A method and system for topology identification of low-voltage power supply networks

Abstract

This invention discloses a method and system for identifying the topology of a low-voltage power supply network, belonging to the field of power supply network technology. It includes: establishing a general structure of the low-voltage power supply network topology; dividing the meter nodes in the low-voltage power supply network topology into branch nodes and user nodes; obtaining the connection relationships of the branch nodes based on their power data; obtaining the connection positions of the user nodes based on their power data and the power data of the user nodes; correcting any abnormal results regarding the connection positions of the user nodes based on their voltage data and the voltage data of the user nodes; and obtaining the low-voltage power supply network topology structure. This method can accurately obtain the low-voltage power supply network topology structure, has strong applicability, and high reliability.

Description

Technical Field

[0001] This application relates to the field of power supply network technology, and in particular to a method and system for identifying the topology of a low-voltage power supply network. Background Technology

[0002] The statements in this section merely refer to the background art relevant to this application and do not necessarily constitute prior art.

[0003] Low-voltage power supply networks refer to the network from 10kV / 400V transformer substations to user meters, and are a crucial infrastructure supporting national economic and social development. The topology information of low-voltage power supply networks is essential for accurate load modeling, power network line loss calculation, fault location, and improving power supply reliability. For many years, the construction of low-voltage power supply networks in my country has suffered from problems such as inconsistent planning, large-scale renovation and expansion projects, and reliance on manual maintenance of topology information. This leads to frequent changes in the actual network topology, inconsistent with the structure maintained by the system, causing difficulties in transformer substation power supply management, outage location determination, and line loss calculation, and seriously affecting the reliability of power supply to users. Therefore, it is necessary to research an automatic power supply network topology identification method to achieve automatic identification and management of power supply network topology.

[0004] Currently, the most widely used method for power supply network topology identification is the signal injection method. This method involves installing signal sensing devices at the user's electricity meter and injecting voltage or current characteristic signals into the power supply area or upstream node. By analyzing the sensing results of these characteristic signals, the connection relationships between meters are determined, thus completing the topology identification. While the signal injection method has a clear principle and good power supply network topology identification capabilities, it is susceptible to interference, requires sophisticated signal processing, and necessitates additional signal injection and detection equipment, leading to increased costs and engineering workload.

[0005] Many scholars have conducted research on power supply network topology identification methods based on electricity meter measurement data, mainly including similarity algorithms, linear programming algorithms, cluster analysis, and artificial intelligence algorithms. The paper "Smart Meter Data Analytics for Distribution Network Connectivity Verification [J]. IEEE Transactions on Smart Grid" determines the topology by calculating the Pearson correlation coefficient of node voltage sequences. This method is simple and easy to implement, but relying solely on node voltage data has unreliable limitations. The paper "Identifying Topology of Low Voltage Distribution Networks Based on Smart Meter Data [J]. IEEE Transactions on Smart Grid" uses principal component analysis and energy conservation to establish linear relationships between adjacent nodes, identifying the network topology layer by layer. This method cannot accurately identify the network topology when the node hierarchical relationships are unknown or when users connect between nodes. The paper "A Method for Verifying the Topology of Low-Voltage Distribution Network Based on Discrete Fréchet Distance and Clipped Nearest Neighbor Method" uses the k-nearest neighbor clustering algorithm to classify low-voltage distribution areas and determine the distribution area to which a user belongs. This method is suitable for networks with simple topologies; its effectiveness in complex networks remains to be verified. The literature, including "Online Topology Identification of Smart Distribution Networks Based on LightGBM and DNN", "Distribution Network Connection Identification Technology Based on Integrated Deep Neural Network", "Structure Learning in Power Distribution Networks [J]. IEEE Transactions on Control of Network Systems.", and "A Data-Driven Parameter and Topology Joint Estimation Framework in Distribution Grids", uses artificial intelligence algorithms to identify the topology of power supply networks. Such methods require a large amount of data for learning, the algorithms are complex, and their adaptability to new topologies is generally limited. Summary of the Invention

[0006] In recent years, Advanced Metering Infrastructure (AMI) has developed rapidly. Smart meters, as terminal devices of AMI, undertake tasks such as collecting and uploading electricity consumption information. Smart meters can collect and upload multiple measurement data, including electrical energy, active power, voltage, and current. This data not only contains electricity consumption information for each node at various time points but also includes the topology information of the power supply network. There is a specific mathematical relationship between electricity consumption information and network topology. Therefore, by inferring the connection relationships between nodes based on the measurement data of smart meters, the topology of the power supply network can be obtained.

[0007] To address the shortcomings of existing technologies, this application provides a method, system, electronic device, and computer-readable storage medium for identifying low-voltage power supply network topology. By establishing a general structure for the low-voltage power supply network topology, the meter nodes in the topology are divided into two main categories: branch meter nodes and user meter nodes. First, wavelet transform is performed on the active power curve to extract features and determine the connection relationships of branch nodes. Then, a 0-1 integer quadratic programming method is used to determine the connection positions of user nodes in the network, completing the initial topology identification. Based on this, the identification results are checked and corrected using the hypothesis testing approach, ultimately obtaining an accurate network topology structure.

