Low-voltage distribution network topological relation identification method and system based on data driving

Abstract

The invention discloses a low-voltage power distribution network topological relation identification method and system based on data driving, and relates to the technical field of power distribution, and the method comprises the steps: obtaining the three-phase voltage data of each power distribution area general meter and the voltage data of each user meter, and carrying out the data decentralization processing, and obtaining an initial voltage data set; calculating the distance between two sample points in the initial voltage data set in the distance matrix, and performing isolated point removal and data deduplication processing for a plurality of times to obtain a preprocessed voltage data set; initializing clustering parameters based on the phase number of the transformer and outlet voltage data, and selecting an initial clustering center; performing clustering analysis on the preprocessed voltage data set by using an improved K-means algorithm to obtain a final clustering result meeting clustering criterion function convergence; determining a specific phase connection relationship between the user table and the transformer area general table according to the final clustering result; the identification accuracy of the topological relation of the low-voltage power distribution network is improved.

Description

Technical Field

[0001] The present application relates to the technical field of power distribution, and in particular, to a method and system for identifying the topological relationship of a low-voltage distribution network based on data driving. Background Art

[0002] As the terminal link of power distribution network power supply, the low-voltage distribution substation area has a complex structure and diverse load types; with the rapid development of social economy and the continuous growth of residents' electricity demand, the power facility structure of the low-voltage distribution substation area frequently changes and expands. However, due to factors such as irregular supervision and lack of professional technology, it is difficult for end-users to accurately register with the distribution transformers under the control of the substation area, and it is difficult to achieve intelligent control of the entire substation area.

[0003] The existing methods for identifying the topological relationship of a low-voltage distribution network mainly include traditional methods and methods based on similarity analysis. Among them, the traditional methods include the matrix method and the tree search method. Due to the complex line structure and large amount of data in the low-voltage distribution network, the matrix method and the tree search method have a large amount of calculation and are not applicable; the method based on similarity analysis identifies the connection relationship by measuring the similarity between voltage data. However, due to the small fluctuation difference of voltage data between different nodes, it is difficult to obtain an accurate topological relationship only relying on the similarity of voltage curves.

[0004] In view of the above related technologies, there are defects in the low accuracy of identifying the topological relationship of the low-voltage distribution network, and it is urgent to improve. Summary of the Invention

[0005] In order to improve the accuracy of identifying the topological relationship of the low-voltage distribution network, the present application provides a method and system for identifying the topological relationship of the low-voltage distribution network based on data driving.

[0006] In a first aspect, the invention object of the present application is realized by adopting the following technical solution: A method for identifying the topological relationship of a low-voltage distribution network based on data driving includes: Obtaining the three-phase voltage data of the main meter of each distribution substation area and the voltage data of each user meter, and performing data decentralization processing to obtain an initial voltage data set; Calculating the distance between two sample points in the initial voltage data set in a distance matrix, and performing several times of outlier removal and data deduplication processing on the initial voltage data set to obtain a preprocessed voltage data set; Initializing clustering parameters based on the phase number and outlet voltage data of the transformer, and selecting an initial clustering center; Using an improved K-means algorithm to perform clustering analysis on the preprocessed voltage data set, identifying the household-transformer relationship of the substation area, and obtaining a final clustering result that satisfies the convergence of the clustering criterion function; Determining the specific phase connection relationship between the user meter and the main meter of the substation area according to the final clustering result.

[0007] By adopting the above technical solutions, to improve data accuracy, this application first performs data decentralization processing and calculates the distances between sample points in the distance matrix, which can effectively remove the noise and outliers (isolated points) in the voltage data. The process of data decentralization not only helps to reduce the interference factors in the original collected data, but also improves the accuracy and reliability of subsequent analysis. Since the traditional K-means algorithm is sensitive to the selection of the initial clustering center and is easily affected by isolated points, this application performs several times of isolated point removal and data deduplication processing on the initial voltage data set to ensure that the preprocessed voltage data set is more stable and reliable. Then, based on the number of phases of the transformer and the outlet voltage data, the clustering parameters are initialized and the initial clustering center is selected to more accurately reflect the topological structure in the actual power system, making the result of the clustering analysis more in line with the actual situation and improving the accuracy of the identification of the household-transformer relationship. At the same time, using the improved K-means algorithm (referring to the above-mentioned isolated point removal and optimization adjustment of the clustering center parameters for the K-means algorithm) to perform clustering analysis on the preprocessed voltage data set can effectively reduce the calculation time. At the same time, due to the adoption of a more reasonable initial clustering center selection strategy, the accuracy of the clustering result is further improved, which is conducive to shortening the total calculation time. Finally, according to the final clustering result, the specific phase connection relationship between the user meter and the substation master meter is determined, realizing the accurate identification of the household-transformer relationship in the low-voltage distribution network, which helps to improve the intelligent management and operation efficiency of the entire distribution system. The technical solution of this application is based on data decentralization, preprocessing and optimization of the clustering algorithm, which significantly improves the accuracy and efficiency of the identification of the household-transformer relationship in the low-voltage distribution network.

