A method and device for identifying full-level topology of low-voltage distribution station area

Abstract

The present invention relates to the technical field of substation topology identification, and discloses a method and device for full-level topology identification of a low-voltage distribution substation. The method includes clustering the voltage sampling data of all users at any sampling time using the density peak method, obtaining three initial cluster centers corresponding to the three-phase voltages, and then using K-means clustering to update the cluster centers and divide the cluster clusters until the clustering objective function converges, and all users are divided into three cluster clusters to complete user phase identification; the membership factors between each user and all meter boxes, and the membership factors between each meter box and all branch boxes are calculated respectively, and the meter box and branch box corresponding to the highest membership factor value are obtained as the meter box to which the user belongs and the branch box to which the meter box belongs, to complete full-level topology identification. The present invention combines the density peak method with K-means clustering, and identifies the physical topology connection relationship by constructing a membership factor based on the similarity of the voltage curve, thereby improving the accuracy of topology identification.

Description

Technical Field

[0001] The present invention relates to the technical field of substation topology identification, and in particular to a method and device for identifying the full-level topology of a low-voltage distribution substation. Background Art

[0002] The low-voltage distribution network is at the end of the power grid and directly serves users. However, due to its complex wiring, errors are inevitable in record archiving. The key task of low-voltage distribution substation topology identification is to sort out the physical connection relationship between "distribution transformer-branch-meter box-user" and the phase sequence relationship of the user. The accuracy of its identification relationship fundamentally affects the stability of the power system and the effectiveness of its work. At present, the methods for low-voltage distribution network topology identification can be divided into signal injection method and data analysis method. The signal injection method is based on hardware equipment and faces problems such as high cost and large investment in practical applications. It is also easily affected by electromagnetic interference, resulting in erroneous results. The low-voltage distribution network is located at the end of the power grid and directly faces users. It is the most valuable component of big data in the construction of smart grids. Therefore, the use of data analysis methods has broad application prospects.

[0003] The existing data analysis method for low-voltage distribution network topology identification uses K-Means algorithm for clustering to identify the phase to which each user belongs. However, the K-Means algorithm needs to set the number of clusters in advance and the random selection of the initial cluster center may cause the clustering result to fall into the local optimum, affecting the clustering result, and thus leading to inaccurate topology identification of the full level of the low-voltage distribution station area. The graph-based method is used for topology identification. In the complex topological structure of the low-voltage distribution station area, the graph-based method has high computational complexity and processing difficulty, and is difficult to adapt to the dynamic changes of the topological structure of the low-voltage distribution station area. It also requires certain prior knowledge and domain knowledge of the topological structure of the low-voltage distribution station area in order to build a suitable graph model. However, in practical applications, it is difficult to obtain complete prior knowledge and domain knowledge, which in turn affects the accuracy and reliability of topological identification.

[0004] In summary, the existing low-voltage distribution substation topology identification methods, when performing user phase identification, mostly use the K-Means clustering algorithm, which requires setting the number of clusters in advance and the clustering results are prone to fall into local optimality, affecting the clustering results and causing inaccurate user phase identification; when identifying topological relationships between levels, they are mostly based on graph theory methods, which require prior knowledge to build a suitable graph model, but prior knowledge is difficult to obtain and difficult to adapt to the dynamic changes of the topological structure, resulting in disordered phase-to-user relationships in the identified low-voltage distribution substation, unclear power topology hierarchical affiliation, and difficulty in updating the topological structure. Summary of the invention

[0005] To this end, the technical problem to be solved by the present invention is to overcome the problems in the prior art of inaccurate clustering results and difficulty in obtaining prior knowledge required for topology identification between levels, resulting in inaccurate user phase identification, unclear power topology hierarchy affiliation and difficulty in updating the topology structure.