[0008] Firstly, this application provides a method for identifying the topology of a low-voltage power supply network;

[0009] A method for identifying the topology of a low-voltage power supply network, comprising:

[0010] A general structure for the low-voltage power supply network topology is established, and the meter nodes in the low-voltage power supply network topology are divided into branch nodes and user nodes;

[0011] Based on the power data of the branch nodes, obtain the connection relationship of the branch nodes;

[0012] Based on the power data of the branch node and the power data of the user node, the connection location of the user node is obtained;

[0013] Based on the voltage data of the branch nodes and the voltage data of the user nodes, the abnormal results of the user node connection positions are corrected, and the low-voltage power supply network topology is obtained.

[0014] Secondly, this application provides a low-voltage power supply network topology identification system;

[0015] A low-voltage power supply network topology identification system, comprising:

[0016] The initial topology acquisition module is used to establish a general structure of the low-voltage power supply network topology, and to divide the meter nodes in the low-voltage power supply network topology into branch nodes and user nodes.

[0017] The branch node connection relationship acquisition module acquires the branch node connection relationship based on the power data of the branch nodes;

[0018] The user node connection location acquisition module is used to acquire the user node connection location based on the power data of the branch node and the power data of the user node.

[0019] The low-voltage power supply network topology acquisition module is used to check and correct abnormal results of user node connection positions based on the voltage data of the branch nodes and the voltage data of the user nodes, and to acquire the low-voltage power supply network topology.

[0020] Thirdly, this application provides an electronic device;

[0021] An electronic device includes a memory and a processor, as well as computer instructions stored in the memory and running on the processor, wherein the computer instructions, when executed by the processor, perform the method described above.

[0022] Fourthly, this application provides a computer-readable storage medium;

[0023] A computer-readable storage medium for storing computer instructions, which, when executed by a processor, perform the above-described method.

[0024] Compared with the prior art, the beneficial effects of this application are:

[0025] 1. When judging the connection relationship of branch nodes, the idea of ​​wavelet transform is used to extract the power change feature of branch nodes, which amplifies the characteristics of branch nodes, makes branch nodes have good separability, and improves the accuracy of similarity judgment.

[0026] 2. When determining the connection location of user nodes, the method of 0-1 integer quadratic programming is used to overcome the influence of the high randomness of user electricity consumption behavior. The method is to check and correct the judgment results by using the idea of ​​hypothesis testing, thus ensuring the accuracy of the method.

[0027] 3. Without losing generality, it addresses complex topologies by employing different identification methods for branch nodes and user nodes, ensuring the applicability and reliability of the methods. Automatic identification of low-voltage power supply network topology serves as the foundation and prerequisite for the intelligent construction of low-voltage distribution networks, which is conducive to further improving the service quality of smart grids and is of great significance for lean management of distribution areas and loss reduction and energy saving. Attached Figure Description

[0028] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments of this application and their descriptions are used to explain this application and do not constitute an undue limitation of this application.

[0029] Figure 1 A flowchart illustrating an embodiment of this application;

[0030] Figure 2 A flowchart of topology identification provided for embodiments of this application;

[0031] Figure 3 A schematic diagram of wavelet decomposition provided for an embodiment of this application;

[0032] Figure 4 This is a schematic diagram of the branch node connection relationship determination process provided in the embodiments of this application;

[0033] Figure 5 A schematic diagram of the branch section provided in the embodiments of this application;

[0034] Figure 6 This is a schematic diagram of an abnormal user provided in an embodiment of this application;

[0035] Figure 7 This is a schematic diagram of the topology provided in an embodiment of this application;

[0036] Figure 8 This is a schematic diagram of the branch node relationship determination process provided in the embodiments of this application;

[0037] Figure 9 This is a schematic diagram of the branch node correlation analysis calculation results provided in an embodiment of this application. Detailed Implementation

[0038] It should be noted that the following detailed descriptions are exemplary and intended to provide further explanation of this application. Unless otherwise specified, all technical and scientific terms used in this application have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0039] It should be noted that the terminology used herein is for the purpose of describing particular implementations only and is not intended to limit the exemplary implementations according to this application. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. Furthermore, it should be understood that the terms “comprising” and “having”, and any variations thereof, are intended to cover non-exclusive inclusion, for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such process, method, product, or apparatus.

[0040] Where there is no conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0041] Example 1

[0042] This embodiment provides a method for identifying low-voltage network topology;

[0043] like Figure 1 As shown, a low-voltage network topology identification method includes:

[0044] Establish a general structure for the low-voltage power supply network topology, dividing the meter nodes in the low-voltage power supply network topology into branch nodes and user nodes;

[0045] Based on the power data of the branch nodes, obtain the connection relationship of the branch nodes;

[0046] Based on the power data of the branch nodes and the power data of the user nodes, obtain the connection location of the user nodes;

[0047] Based on the voltage data of the branch nodes and the voltage data of the user nodes, the abnormal results of the user node connection positions are corrected, and the low-voltage power supply network topology is obtained.