[0008] In a preferred example of this application: the isolated point removal and data deduplication processing include: Calculate the distances between each sample point and the rest of the sample points and store them in the distance matrix; Determine the number of isolated points, find the isolated point with the largest distance by calculating the sum of the distances between each point and other points, and remove the found several isolated points from the initial voltage data set to obtain a new voltage data set; After the initial removal of isolated points is completed, recalculate the distances between the sample points in the new voltage data set and store them in a new distance matrix, and find the two sample points with the largest distance from the new distance matrix; and determine the data sample point center and the distance radius based on the maximum distance; Calculate the distance between each sample point in the new voltage data set and the data sample point center again, and remove the sample points whose distances are greater than the distance radius from the new voltage data set to obtain a first voltage data set; store several sample points whose distances are greater than the distance radius in a second voltage data set; The preprocessed voltage dataset is obtained based on the first voltage dataset and the second voltage dataset.

[0009] By adopting the above technical solution, two independent point removal methods of comprehensive distance screening and dynamic radius filtering are used to effectively eliminate noise points and edge outliers in the voltage data, greatly improving the data purity; the data deduplication process is also beneficial to eliminating redundant samples and reducing the subsequent clustering calculation complexity.

[0010] In a preferred example of the present application: the clustering parameters are initialized based on the number of phases of the transformer and the outlet voltage data, and the initial clustering centers are selected, specifically including: Determine the number of clusters based on the number of phases of the transformer, and select the number of clusters to be three times the total number of phases of the transformer; Take the outlet voltage data of each phase of the transformer as the initial clustering centers and store them in the initial clustering center matrix, so that the voltage data of each phase of different transformers belong to different classes.

[0011] By adopting the above technical solution, when the present application initializes and optimizes the clustering parameters, it will determine the number of clusters to be 3N based on the number of phases of the transformer (N), ensuring that the voltage data of each transformer phase forms an independent cluster, conforming to the physical topology law of the distribution network. At the same time, the outlet voltage of the transformer is used as the initial clustering center, making the clustering result highly consistent with the actual electrical connection characteristics, which is beneficial to improving the convergence speed of the clustering algorithm. The improved K-means algorithm of the present application has excellent performance and a high accuracy rate for identifying phase connection relationships.

[0012] In a preferred example of the present application: the improved K-means algorithm includes the following steps: Divide the first voltage dataset into several subsets, and each subset corresponds to an initial clustering center in the initial clustering center matrix; Perform clustering analysis using the selected initial clustering centers until the clustering centers no longer change or reach a predetermined number of iterations, and obtain the final clustering result that satisfies the convergence of the clustering criterion function; Calculate the distance between each sample point in the second voltage dataset and each initial clustering center, and reassign the sample points in the second voltage dataset to the clusters according to the principle of the nearest distance to generate the final topology relationship matrix and obtain the final clustering result; According to the final clustering result, determine the specific phase connection relationship between each user meter and the distribution substation main meter, and obtain the recognition result of the household-transformer relationship in the low-voltage distribution network.

[0013] By adopting the above technical solution, the first voltage data set is divided into multiple subsets, and an initial clustering center is assigned to each subset, making the clustering process more refined and accurate. Using the selected initial clustering centers for clustering analysis until the clustering centers no longer change or reach a predetermined number of iterations ensures the stability and accuracy of the final clustering result; calculating the distance between each sample point in the second voltage data set and each initial clustering center, and reassigning them to the clusters according to the principle of the closest distance, generating the final topological relationship matrix. Based on the dynamic adjustment mechanism, the dynamic adjustment of the sample point allocation method can more accurately reflect the connection relationship in the actual power system.

[0014] In a preferred example of this application: The determination of the data sample point center and the distance radius based on the maximum distance includes: finding two sample points A and B with the largest distance from the new distance matrix, and the corresponding maximum distance is maxinD. Calculate the data sample point center center and the radius distance (r = maxinD / 2) according to the following formula; Set the sample point object xi(xi1, xi2,..., xip) T , xj(xj1, xj2,..., xjp) T The distance between them is represented by the Euclidean distance d(x i , x j ) as follows: Define the data sample point center center(x i , x j ):

[0015] By adopting the above technical solution, finding two sample points A and B with the largest distance from the new distance matrix again, and calculating the data sample point center center and the radius r according to the maximum distance can effectively eliminate the influence of isolated points, improve the stability of the clustering result, and use the Euclidean distance to measure the distance between sample points, ensuring the consistency and standardization of distance calculation, which helps to improve the generality and applicability of the clustering algorithm.