[0006] In order to solve the above technical problems, the present invention provides a method for identifying the full-level topology of a low-voltage distribution station area, comprising:

[0007] Obtain the voltage sampling data of all branch boxes, meter boxes and users in the low-voltage distribution station area to be identified at each sampling time in the preset sampling period, and form the voltage time series data corresponding to each branch box, each meter box and each user;

[0008] Based on the voltage sampling data of all users at any sampling time, the density peak method is used for clustering to obtain three initial cluster centers corresponding to the three-phase voltages respectively;

[0009] Based on the three initial cluster centers, K-means clustering is used to update the cluster centers and divide the clusters until the clustering objective function converges. All users are divided into three clusters, and the phase of each user is identified according to the phase corresponding to the cluster to which each user belongs.

[0010] Calculate the membership factor between each user and all meter boxes respectively, obtain the meter box corresponding to the highest membership factor value as the meter box to which the user belongs, and complete the user-meter box topology identification, including: constructing a first mean coefficient according to the standardized user voltage time series data and the meter box voltage time series data; obtaining a first correlation coefficient between the user and the meter box based on the first mean coefficient, the standard deviation of the user voltage time series data and the standard deviation of the meter box voltage time series data; and calculating the membership factor between the user and the meter box based on the first correlation coefficient;

[0011] Calculate the membership factors between each meter box and all branch boxes respectively, obtain the branch box corresponding to the highest membership factor value as the branch box to which the meter box belongs, and complete the topological identification of the meter box-branch box, including: constructing a second mean coefficient based on the standardized meter box voltage time series data and the branch box voltage time series data; obtaining a second correlation coefficient between the meter box and the branch box based on the second mean coefficient, the standard deviation of the meter box voltage time series data and the standard deviation of the branch box voltage time series data; and calculating the membership factor between the meter box and the branch box based on the second correlation coefficient;

[0012] Based on the user's phase identification, user-meter box topology identification, and meter box-branch box topology identification, the full-level topology identification of the low-voltage distribution station area to be identified is completed.

[0013] In one embodiment of the present invention, the voltage sampling data of all users at any sampling time is clustered using the density peak method to obtain three initial cluster centers corresponding to the three-phase voltages, including:

[0014] Obtain the voltage sampling data of each user at any sampling time in a preset sampling period, and use Gaussian Kernel to calculate the local density of each user;

[0015] For each user, obtain users whose Euclidean distance exceeds a preset distance and whose local density is higher than that of the user, and form a set of users to be clustered for the user;

[0016] If the user set to be clustered is empty, the maximum distance between the user and the remaining users is obtained as the cluster center distance of the user; if the user set to be clustered is not empty, the minimum distance between the user and all users in the user set to be clustered is obtained as the cluster center distance of the user;

[0017] Based on the product of each user's local density and the distance to the cluster center, the density peak of each user is obtained;

[0018] The density peaks of all users are sorted from large to small, and the users corresponding to the first three density peaks are selected as the three initial clustering centers corresponding to the three-phase voltages respectively.

[0019] In one embodiment of the present invention, the local density is expressed as:

[0020]

[0021] Among them, ρ i represents the local density of the i-th user; d ij Represents the distance between the i-th user and the j-th user, and the expression is: N represents the total number of users in the low-voltage distribution station area to be identified; voltage sampling data set represents the voltage sampling data of the i-th user at the T-th sampling time; d c Indicates the preset cutoff distance.

[0022] In one embodiment of the present invention, the cluster center distance includes:

[0023] If the set of users to be clustered is empty, the maximum distance between the user and the other users is obtained as the cluster center distance δ of the user i , expressed as:

[0024] If the set of users to be clustered is not empty, the minimum distance between the user and all users in the set of users to be clustered is obtained as the cluster center distance δ of the user i , expressed as:

[0025] In one embodiment of the present invention, based on the three initial cluster centers, K-means clustering is used to update the cluster centers and divide the cluster clusters until the clustering objective function converges, all users are divided into three cluster clusters, and the phase identification of each user is completed according to the phase corresponding to the cluster cluster to which each user belongs, including:

[0026] Based on the voltage sampling data of all users at any sampling time and three initial cluster centers, a clustering objective function is constructed;

[0027] Using K-means clustering, based on the three initial clusters, recalculate the centroid of each cluster as the new cluster center, and divide and generate new clusters;

[0028] Based on the new clusters, new cluster centers are repeatedly selected until the clustering objective function converges;

[0029] Three clusters are obtained when the clustering objective function converges, and the phase of the users in each cluster is consistent with the phase of the cluster to which it belongs.