[0048] Furthermore, based on the power data of the branch nodes, the connection relationships of the branch nodes are obtained, including:

[0049] Based on the power data of the branch nodes, obtain the possible parent nodes of each branch node;

[0050] For the power data of branch nodes and the power data of possible parent nodes, obtain the similarity of possible parent nodes;

[0051] Based on similarity, obtain the parent node of each branch node.

[0052] Furthermore, wavelet transform features are extracted from the branch nodes and their possible parent nodes to obtain feature power data;

[0053] Based on the characteristic power data, obtain the correlation coefficient of each possible parent node;

[0054] Based on the correlation coefficient, obtain the parent node of the branch node.

[0055] Furthermore, the formula for wavelet transform is defined as follows:

[0056]

[0057] in, Indicates the original input; Represents the mother wavelet function; , is the scale factor; , where m and n are translation parameters; m and n are both positive integers; T is the number of sampling points.

[0058] Furthermore, obtaining the user node connection location based on the power data of the branch node and the power data of the user node includes:

[0059] Based on the power data of the branch nodes, obtain the power data of the branch section;

[0060] Based on the power data of the branch segment and the power data of the user node, obtain the connection location of the user node.

[0061] Furthermore, based on power conservation, the relationship between the power data of the branch section and the power data of the user node is obtained;

[0062] Based on the relationship between the power data of the branch segment and the power data of the user node, the connection position of the user node is obtained through 0-1 integer quadratic programming.

[0063] Furthermore, the relationship between the power data of the branch segment and the power data of the user node is as follows:

[0064]

[0065] in, Let L be the segment power of the branch segment L at time t; Let R be the power value of user R at time t; The set of all users on segment L; This refers to the power loss in section L and the meter measurement error.

[0066] Next, combined Figure 1-6 This embodiment provides a detailed description of a low-voltage power supply network topology identification method.

[0067] A method for identifying the topology of a low-voltage power supply network involves: calculating the mean of the power sequence of each branch node; identifying possible parent nodes for each branch node in descending order of mean; extracting features using wavelet transform on the power sequences of the branch node and all possible parent nodes; calculating the correlation coefficient between the power sequences and each possible parent node; and identifying the parent node with the largest correlation coefficient. This completes the determination of the connection relationships between branch nodes. Based on this, the segment power of each branch section is calculated. Utilizing the ideal relationship that segment power equals the sum of the segment's user power, a 0-1 integer quadratic programming method is used to calculate the column vector reflecting the user connection locations. Then, hypothesis testing is applied to verify and correct the judgment results, completing the determination of user node connection locations and obtaining the topology of the entire power supply network. This includes:

[0068] S1. Establish a general structure for the low-voltage power supply network topology, and divide the meter nodes in the low-voltage power supply network topology into branch nodes and user nodes.

[0069] The general structure of low-voltage power supply network topology includes regular and complex topologies. Regular topologies are commonly found in urban buildings and newly built communities, employing radial wiring. The 400V low-voltage busbar forms multiple branch lines through a primary distribution device, and each branch line then distributes power to different end users through a secondary distribution device. Complex topologies are common in rural areas or peri-urban areas, encompassing various forms. Unlike regular topologies, in complex topologies, user nodes are not only located at the end of the power supply line but may also be located on branch sections between any two branch nodes. The difference in topologies between urban and rural power supply networks stems from the difference in user distribution. Urban users are relatively concentrated, resulting in heavy loads, large currents, and relatively significant voltage drops. Therefore, users are connected in parallel at the end of the power supply line to ensure voltage quality for each user. Rural users are often more dispersed, with lighter loads, smaller currents, and less significant voltage drops, thus allowing users to be irregularly connected to various branch sections.

[0070] Regular topologies are actually a simplified form of complex topologies. Without loss of generality, this embodiment focuses on complex topologies. For ease of analysis, this embodiment makes the following simplifications and assumptions:

[0071] (1) Since branch meters and user meters are distinguished in the actual system, it is assumed that the types and quantities of meters are known;

[0072] (2) In reality, there may be three-phase users, but in theoretical analysis, they can be regarded as three single-phase users. Therefore, in order to simplify the theoretical analysis process, it is assumed that all users are single-phase users.

[0073] (3) Assume the user's phase is known;

[0074] (4) Power backfeed phenomenon in the power supply network is not considered for the time being. If there is photovoltaic on the user side, the impact can be shielded by nighttime data.

[0075] (5) Assume that the smart meter works normally during the sampling period and there are no users whose power is always zero.