[0016] In a preferred example of this application: The method further includes: In the data acquisition stage, continuously sample the user-side voltage at a predetermined frequency; When performing data decentralization processing, subtract the sample average value from each voltage sequence sample value; And / or, The improved K-means algorithm further includes: Adapting the clustering weight ω according to the sample point density i , and the formula is: ωi = 1 / (1 + α × d) i (3); Where i is the identifier of the sample point in the voltage dataset, α is the adjustment coefficient, and d i is the sample point density; Divide the preprocessed voltage dataset into multiple subspaces, and perform clustering iteration in parallel using the MapReduce mode. The formula is: Where represents the final set of cluster centers obtained by the K-means algorithm after (t + 1) iterations; k is the total data of the clusters, which is three times the total number of phases of the transformer; t is the number of iterations; C k is a set containing k cluster centers, and each cluster center c j is a data point; m is the number of parallel threads; argmin cj represents selecting the cluster center that minimizes the objective function among all c j ; S j is the j-th subset among the m subspaces into which the preprocessed voltage dataset is divided.

[0017] By adopting the above technical solution, decentralization processing is performed in the data collection stage, which is beneficial to reducing data deviation. The clustering weight is adaptively adjusted according to the sample point density, and formula (3) can better process datasets with different densities, improving the rationality of the clustering results; clustering iteration is performed in parallel using the MapReduce mode, and formula (4) improves the calculation efficiency under large-scale datasets.

[0018] In a preferred example of the present application: The improved K-means algorithm further includes real-time constraint processing and adaptive iteration termination, specifically including; In the data preprocessing stage, set the sampling frequency f max and the upper limit of calculation delay T lim , and dynamically adjust the data processing batch through a sliding window to satisfy: N batc h / f max ≤ T lim Where N batch is the sample size for single processing; When the movement amount of the cluster center in two consecutive iterations ΔC < ∈1 and the change rate of the objective function ΔJ / Δt < ∈2, terminate the iteration in advance, where ∈1 and ∈2 are dynamic thresholds related to the data volume N: Where σ is the global data standard deviation, μ is the average calculation time-consuming; γ and δ are empirical coefficients.

[0019] By adopting the above technical solution, the sampling frequency and the upper limit of the calculation delay are set, and the data processing batches are dynamically adjusted through a sliding window to ensure the timeliness and efficiency of data processing. It meets the real-time requirements and is especially suitable for scenarios that require quick response; when the movement amount of the cluster center in two consecutive iterations is less than the threshold and the change rate of the objective function is lower than the dynamic threshold, the iteration is terminated in advance, reducing unnecessary calculations and improving the overall efficiency of the algorithm. Among them, the dynamic threshold is related to the data volume and can be automatically adjusted according to the actual situation, ensuring the flexibility and robustness of the algorithm.

[0020] In a second aspect, the invention object of the present application is achieved by adopting the following technical solution: A data-driven low-voltage distribution network topology relationship identification system for executing a data-driven low-voltage distribution network topology relationship identification method as described above, the system includes: A data acquisition module for acquiring three-phase voltage data of the main meters of each distribution substation area and voltage data of each user meter; A data preprocessing module connected to the data acquisition module for performing decentralization processing on the acquired voltage data, calculating the distance between two sample points in the initial voltage data set, and obtaining a preprocessed voltage data set through several outlier removal and data deduplication processes; A parameter initialization module connected to the data preprocessing module for initializing clustering parameters according to the number of phases of the transformer and the outlet voltage data and selecting an initial clustering center; A clustering analysis module connected to the parameter initialization module for performing clustering analysis on the preprocessed voltage data set by using an improved K-means algorithm, identifying the household-transformer relationship in the substation area, and obtaining a final clustering result that satisfies the convergence of the clustering criterion function; a relationship determination module connected to the clustering analysis module for determining the specific phase connection relationship between the user meter and the main meter of the substation area according to the final clustering result.