[0030] In one embodiment of the present invention, the clustering objective function is constructed based on the voltage sampling data of all users at any sampling time and the three initial clustering centers, and the centroid of each cluster is recalculated as a new clustering center based on the three initial clustering clusters using K-means clustering, including:

[0031] The clustering objective function is expressed as:

[0032] Among them, r ij is a hidden variable, whose value is α k represents the kth cluster center, 1≤k≤K, K represents the number of cluster centers; N represents the total number of users in the low-voltage distribution station area to be identified;

[0033] Using K-means clustering, based on the three initial clusters, recalculate the centroid of each cluster as the new cluster center:

[0034] C k represents the data set of the kth cluster; x i Represents the voltage sampling data of the i-th user.

[0035] In one embodiment of the present invention, the calculating of the membership factor between the user and the meter box includes:

[0036] The average voltage X based on the user's voltage time series data mean With standard deviation X std , standardize the user voltage time series data, and obtain the standardized user voltage time series data A t , expressed as:

[0037] The mean voltage Y based on the voltage time series data of the meter box mean and standard deviation Y std , standardize the meter box voltage time series data, and obtain the standardized meter box voltage time series data B t , expressed as:

[0038] Using the standardized user voltage timing data A t With meter box voltage timing data B t , construct the first mean coefficient C, expressed as:

[0039] Based on the first mean coefficient C, the standard deviation X of the user voltage time series data std The standard deviation Y of the meter box voltage time series data std , get the first correlation coefficient R between the user and the table box, expressed as:

[0040] Based on the first correlation coefficient R, the minimum value R of the first correlation coefficient min The maximum value of the first correlation coefficient R max , construct the membership factor R between the user and the table box subordinate , expressed as:

[0041]

[0042] Among them, X t Indicates the voltage sampling data of the user at the tth sampling time within the preset sampling period, Y t represents the voltage sampling data of the meter box at the tth sampling time within the preset sampling period, m represents the total number of sampling times within the preset sampling period; X mean It represents the mean voltage of the user voltage time series data within the preset sampling period, expressed as Y mean It represents the mean voltage of the meter box voltage time series data within the preset sampling period, expressed as X std It represents the standard deviation of the user voltage timing data within the preset sampling period, expressed as Y stdIt represents the standard deviation of the meter box voltage time series data within the preset sampling period, expressed as

[0043] In one embodiment of the present invention, the calculating the membership factor between the meter box and the branch box includes:

[0044] The mean voltage Y based on the voltage time series data of the meter box mean and standard deviation Y std , standardize the meter box voltage time series data, and obtain the standardized meter box voltage time series data B t , expressed as:

[0045] Mean voltage Z based on branch box voltage time series data mean and standard deviation Z std , standardize the branch box voltage time series data, and obtain the standardized branch box voltage time series data D t , expressed as:

[0046] Using the standardized meter box voltage time series data B t With branch box voltage timing data D t , construct the second mean coefficient C', expressed as:

[0047] Based on the second mean coefficient C', the standard deviation Y of the meter box voltage time series data std The standard deviation Z of the branch box voltage time series data std , obtain the second correlation coefficient R' between the meter box and the branch box, expressed as:

[0048] Based on the second correlation coefficient R', the minimum value R' of the second correlation coefficient min The maximum value of the second correlation coefficient R' max , construct the membership factor R' between the user and the table box subordinate , expressed as:

[0049]