[0076] S2. Based on the power data of the branch nodes, obtain the connection relationship of the branch nodes. Due to the user access in the branch section between two branch nodes in the complex topology, the electrical energy of a certain branch node is not equal to the sum of the electrical energy of its connected child branch nodes, and the same applies to power and current. This makes the equality relationship between upper and lower level branch nodes unclear, and methods such as regression analysis and linear programming are not applicable to the determination of the connection relationship of branch nodes. The characteristics of user electricity consumption determine that the power curve will have multiple power value abrupt change points. For branch nodes at the same level or without direct connection relationship, the power abrupt change of each node is random, but for branch nodes with direct upper and lower level connection relationship, the power abrupt change of the two nodes is similar. This paper uses power abrupt change as a feature to analyze whether there are similar abrupt change features in its possible parent nodes, so as to find the parent node corresponding to the child node, and then determine the connection relationship of the branch node. Specifically, it includes:

[0077] S201. Perform wavelet transform feature extraction on the branch node and its possible parent node to obtain feature power data;

[0078] Specifically, wavelet transform possesses excellent localization properties in both the time and frequency domains, thus it is widely used in signal decomposition and reconstruction, signal-to-noise separation, and data feature extraction. This embodiment employs Discrete Wavelet Transform (DWT), defined as:

[0079] (1)

[0080] in, Indicates the original input; Represents the mother wavelet function; , is the scale factor; is the translation parameter; m and n are both positive integers; T is the number of sampling points.

[0081] The power sequence is decomposed using DWT, assuming a 3-level decomposition process, as follows: Figure 3 As shown.

[0082] A3 represents the low-frequency component after three-level decomposition, reflecting the overall trend of the power sequence, and is called the "approximate component." D1, D2, and D3 represent the high-frequency components in the decomposition process, reflecting the detailed features of the power sequence at different frequencies, and are called "detail components." Power abrupt changes belong to the high-frequency components in the power sequence and exist in the detail components obtained from wavelet decomposition. In this embodiment, the detail components are extracted as the abrupt change features of the power sequence.

[0083] The decomposition results are represented by row vectors, as shown in equation (2):

[0084] (2)

[0085] in, These are approximate wavelet coefficients; These are the detailed wavelet coefficients.

[0086] Wavelet coefficients represent the degree of matching between the original data sequence and the wavelet basis function at a certain moment. The larger the wavelet coefficients, the more components with the same frequency as the wavelet basis function are contained in the original data sequence. By filtering the wavelet coefficients, retaining the larger wavelet coefficients and removing the smaller ones, the power curves at non-abrupt points can be smoothed while preserving power abrupt changes, thus completing the feature extraction of the power sequence.

[0087] right Feature extraction is performed to obtain the processed detail wavelet coefficient row phasors. The feature extraction process is as follows:

[0088] (3)

[0089] in, This is the threshold value for wavelet coefficients.

[0090] Will Medium to large Wavelet coefficients are preserved, less than Setting the wavelet coefficients to zero completes the feature extraction of detail components.

[0091] Since the approximate components reflect the overall trend of the power sequence and do not exhibit abrupt power changes, therefore... All wavelet coefficients in the vector are set to zero, resulting in a row vector of all zeros. .

[0092] The wavelet coefficient row vector after feature extraction is:

[0093] (4)

[0094] The detail wavelet components after feature extraction , , and the approximate wavelet components after complete nullification Wavelet reconstruction is performed to finally obtain the feature power sequence after feature extraction. .

[0095] S202. Based on the feature power data, obtain the correlation coefficient of each possible parent node; specifically, before judging the similarity of the power sequence of the branch nodes, perform feature extraction on the sub-branch nodes and all possible parent nodes.

[0096] Because power sequences from different user nodes may annihilate each other within a certain time period, resulting in unclear power fluctuation characteristics of upper-level branch nodes, a preliminary selection of power data for the nodes to be analyzed can be performed. The node's power sampling value P is divided into multiple time periods, and the variance of the power value in each period is calculated. If the variance value of a certain period is small, it indicates that the power change characteristics are not obvious and it is not suitable for similarity analysis. Therefore, based on the variance results, only power sequences with stronger fluctuations can be selected for calculation.

[0097] Find the time when the power change occurs in the sub-branch node and the corresponding spectral components. For all possible parent nodes, only the spectral components at the corresponding time are retained, and the spectral components at other times are ignored. The extraction process is shown in Equation (5):

[0098] (5)

[0099] in, The row vector of feature wavelet coefficients for the branch node to be identified; The row vector of feature wavelet coefficients of the possible parent nodes of the branch node to be identified; for Possible parent nodes, let There are M possible parent nodes, then .

[0100] Will , All approximate wavelet coefficients in the sequence are set to zero, and wavelet reconstruction is performed separately to finally obtain the characteristic power sequence. and .