[0021] By adopting the above technical solutions, the data acquisition module includes a smart meter installed on the user side. Through the extensive application of the smart meter and the user information acquisition system, a large amount of detailed electrical data is provided, laying a data foundation for the identification of the topological relationship of the low-voltage distribution network; the data preprocessing module obtains a preprocessed voltage data set through decentralized processing, several outlier removals, and data deduplication, enhancing the stability and consistency of the data, contributing to improving the accuracy and stability of the clustering results, and reducing misjudgments caused by data fluctuations or anomalies; the parameter initialization module selects initial clustering centers to better reflect the topological structure in the actual power system, ensuring the rationality of the clustering analysis; the clustering analysis module uses an improved K-means algorithm to perform clustering analysis on the preprocessed voltage data set to identify the household-transformer relationship. The improved K-means algorithm is optimized on the basis of the traditional algorithm, solving the instability problem caused by the random selection of the initial clustering center; then the relationship determination module realizes the accurate identification of the household-transformer relationship in the low-voltage distribution network.

[0022] In the third aspect, the invention object of the present application is realized by adopting the following technical solutions: A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the above method for identifying the topological relationship of a low-voltage distribution network based on data driving are realized.

[0023] In the fourth aspect, the invention object of the present application is realized by adopting the following technical solutions: A computer-readable storage medium stores a computer program. When the computer program is executed by a processor, the steps of the above method for identifying the topological relationship of a low-voltage distribution network based on data driving are realized.

[0024] In summary, the present application includes at least one of the following beneficial technical effects: 1. By performing several outlier removals and data deduplication on the initial voltage data set, the present application ensures that the preprocessed voltage data set is more stable and reliable. Then, based on the phase number of the transformer and the outlet voltage data, the clustering parameters are initialized, and the initial clustering centers are selected to more accurately reflect the topological structure in the actual power system, making the results of the clustering analysis more in line with the actual situation and improving the accuracy of the household-transformer relationship identification; 2. Through the collaborative work of each module, not only the accuracy and efficiency of the identification of the topological relationship of the low-voltage distribution network are improved, but also the real-time performance and adaptability of the system are enhanced. Description of the Drawings

[0025] Figure 1It is a flowchart of a method for identifying the topological relationship of a low-voltage distribution network based on data driving in an embodiment of the present application; Figure 2 It is an application flowchart of one of the embodiments of a method for identifying the topological relationship of a low-voltage distribution network based on data driving in an embodiment of the present application; Figure 3 It is a schematic diagram of the device in an embodiment of the present application. Detailed implementation manners

[0026] The following further describes the present application in detail with reference to the accompanying drawings.

[0027] In one embodiment, as Figure 1 and Figure 2 shown, the present application discloses a method for identifying the topological relationship of a low-voltage distribution network based on data driving, which specifically includes the following steps: S1: Obtain the three-phase voltage data of the total meters of each distribution substation area and the voltage data of each user meter, and perform data decentralization processing to obtain an initial voltage data set.

[0028] In this embodiment, data acquisition refers to obtaining voltage data from the total meters and user meters of each distribution substation area through smart meters or data acquisition systems. In the data acquisition stage, the user voltage is continuously sampled at a predetermined frequency (for example, the sampling frequency is once per hour and the duration is 7 days); as Figure 2 shown is an application flowchart of one of the embodiments of the method for identifying the topological relationship of a low-voltage distribution network based on data driving.

[0029] Specifically, when performing data decentralization processing, subtract the sample average value from each voltage sequence sample value.

[0030] S2: Calculate the distance between two sample points in the initial voltage data set in the distance matrix, and perform several times of outlier removal and data deduplication processing on the initial voltage data set to obtain a preprocessed voltage data set.

[0031] In this embodiment, noise points and duplicate data are removed through two iterative screenings to construct a high-quality preprocessed data set; specifically, outlier removal and data deduplication processing include: S21: Calculate the distance between each sample point and the remaining sample points, and store it in the distance matrix.

[0032] In this embodiment, the distance matrix is an Euclidean distance matrix, which is used to quantify the similarity between sample points (also called data points).

[0033] S22: Determine the number of isolated points. Find the isolated point with the maximum distance by calculating the sum of the distances between each point and other points, and remove the found isolated points from the initial voltage dataset to obtain a new voltage dataset.

[0034] Specifically, determine the number of isolated points to be removed. Based on the distances calculated between each sample point and all other sample points, select the top M points with the maximum distances as the several isolated points to be removed (where M is the determined number of isolated points to be removed), and remove the M determined isolated points from the initial voltage dataset to generate a new voltage dataset. Then store the M initially removed isolated points in a preset second voltage dataset.

[0035] S23: After initially removing the isolated points, recalculate the distances between the sample points in the new voltage dataset and store them in a new distance matrix. Find the two sample points with the maximum distance from the new distance matrix; and determine the data sample point center and distance radius based on the maximum distance.