[0050] Among them, Z t It represents the voltage sampling data of the branch box at the tth sampling time within the preset sampling period, Y t represents the voltage sampling data of the meter box at the tth sampling time within the preset sampling period, m represents the total number of sampling times within the preset sampling period; Z mean It represents the mean voltage of the branch box voltage time series data within the preset sampling period, expressed as Y mean It represents the mean voltage of the meter box voltage time series data within the preset sampling period, expressed as Z std It represents the standard deviation of the branch box voltage time series data within the preset sampling period, expressed as Y std It represents the standard deviation of the meter box voltage time series data within the preset sampling period, expressed as

[0051] In one embodiment of the present invention, after obtaining the voltage sampling data of all branch boxes, meter boxes and users in the low-voltage distribution station area to be identified at each sampling moment in the preset sampling period, it also includes: preprocessing all voltage sampling data using the Z-Score standardization method.

[0052] The embodiment of the present invention further provides a device for identifying a full-level topology of a low-voltage distribution station area, comprising:

[0053] An acquisition module is used to acquire the voltage sampling data of all branch boxes, meter boxes and users in the low-voltage distribution station area to be identified at each sampling moment in a preset sampling period, and to form voltage time series data corresponding to each branch box, each meter box and each user;

[0054] The phase identification module is used to cluster the voltage sampling data of all users at any sampling time using the density peak method to obtain three initial cluster centers corresponding to the three-phase voltages respectively; based on the three initial cluster centers, K-means clustering is used to update the cluster centers and divide the cluster clusters until the clustering objective function converges, and all users are divided into three cluster clusters. According to the phase corresponding to the cluster cluster to which each user belongs, the phase identification of each user is completed;

[0055] A user-meter box topology identification module is used to respectively calculate the membership factor between each user and all meter boxes, obtain the meter box corresponding to the highest membership factor value as the meter box to which the user belongs, and complete the user-meter box topology identification, including: constructing a first mean coefficient according to the standardized user voltage time series data and the meter box voltage time series data; obtaining a first correlation coefficient between the user and the meter box based on the first mean coefficient, the standard deviation of the user voltage time series data and the standard deviation of the meter box voltage time series data; and calculating the membership factor between the user and the meter box based on the first correlation coefficient;

[0056] The meter box-branch box topology identification module is used to respectively calculate the membership factors between each meter box and all branch boxes, obtain the branch box corresponding to the highest membership factor value as the branch box to which the meter box belongs, and complete the topology identification of the meter box-branch box, including: constructing a second mean coefficient based on the standardized meter box voltage time series data and the branch box voltage time series data; obtaining a second correlation coefficient between the meter box and the branch box based on the second mean coefficient, the standard deviation of the meter box voltage time series data and the standard deviation of the branch box voltage time series data; and calculating the membership factor between the meter box and the branch box based on the second correlation coefficient;

[0057] The full-level topology identification module is used to complete the full-level topology identification of the low-voltage distribution station area to be identified based on the user's phase identification, the user-meter box topology identification, and the meter box-branch box topology identification.

[0058] The above technical solution of the present invention has the following advantages compared with the prior art:

[0059] The method for identifying the full-level topology of a low-voltage distribution station area described in the present invention uses a density peak method for clustering. After obtaining three initial cluster centers, K-means clustering is used to update the cluster centers and divide the cluster clusters, and three cluster clusters corresponding to three-phase voltages are obtained to complete the phase identification of each user; the density peak method is used to construct a density peak according to the local density of the user and the distance from the cluster center for clustering, and the spatial information and density information of the user's voltage data are taken into account, and the similarities and differences of different voltage fluctuation curves can be accurately identified, thereby improving the accuracy of topology identification; the present invention combines the density peak method with K-means clustering, and after obtaining the initial cluster center, guides the K-means algorithm to search in the direction of the global optimal solution, further optimizes the clustering results, and greatly reduces the number of iterations of the K-means algorithm, thereby reducing the computational complexity, and avoiding the problem of local optimal solutions, thereby improving the accuracy and stability of clustering. The present invention determines the physical topological connection relationship within the substation under similar voltage fluctuation amplitudes by constructing a membership factor based on the similarity of voltage curves; the higher the membership factor value between the user and the meter box, and between the meter box and the branch box, the stronger the correlation between the corresponding two parties, that is, the two parties have a topological connection relationship. The use of the membership factor is more helpful to intuitively determine the full-level structure of the low-voltage distribution substation; and based on the membership factor, through real-time monitoring and analysis of voltage changes, the physical topological connection relationship within the low-voltage distribution substation can be dynamically evaluated and determined, and the topological structure of the low-voltage distribution substation can be accurately identified in real time, thereby improving the adaptability and reliability of the identification method. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In order to make the content of the present invention more clearly understood, the present invention is further described in detail below according to specific embodiments of the present invention in conjunction with the accompanying drawings, wherein