[0101] Using Pearson correlation coefficient As a metric for the similarity of the feature power sequences of branch nodes, its expression is as follows:

[0102] (6)

[0103] in, Let X be the covariance of Y; Let X be the variance; The value of is between -1 and 1. The larger the value, the higher the positive correlation between X and Y, and the more likely the two nodes are to be parent and child nodes.

[0104] S3. Based on the correlation coefficient, obtain the parent node of each branch node. Specifically, for any branch node, the power value at each time step will not be greater than the power value of the parent node at the corresponding time step, that is:

[0105] (7)

[0106] in, for The parent node; Branch nodes and The power value at sampling time t.

[0107] So:

[0108] (8)

[0109] Right now:

[0110] (9)

[0111] Given N branch nodes, sort the power sequence mean of each branch node in descending order to obtain an array:

[0112] (10)

[0113] in, Let Fn be the mean of the branch nodes.

[0114] For branch nodes Its power sequence has the largest mean, and it is determined to be the root node. root node child nodes, such as Figure 4 As shown in (a).

[0115] For branch nodes , and Both could be its parent node, such as Figure 4 As shown in (b). In order to determine and , It has better distinguishability when the similarity is high, for Update the power sequence by subtracting child nodes. The power, that is .by For objects, to , , power sequence , , Feature extraction is performed to obtain the feature power sequence. , , Calculate them separately. and , Pearson correlation coefficient , ,like ,but for The child nodes; conversely, for The child nodes.

[0116] Similarly, for branch nodes , … Both could be its parent node, such as Figure 4 As shown in (c). For … The power sequence is updated to obtain a new power sequence. … .by For objects, to And the updated … power sequence … Feature extraction is performed to obtain the feature power sequence. and … Calculate them separately. and … Pearson correlation coefficient … Select the maximum value among them. ,So That is The child nodes. And so on, until the connection relationships of all branch nodes have been determined, such as... Figure 4 As shown in (d).

[0117] S3. Based on the power data of the branch nodes and the power data of the user nodes, obtain the connection location of the user nodes; including:

[0118] S301. Based on the power data of the branch nodes, obtain the power data of the branch sections. The lines between two branch nodes and the lines connected to the terminal branch node are collectively referred to as branch sections. After determining the connection relationships of the branch nodes, the power of all branch sections can be calculated based on the power data of each branch node. For a branch section with branch nodes at both ends, its section power is equal to the power of the first branch node minus the power of all the last branch nodes. For example... Figure 5 As shown, the segment power of segment L1 at time t is L2 section power .

[0119] S302. Obtain the connection location of the user node based on the power data of the branch section and the power data of the user node. Specifically, since the user phases are known, the connection locations of the user nodes are determined separately for phases A, B, and C. By default, the power of the branch section is the power of a certain phase, and the user in the section is the user of that phase. According to the law of conservation of power, the relationship between the power of the branch section and the power of the user is shown in equation (11):

[0120] (11)

[0121] in, Let L be the segment power of the branch segment L at time t; Let R be the power value of user R at time t; The set of all users on segment L; This refers to the power loss in section L and the meter measurement error.

[0122] Introducing a 0-1 variable x to represent the connection relationship between a user node and a branch segment: if user R is connected to segment L, then... ,on the contrary Equation (11) can be transformed into:

[0123] (12)

[0124] in, This refers to the set of all users of a specific phase within a given transformer area.

[0125] Define matrix:

[0126] (13)

[0127] Where N is the total number of user nodes in the distribution area, and T is the number of meter sampling points.

[0128] The matrix expression of equation (12) is as follows:

[0129] (14)

[0130] By trying different values ​​of x, find the one that makes The smallest possible set of X values ​​is sufficient to determine the connection location of the user node. The following optimization model is constructed to solve for X:

[0131] (15)

[0132] Equation (15) is a 0-1 integer quadratic programming problem. Solving this problem yields the optimal solution for X, thereby finding the users connected to segment L. By solving each branch segment separately, the branch segment location of all users can be completed.

[0133] Considering the special case where the power characteristics of different users may be highly similar within a certain time period, thus affecting the topology identification results, this paper uses a power sequence of no less than 100 data points or a corresponding time period of no less than 24 hours for each calculation. By using longer time series data and varying the length of the data sequence for each calculation, highly similar power characteristics of different users can be avoided as much as possible. Furthermore, multiple calculations can be performed using multiple sets of data to continuously refine the topology judgment results, ultimately obtaining an accurate topology structure.

[0134] S4. Based on the voltage data of the branch nodes and the voltage data of the user nodes, correct any abnormal results in the connection positions of the user nodes and obtain the low-voltage power supply network topology.

[0135] Due to factors such as line losses and asynchronous meter data collection, the power of a line segment is not strictly equal to the sum of the power of users in that segment; there is a certain difference between the two, which is generally less than the power of most users. However, within a transformer substation, some users may have lower power consumption levels during the sampling period, or different users may have highly similar power characteristics, which could lead to misjudgments of user connection locations. Therefore, the results of the 0-1 integer quadratic programming need to be checked and corrected.