[0036] Specifically, determining the data sample point center and distance radius based on the maximum distance includes: Find two sample points A and B with the maximum distance from the new distance matrix, and the corresponding maximum distance is maxinD. Calculate the data sample point center center and radius distance (r = maxinD / 2) according to the following formula; Set the sample point object xi(xi1, xi2,..., xip) T and xj(xj1, xj2,..., xjp) T The distance between them is represented by the Euclidean distance d(x i , x j ) as follows: Define the data sample point center center(x i , x j ):

[0037] S24: Calculate the distance between each sample point in the new voltage dataset and the data sample point center again. Remove the sample points whose distances are greater than the distance radius from the new voltage dataset to obtain the first voltage dataset; store the several sample points whose distances are greater than the distance radius in the second voltage dataset.

[0038] In this embodiment, move the sample points whose distances are greater than the distance radius into the second voltage dataset and remove them from the first voltage dataset for the second time to obtain a preprocessed voltage dataset.

[0039] S25: Obtain a preprocessed voltage dataset based on the first voltage dataset and the second voltage dataset.

[0040] Specifically, the data deduplication process includes using the unique function to detect duplicate sample points (points with an Euclidean distance of 0), retaining the first-occurring sample points, and removing the remaining duplicate points.

[0041] S3: Initialize the clustering parameters based on the number of phases and the outlet voltage data of the transformer, and select the initial clustering centers.

[0042] In this embodiment, parameter initialization refers to determining the clustering parameters and the initial clustering centers according to the number of phases and the outlet voltage of the transformer; since the number of phases and the outlet voltage data of the transformer are known, this information can be used to initialize the clustering parameters, and this initialization method ensures that the voltage data of each phase of different transformers are correctly assigned to different categories.

[0043] Specifically, step S3 includes: S31: Determine the number of clusters based on the number of phases of the transformer, and select the number of clusters to be three times the total number of phases of the transformer.

[0044] S32: Use the outlet voltage data of each phase of the transformer as the initial clustering centers and store them in the initial clustering center matrix, so that the voltage data of each phase of different transformers belong to different classes.

[0045] In this embodiment, the number of phases of the transformer is automatically identified through the three-phase voltage amplitude and phase difference of the master table (for example, a Y / Yn0 connection is identified as 3 phases).

[0046] S4: Use the improved K-means algorithm to perform clustering analysis on the preprocessed voltage data set, identify the relationship between the transformer and the household transformers in the substation area, and obtain the final clustering result that satisfies the convergence of the clustering criterion function.

[0047] In this embodiment, the improved K-means algorithm is obtained by optimizing the selection of the initial clustering centers and the handling of outliers in the traditional K-means algorithm; the improved K-means algorithm includes adaptively adjusting the clustering weights and performing clustering iterations in parallel to improve efficiency.

[0048] Specifically, the improved K-means algorithm includes the following steps: S41: Divide the first voltage data set into several subsets, and each subset corresponds to an initial clustering center in the initial clustering center matrix.

[0049] Specifically, based on the hash function, the data in the first voltage data set obtained after secondary outlier removal and deduplication are evenly distributed into m subspaces, and each subset contains sample points adjacent to the corresponding clustering center (such as an Euclidean distance less than the initial radius r).

[0050] S42: Perform clustering analysis using the selected initial cluster centers until the cluster centers no longer change or reach a predetermined number of iterations, obtaining the final clustering result that satisfies the convergence of the clustering criterion function.

[0051] Specifically, the improved K-means algorithm also includes real-time constraint processing and adaptive iteration termination, specifically including: in the data preprocessing stage, setting the sampling frequency f max and the upper limit of the calculation delay T lim , and dynamically adjusting the data processing batch through a sliding window to satisfy: N batc h / f max ≤T lim where N batch is the sample size for a single processing; when the movement amount of the cluster center for two consecutive iterations ΔC < ∈1 and the change rate of the objective function ΔJ / Δt < ∈2, terminate the iteration in advance, where ∈1 and ∈2 are dynamic thresholds related to the data volume N: where σ is the global data standard deviation, μ is the average calculation time-consuming; γ, δ are empirical coefficients.

[0052] In this embodiment, the change rate of the objective function ΔJ / J is the difference between the objective function value J(t) of the current iteration and the value of the previous iteration J(t - 1) divided by the value of the previous iteration, that is, (ΔJ) / J = (J(t) - J(t - 1)) / J(t - 1). This ratio reflects the change amplitude of the objective function value. If the change rate is very small, it indicates that the algorithm has converged.

