matrix according to the number of users who need to perform branch relationship verification, and all initial elements of the matrix are 0; A numbering unit is used to assign different numbers to the users to be verified; A statistical unit, used to count the intersection of the time points of the cross-phase user current rise and the user voltage rise, and calculate the intersection of the two sets; The assignment unit is used to measure the similarity between user feature vectors using the Pearson correlation coefficient, convert the Pearson correlation coefficient into a distance, and assign values ​​to the elements in the current and voltage change association matrix according to the records.

10. The system according to claim 9, characterized in that The distance matrix construction module further comprises: A feature vector obtaining unit, used to define the feature vector of each user as the corresponding row of the user in the association matrix according to the association matrix, and obtain a feature vector representing the distribution of the association strength of the voltage and current changes of the user with all other users; A coefficient calculation unit, used to calculate the Pearson correlation coefficient according to the feature vector to measure the similarity between the feature vectors of different users; The distance calculation unit is used to convert the Pearson correlation coefficient into a distance and construct a symmetric distance matrix.

11. The system according to claim 10, characterized in that The cluster analysis module further comprises: A clustering unit, used for performing clustering analysis using a density-based clustering method DBSCAN based on the constructed distance matrix between users; The parameter setting unit is used to set the DBSCAN clustering parameters, including the domain radius and the minimum number of cores. The method for selecting the domain radius is as follows: sorting the kth nearest neighbor distance of the distance matrix, finding the natural jump point of the distance according to the "inflection point method", and setting it as the domain radius. The minimum number of cores is adjusted according to the user scale and data distribution of the station area. The clustering execution unit is used to execute the DBSCAN clustering process, mark all user points as unvisited, define core points, boundary points and noise points, check the neighborhood, and output the clustering results.

12. The system according to claim 11, characterized in that The clustering fusion module does not include: A minimum spanning tree construction unit is used to construct a minimum spanning tree based on a distance matrix, including: creating an empty edge set and vertex set, sorting all edges in the edge set in ascending order according to weight, traversing the sorted edge set, and for each edge, checking whether the vertex belongs to the same connected component, if not, adding the edge to the edge set and merging the connected components where the vertex is located, and stopping the algorithm when the number of edges reaches the number of user points minus one, thereby generating a minimum spanning tree; The checking unit is used to fuse the clustering results with the minimum spanning tree, including: clustering consistency check, cross-clustering connection optimization, and noise point merging to determine abnormal users.

Images