[0136] Since the above topology identification methods are all based on electricity meter power data, voltage data is used to correct abnormal results in order to comprehensively utilize the value of electricity meter measurement data and improve the reliability of the topology identification methods. The voltage of a single-source radial power supply network exhibits the following characteristics:

[0137] (1) Without considering the photovoltaic access on the user side, the node voltage gradually decreases from the beginning to the end of the same time section.

[0138] (2) The voltage similarity of users located on the same branch is higher than that of users located on different branches, and the closer the electrical distance between two nodes, the closer the voltages are.

[0139] Based on the above characteristics, the voltage of the user node is close to and slightly lower than the voltage of the adjacent upstream branch node at the same time section. In reality, due to factors such as meter measurement errors, there may be individual sampling moments where the user voltage is higher than the adjacent upstream branch node, but this does not affect the characteristic that the overall voltage sequence over a period of time is lower than that of the adjacent upstream branch node.

[0140] To compare the voltage levels of users and adjacent upstream branch nodes over a period of time, a paired-samples t-test is performed using the voltage sequences of both over a period of time as samples to determine whether there is a significant difference between the two. The specific steps are as follows:

[0141] S401, Define variables .in, For user node voltage sequences; The voltage sequence of the user's adjacent upstream branch nodes; n is the number of sampling points, i.e., the number of samples. Follows a normal distribution. For variables mean Perform a check to determine if the value is less than or equal to zero. If it is, then the user is considered to be located under the current branch segment.

[0142] S402. Propose the null hypothesis H0: Alternative hypothesis H1: .

[0143] S403. Constructing the test statistic In the formula: , Samples The mean and variance of.

[0144] S404, from the t-distribution table, we get... The value of . The significance level is typically set to 0.05. If... If the condition is met, the null hypothesis is rejected, and the user is considered not to be in the current branch segment; otherwise, the null hypothesis is accepted, and the user is considered to be in the current branch segment.

[0145] like Figure 6 As shown, if the inspection finds that user R is connected to both branch segment Lm and branch segment Ln, then a paired sample t-test is performed on user R and its upstream branch nodes Fm and Fn according to steps (1) to (4). The branch node that satisfies the null hypothesis is the true upstream branch node of the user. Furthermore, if the paired sample t-tests of user R with Fm and Fn all satisfy the null hypothesis, that is, the user's voltage level is not higher than that of branch nodes Fm and Fn, then the Euclidean distance between the voltage sequences of user R and branch nodes Fm and Fn is calculated. The branch node with the smallest Euclidean distance to user R is its true upstream branch node, thereby determining the branch segment where user R is located.

[0146] By correcting the connection locations of all abnormal users, the accurate determination of user node connection locations can be achieved. By locating users on phases A, B, and C respectively using the method described above, the identification of the entire low-voltage power supply network topology can be completed.

[0147] To verify the effectiveness of the low-voltage power supply network topology identification method described in this embodiment, this embodiment uses... Figure 7The power supply network shown is used as the research object. The network structure and smart meter data are derived from an actual 400V power supply system in Yantai City, Shandong Province, including 12 branch meters and 114 user meters. The minimum sampling interval for the meters is 15 minutes, with 96 samples taken per day from September 13th to September 20th, 2020, totaling 672 sampling points. After removing invalid data, 482 valid sampling points were obtained.

[0148] First, determine the connection relationship between the branch nodes. The determination process is as follows:

[0149] (1) Take the total three-phase power value of all valid sampling points of the smart meter as the power sequence of each branch node. .

[0150] (2) For the 12 branch nodes … The power sequences were calculated by taking their mean values ​​and then sorted from largest to smallest. The sorting results are as follows: The connection relationships of the branch nodes are judged in this order.

[0151] (3) First, it can be determined that for child nodes, such as Figure 3 As shown in (a), then judge in sequence. arrive The connection relationship is shown in Appendix Table 1. The correlation coefficient of the power characteristic sequence of the child node and its possible parent node in each step is shown in Appendix Table 1. The node with the largest correlation coefficient with the child node is the parent node of the child node. The judgment process is shown in Figure 8.

[0152] Figure 9 The correlation coefficients of power feature sequences between each branch node and the true parent node, the correlation coefficients of power feature sequences between each branch node and the next possible parent node, and the difference between the two are presented in the form of a bar chart during the topology identification process.

[0153] like Figure 9 As shown, the feature power sequences between child nodes and their true parent nodes exhibit a strong correlation, with correlation coefficients mostly around 0.8 and even the smallest correlation coefficient above 0.6. Furthermore, there are certain differences in the correlation coefficients between child nodes and their true parent nodes and the next most likely parent nodes, with the smallest difference being above 0.2. There is a clear distinction between child nodes and their true parent nodes and non-parent nodes; therefore, this method has good reliability.