[0053] S43: Calculate the distance between each sample point in the second voltage dataset and each initial cluster center, reassign the sample points in the second voltage dataset to the clusters according to the principle of the nearest distance, generate the final topological relationship matrix, and obtain the final clustering result.

[0054] S44: According to the final clustering result, determine the specific phase connection relationship between each user table and the total table of the distribution transformer area, and obtain the identification result of the household-transformer relationship of the low-voltage distribution network.

[0055] Specifically, map the final clustering result to the physical phase to generate a household-transformer relationship matrix: specifically including phase label mapping, conflict detection, and topological graph generation. Phase label mapping includes mapping the cluster label (such as Cluster_1) to the transformer phase sequence (such as Phase A); conflict detection includes triggering an alarm and manual review if the same user table is assigned to multiple phases; topological graph generation includes outputting an adjacency matrix containing phase information and being compatible with the grid dispatching system interface to obtain a household-transformer relationship identification result with a relatively high accuracy of the phase connection relationship and the household-transformer connection relationship.

[0056] S5: Determine the specific phase connection relationship between the user table and the substation total table according to the final clustering result.

[0057] In this embodiment, since each cluster corresponds to a specific phase, the specific phase connection relationship is confirmed by comparing the voltage data characteristics within the cluster; analyze the final clustering result, determine which cluster each user table belongs to, and determine the specific phase connection relationship between the user table and the substation total table according to the identifier of the cluster (such as phase A, phase B, or phase C), and output the identification result of the user-transformer relationship in the low-voltage distribution network.

[0058] In one embodiment, the improved K-means algorithm further includes: Adaptive adjustment of the clustering weight ω according to the sample point density i , the formula is: ω i = 1 / (1 + α × d) i (3); In the formula, i is the identifier of the sample point in the voltage dataset, α is the adjustment coefficient, and d i is the sample point density; Divide the preprocessed voltage dataset into multiple subspaces, and perform clustering iterations in parallel using the MapReduce mode. The formula is: In the formula, represents the set of final cluster centers obtained by the K-means algorithm after (t + 1) iterations; k is the total data of the clusters, which is three times the total number of phases of the transformer; t is the number of iterations; C k is a set containing k cluster centers, and each cluster center c j is a data point; m is the number of parallel threads; argmin cj represents selecting the clustering center that minimizes the objective function from all c j ; S j is the j-th subset among the m subspaces into which the preprocessed voltage dataset is divided.

[0059] It should be understood that the sequence numbers of the steps in the above embodiments do 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 to the implementation process of the embodiments of this application.

[0060] In one embodiment, a data-driven low-voltage distribution network topology relationship identification system is provided. This data-driven low-voltage distribution network topology relationship identification system corresponds to the data-driven low-voltage distribution network topology relationship identification method in the above embodiment.

[0061] A data-driven low-voltage distribution network topology relationship identification system, including a data acquisition module, a data preprocessing module, a parameter initialization module, a clustering analysis module, and a relationship determination module. The detailed descriptions of each functional module are as follows: The data acquisition module is used to obtain the three-phase voltage data of the main meters of each distribution area and the voltage data of each user meter; The data preprocessing module is connected to the data acquisition module, and is used to perform decentralization processing on the acquired voltage data, calculate the distance between two sample points in the initial voltage data set, and obtain a preprocessed voltage data set through several outlier removal and data deduplication processes; The parameter initialization module is connected to the data preprocessing module, and is used to initialize the clustering parameters according to the number of phases of the transformer and the outlet voltage data, and select the initial clustering center; The clustering analysis module is connected to the parameter initialization module, and is used to perform clustering analysis on the preprocessed voltage data set by using an improved K-means algorithm, identify the household-transformer relationship in the area, and obtain the final clustering result that satisfies the convergence of the clustering criterion function; The relationship determination module is connected to the clustering analysis module, and is used to determine the specific phase connection relationship between the user meter and the main meter of the area according to the final clustering result.

[0062] For the specific limitations of a data-driven low-voltage distribution network topology relationship identification system, reference can be made to the limitations of a data-driven low-voltage distribution network topology relationship identification method in the above text, which will not be elaborated here; each module in the above data-driven low-voltage distribution network topology relationship identification system can be implemented in whole or in part by software, hardware, and their combination; the above modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0063] In one embodiment, a computer device is provided. The computer device can be a server, and its internal structure diagram can be as Figure 3 shown. The computer device includes a processor, a memory, a network interface, and a database connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store the three-phase voltage data of the main meters of the radio area, the voltage data of each user meter, and the voltage data set. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, it implements a data-driven low-voltage distribution network topology relationship identification method.

[0064] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the following steps are implemented: S1: Obtain the three-phase voltage data of the main meters of each distribution transformer area and the voltage data of each user meter, and perform data decentralization processing to obtain an initial voltage data set.

[0065] S2: Calculate the distance between two sample points in the initial voltage data set in the distance matrix, and perform several outlier removal and data deduplication processes on the initial voltage data set to obtain a preprocessed voltage data set.

[0066] S3: Initialize the clustering parameters based on the number of phases of the transformer and the outlet voltage data, and select the initial clustering center.

[0067] S4: Use the improved K-means algorithm to perform clustering analysis on the preprocessed voltage data set, identify the relationship between the user transformers in the transformer area, and obtain the final clustering result that satisfies the convergence of the clustering criterion function.

[0068] S5: Determine the specific phase connection relationship between the user meter and the main meter of the transformer area according to the final clustering result.

[0069] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented: S1: Obtain the three-phase voltage data of the main meters of each distribution transformer area and the voltage data of each user meter, and perform data decentralization processing to obtain an initial voltage data set.

[0070] S2: Calculate the distance between two sample points in the initial voltage data set in the distance matrix, and perform several outlier removal and data deduplication processes on the initial voltage data set to obtain a preprocessed voltage data set.

[0071] S3: Initialize the clustering parameters based on the number of phases of the transformer and the outlet voltage data, and select the initial clustering center.

[0072] S4: Use the improved K-means algorithm to perform clustering analysis on the preprocessed voltage data set, identify the relationship between the user transformers in the transformer area, and obtain the final clustering result that satisfies the convergence of the clustering criterion function.

[0073] S5: Determine the specific phase connection relationship between the user meter and the main meter of the transformer area according to the final clustering result.

[0074] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, storage, database, or other medium used in the embodiments provided in the present application can include non-volatile and / or volatile memories. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and Rambus dynamic RAM (RDRAM), etc.

[0075] Those skilled in the art can clearly understand that, for the convenience and simplicity of description, only the above-mentioned division of each functional unit and module is used as an example. In actual applications, the above functions can be allocated to different functional units and modules according to needs, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0076] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A data-driven method for identifying the topological relationship of a low-voltage distribution network, characterized in that Including: Obtain the three-phase voltage data of the total meter of each distribution transformer area and the voltage data of each user meter, and perform data decentralization processing to obtain an initial voltage data set; Calculate the distance between two sample points in the initial voltage data set in the distance matrix, and perform several outlier removal and data deduplication processes on the initial voltage data set to obtain a preprocessed voltage data set; Initialize the clustering parameters based on the number of phases of the transformer and the outlet voltage data, and select the initial clustering center; Use the improved K-means algorithm to perform clustering analysis on the preprocessed voltage data set, identify the relationship between the user transformer and the area transformer, and obtain the final clustering result that satisfies the convergence of the clustering criterion function; Determine the specific phase connection relationship between the user meter and the total meter of the transformer area according to the final clustering result.

2. The method for identifying the topological relationship of a low-voltage distribution network based on data driving according to claim 1, characterized in that, The outlier removal and data deduplication process includes: Calculate the distance between each sample point and the remaining sample points, and store it in the distance matrix; Determine the number of outliers, find the outlier with the largest distance by calculating the sum of the distances between each point and other points, and remove the found several outliers from the initial voltage data set to obtain a new voltage data set; After the initial outlier removal is completed, recalculate the distance between the sample points in the new voltage data set and store it in the new distance matrix, and find the two sample points with the largest distance from the new distance matrix; and determine the data sample point center and distance radius based on the maximum distance; Calculate the distance between each sample point in the new voltage data set and the data sample point center again, and remove the sample points greater than the distance radius from the new voltage data set to obtain a first voltage data set; store several sample points greater than the distance radius in a second voltage data set; Obtain the preprocessed voltage data set based on the first voltage data set and the second voltage data set.

3. A data-driven method for identifying the topological relationship of a low-voltage distribution network according to claim 2, characterized in that, The initializing the clustering parameters based on the number of phases of the transformer and the outlet voltage data and selecting the initial clustering center specifically includes: Determine the number of clusters based on the number of phases of the transformer, and select the number of clusters to be three times the total number of phases of the transformer; Use the outlet voltage data of each phase of the transformer as the initial clustering center and store it in the initial clustering center matrix, so that the voltage data of different phases of different transformers belong to different classes.