[0154] Secondly, the connection location of the user nodes is determined. Since the user phase is known, the connection location is determined by phase. Taking 38 B-phase users as an example, the determination steps are as follows:

[0155] (1) Based on the identified branch node connection relationships, calculate the power sequence of all branch segments in phase B, and use... … express.

[0156] (2) The 12 branch segments are calculated according to the 0-1 integer quadratic programming method described in Section 3.1 to obtain 12 column vectors reflecting the connection positions of user nodes, as shown in Table 1.

[0157] Table 1 User Node Connection Locations

[0158] Table 1 Location of user nodes

[0159]

[0160] Note: Values ​​of "0" are omitted as blanks in the table.

[0161] (3) Checking the user connection location judgment results shown in Table 1, it was found that the connection location judgment of user B37 was abnormal, being judged as connected to both segment L6 and segment L12. Comparing the voltage sequences of user B37 and the branch nodes F6 and F12 at the beginning of branch segments L6 and L12, the voltage of B37 is generally higher than the voltage of F6 and similar to the voltage of F12. Furthermore, assuming that the voltage level of B37 is lower than that of F6, a paired-samples t-test was performed on it, and the significance level was taken. From the t-distribution table, we can obtain... ; After calculation, Therefore, we reject the null hypothesis; the actual voltage level of B37 is higher than that of F6. Assuming the voltage level of B37 is lower than that of F12, we perform a paired-samples t-test on it. Therefore, the original assumption is true, and the actual situation is that the voltage level of B37 is lower than that of F12. Since the voltage of the radial power supply network gradually decreases from the beginning to the end, user B37 is connected at the end of branch node F12, located in branch segment L12.

[0162] At this point, the connection locations of Phase B users have been determined, and the resulting topology is shown in the attached figure. Figure 3 As shown.

[0163] Similarly, following the steps above, the connection locations of phase A and phase C users are determined. After completing step (2), the determination results for phase A and phase C are checked, and no abnormalities are found. The determination results for phases A, B, and C are integrated to obtain the final power supply network topology and its associated structure. Figure 7 same.

[0164] To further verify the reliability of the proposed method, eight other different topologies were also analyzed, and the results are shown in Table 2.

[0165] Table 2 Topology Recognition Results

[0166] Table 1Result of topology identification

[0167] Topology number Number of branch nodes Number of user nodes Number of correctly identified branch nodes The number of user nodes was correctly identified. Topology recognition accuracy % 1 8 92 8 92 100 2 9 102 9 102 100 3 11 108 11 108 100 4 13 126 13 126 100 5 13 128 13 128 100 6 14 140 14 140 100 7 15 156 15 154 99 8 15 168 15 166 99

[0168] The results show that the proposed method achieves an accuracy of over 99%, correctly identifying the branch node connections for all eight topologies. Two errors were found in user node identification for topologies 7 and 8, because both topologies had an inactive user whose power remained consistently low throughout the sampling period. This can be mitigated by extending the sampling time or selecting different sampling periods.

[0169] The method presented in this paper is compared with commonly used regression analysis methods. Figure 8 The recognition performance of the topological structures shown is compared, and the results are shown in Table 3.

[0170] Table 3 Comparison of Topology Recognition Results

[0171] Table 2Comparison of topology identification results

[0172] Topology identification methods Topology recognition accuracy % Regression analysis 92.1 This article's method 100

[0173] While regression analysis achieves high accuracy in identifying regular topologies, the method in this embodiment demonstrates even higher accuracy for identifying complex topologies. Furthermore, regression analysis requires prior knowledge of the connection relationships between branch nodes, whereas the method in this embodiment can perform topology identification even when the topology is completely unknown. Therefore, the method proposed in this embodiment exhibits greater reliability and applicability.

[0174] Example 2

[0175] This embodiment provides a low-voltage power supply network topology identification system;

[0176] A low-voltage power supply network topology identification system, comprising:

[0177] The initial topology acquisition module is used to establish a general structure of the low-voltage power supply network topology, and to divide the meter nodes in the low-voltage power supply network topology into branch nodes and user nodes.

[0178] The branch node connection relationship acquisition module acquires the branch node connection relationship based on the power data of the branch nodes;

[0179] The user node connection location acquisition module is used to acquire the user node connection location based on the power data of the branch node and the power data of the user node.

[0180] The low-voltage power supply network topology acquisition module is used to check and correct abnormal results of user node connection positions based on the voltage data of the branch nodes and the voltage data of the user nodes, and to acquire the low-voltage power supply network topology.

[0181] It should be noted that the initial topology acquisition module, branch node connection relationship acquisition module, user node connection location acquisition module, and low-voltage power supply network topology acquisition module mentioned above correspond to the steps in Embodiment 1. The examples and application scenarios implemented by these modules and their corresponding steps are the same, but they are not limited to the content disclosed in Embodiment 1. It should be noted that these modules, as part of the system, can be executed in a computer system such as a set of computer-executable instructions.

[0182] Example 3

[0183] This embodiment also provides an electronic device, including a memory and a processor, as well as computer instructions stored in the memory and running on the processor. When the computer instructions are executed by the processor, they complete the method described in Embodiment 1.

[0184] Example 4

[0185] This embodiment also provides a computer-readable storage medium for storing computer instructions, which, when executed by a processor, complete the method described in Embodiment 1.

[0186] The descriptions of each embodiment in the above embodiments have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0187] The proposed system can be implemented in other ways. For example, the system embodiments described above are merely illustrative. For instance, the division of modules described above is only a logical functional division. In actual implementation, there may be other division methods. For example, multiple modules may be combined or integrated into another system, or some features may be ignored or not executed.

[0188] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A method for identifying the topology of a low-voltage power supply network, characterized in that, include: A general structure for the low-voltage power supply network topology is established, and the meter nodes in the low-voltage power supply network topology are divided into branch nodes and user nodes; Based on the power data of the branch nodes, the connection relationship of the branch nodes is obtained, including: based on the power data of the branch nodes, obtaining the possible parent nodes of each branch node; performing wavelet transform feature extraction on the branch nodes and their possible parent nodes; decomposing the power data through wavelet transform; obtaining wavelet coefficients containing detail components and approximate components after decomposition; performing threshold filtering on the decomposed wavelet coefficients: retaining the large wavelet coefficients in the detail components, filtering out the small wavelet coefficients, and setting all coefficients of the approximate components to zero; and then reconstructing to obtain a power sequence containing only abrupt change features, thereby obtaining the feature power data. Obtaining the user node connection location based on the power data of the branch node and the power data of the user node includes: obtaining the power data of the branch segment based on the power data of the branch node; obtaining the user node connection location based on the power data of the branch segment and the power data of the user node; obtaining the relationship between the power data of the branch segment and the power data of the user node based on power conservation; and obtaining the user node connection location through 0-1 integer quadratic programming based on the relationship between the power data of the branch segment and the power data of the user node. Based on the voltage data of the branch nodes and the voltage data of the user nodes, the abnormal results of the user node connection positions are corrected by paired-samples t-test to obtain the low-voltage power supply network topology: Define variables. ,in, For user node voltage sequences; The voltage sequence of the user's adjacent upstream branch nodes; n is the number of sampling points. Follows a normal distribution, for variables mean Perform a test to determine if the value is less than or equal to zero. If so, consider the user to be below the current branch segment; propose the null hypothesis H0: Alternative hypothesis H1: Construct the test statistic In the formula: , Samples The mean and variance; find the t-distribution table. The value, For significance level, if If the condition is met, the null hypothesis is rejected, and the user is considered not to be in the current branch segment; otherwise, the null hypothesis is accepted, and the user is considered to be in the current branch segment.

2. The low-voltage power supply network topology identification method as described in claim 1, characterized in that, The step of obtaining the branch node connection relationship based on the power data of the branch node further includes: For the power data of the branch nodes and the power data of the possible parent nodes, obtain the similarity of the possible parent nodes; Based on the similarity, obtain the parent node of each branch node.

3. The low-voltage power supply network topology identification method as described in claim 2, characterized in that, Based on the characteristic power data, obtain the correlation coefficient of each possible parent node; Based on the correlation coefficient, obtain the parent node of the branch node.

4. The low-voltage power supply network topology identification method as described in claim 3, characterized in that, The formula for wavelet transform is defined as follows: in, Indicates the original input; Represents the mother wavelet function; , is the scale factor; , where m and n are translation parameters; m and n are both positive integers; T is the number of sampling points.

5. The low-voltage power supply network topology identification method as described in claim 1, characterized in that, The relationship between the power data of the branch segment and the power data of the user node is as follows: in, Let L be the segment power of the branch segment L at time t; Let R be the power value of user R at time t; The set of all users on segment L; This refers to the power loss in section L and the meter measurement error.

6. A low-voltage power supply network topology identification system, employing the low-voltage power supply network topology identification method as described in any one of claims 1-5, characterized in that, include: The initial topology acquisition module is used to establish a general structure of the low-voltage power supply network topology, and to divide the meter nodes in the low-voltage power supply network topology into branch nodes and user nodes. The branch node connection relationship acquisition module acquires the branch node connection relationship based on the power data of the branch nodes; The user node connection location acquisition module is used to acquire the user node connection location based on the power data of the branch node and the power data of the user node. The low-voltage power supply network topology acquisition module is used to check and correct abnormal results of user node connection positions based on the voltage data of the branch nodes and the voltage data of the user nodes, and to acquire the low-voltage power supply network topology.

7. An electronic device, characterized in that, It includes a memory and a processor, as well as computer instructions stored in the memory and running on the processor, which, when executed by the processor, perform the method according to any one of claims 1-5.

8. A computer-readable storage medium for storing computer instructions, which, when executed by a processor, perform the method according to any one of claims 1-5.

Images