4. A method for identifying the topological relationship of a low-voltage distribution network based on data-driven, as claimed in claim 3, wherein The improved K-means algorithm includes the following steps: Divide the first voltage data set into several subsets, and each subset corresponds to an initial clustering center in the initial clustering center matrix; Use the selected initial clustering center to perform clustering analysis until the clustering center no longer changes or reaches a predetermined number of iterations, and obtain the final clustering result that satisfies the convergence of the clustering criterion function; Calculate the distance between each sample point in the second voltage data set and each initial clustering center, and reassign the sample points in the second voltage data set to the clusters according to the principle of the nearest distance to generate the final topology relationship matrix and obtain the final clustering result; According to the final clustering result, determine the specific phase connection relationship between each user meter and the total meter of the distribution transformer area, and obtain the recognition result of the user-transformer relationship of the low-voltage distribution network.

5. A method for identifying the topological relationship of a low-voltage distribution network based on data-driven according to claim 2, characterized in that The determining the data sample point center and distance radius based on the maximum distance includes: Find two sample points A and B with the largest distance from the new distance matrix, and the corresponding maximum distance is maxinD. Calculate the center center of the data sample points and the radius distance (r = maxinD / 2) according to the following formula; Set the sample point object \(x_i(x_{i1}, x_{i2}, \ldots, x_{ip})\) T and \(x_j(x_{j1}, x_{j2}, \ldots, x_{jp})\) T The distance between them is represented by the Euclidean distance \(d(x\) i , x j ) as follows: Define the center of the data sample point center(x i , x j ):

6. A data-driven method for identifying the topological relationship of a low-voltage distribution network according to claim 2, characterized in that, The method further includes: In the data acquisition stage, continuously sample the user voltage at a predetermined frequency; When performing data decentralization processing, subtract the sample average value from each voltage sequence sample value; And / or The improved K-means algorithm further includes: Adaptive adjustment of the clustering weight ω according to the sample point density i , the formula is: ω i = 1 / (1 + α×d) i (3); where i is the identifier of the sample point in the voltage dataset, α is the adjustment coefficient, and d i is the sample point density; Divide the preprocessed voltage data set into multiple subspaces, and perform clustering iteration in parallel using the MapReduce mode. The formula is: Wherein, represents the final set of cluster centers obtained by the K-means algorithm after (t+1) iterations; k is the total data of the clusters, which is three times the total number of phases of the transformer; t is the number of iterations; C k is a set containing k cluster centers, and each cluster center c j is a data point; m is the number of parallel threads; argmin cj represents selecting the clustering center that minimizes the objective function from all c j ; S j is the j-th subset among the m subspaces into which the preprocessed voltage data set is divided.

7. A data-driven method for identifying the topological relationship of a low-voltage distribution network according to claim 6, characterized in that The improved K-means algorithm further includes real-time constraint processing and adaptive iteration termination, specifically including; In the data preprocessing stage, set the sampling frequency f max and the upper limit T of the calculation delay lim , and dynamically adjust the data processing batch through a sliding window to satisfy: N batc h / f max ≤T lim where N batch is the sample size for single processing; When the movement amount ΔC of the cluster center in two consecutive iterations is <∈1 and the change rate ΔJ / Δt of the objective function is <∈2, terminate the iteration in advance, where ∈1 and ∈2 are dynamic thresholds related to the data volume N: ∈2 = δ × μ / T lim ; Where σ is the global data standard deviation, μ is the average calculation time; γ and δ are empirical coefficients.

8. A data-driven low-voltage distribution network topology relationship identification system, characterized in that For implementing a data-driven low-voltage distribution network topology relationship recognition method according to any one of claims 1-7, the system includes: A data acquisition module for obtaining three-phase voltage data of the total meter of each distribution transformer area and voltage data of each user meter; A data preprocessing module connected to the data acquisition module for performing decentralization processing on the acquired voltage data, calculating the distance between two sample points in the initial voltage data set, and obtaining a preprocessed voltage data set through several outlier removal and data deduplication processes; A parameter initialization module connected to the data preprocessing module for initializing clustering parameters according to the number of phases of the transformer and the outlet voltage data, and selecting an initial clustering center; A clustering analysis module connected to the parameter initialization module for performing clustering analysis on the preprocessed voltage data set using an improved K-means algorithm, identifying the household-transformer relationship in the transformer area, and obtaining a final clustering result that satisfies the convergence of the clustering criterion function; A relationship determination module connected to the clustering analysis module for determining the specific phase connection relationship between the user meter and the total meter of the transformer area according to the final clustering result.

9. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of a data-driven low-voltage distribution network topology relationship recognition method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of a data-driven low-voltage distribution network topology relationship recognition method according to any one of claims 1 to 7.

Images