[0061] Figure 1 It is a flowchart of the steps of the full-level topology identification method of the low-voltage distribution station area provided by the present invention;

[0062] Figure 2 It is a structural schematic diagram of the low-voltage distribution station area provided by the present invention;

[0063] Figure 3 It is a flowchart of specific steps for topology identification of a distribution station area provided by the present invention;

[0064] Figure 4 It is a voltage change trend curve diagram of the meter box and the user provided by the present invention;

[0065] Figure 5 It is the ρ-δ diagram obtained by clustering the phase relationship of the low-voltage distribution station area provided by the present invention;

[0066] Figure 6 It is a peak value descending graph of density of users provided by the present invention;

[0067] Figure 7 It is a user voltage curve diagram of phase A in the three-phase user voltage curve provided by the present invention;

[0068] Figure 8 It is a B-phase user voltage curve diagram among the three-phase user voltage curves provided by the present invention;

[0069] Fig. 9 It is a C-phase user voltage curve diagram among the three-phase user voltage curves provided by the present invention;

[0070] Fig.10 It is a diagram of the affiliation between the branch box and the meter box provided by the present invention;

[0071] Fig.11 It is a hierarchical topological relationship diagram between the branch box, the meter box and the user provided by the present invention. DETAILED DESCRIPTION

[0072] The present invention is further described below in conjunction with the accompanying drawings and specific embodiments so that those skilled in the art can better understand the present invention and implement it, but the embodiments are not intended to limit the present invention.

[0073] Reference Figure 1 As shown, the flowchart of the method for identifying the full-level topology of a low-voltage distribution station area of ​​the present invention includes the following specific steps:

[0074] S101: Obtain voltage sampling data of all branch boxes, meter boxes and users in the low-voltage distribution station area to be identified at each sampling time in a preset sampling period, and form voltage time series data corresponding to each branch box, each meter box and each user;

[0075] S102: Based on the voltage sampling data of all users at any sampling time, clustering is performed using a density peak method to obtain three initial cluster centers corresponding to the three-phase voltages respectively;

[0076] S103: Based on the three initial cluster centers, K-means clustering is used to update the cluster centers and divide the clusters until the clustering objective function converges, and all users are divided into three clusters. According to the phase corresponding to the cluster to which each user belongs, the phase of each user is identified;

[0077] S104: Calculate the membership factors between each user and all meter boxes respectively, obtain the meter box corresponding to the highest membership factor value as the meter box to which the user belongs, and complete the user-meter box topology identification, including:

[0078] According to the standardized user voltage time series data and the meter box voltage time series data, a first mean coefficient is constructed; based on the first mean coefficient, the standard deviation of the user voltage time series data and the standard deviation of the meter box voltage time series data, a first correlation coefficient between the user and the meter box is obtained; based on the first correlation coefficient, a membership factor between the user and the meter box is calculated;

[0079] S105: Calculate the membership factors between each meter box and all branch boxes respectively, obtain the branch box corresponding to the highest membership factor value, and use it as the branch box to which the meter box belongs, and complete the topology identification of the meter box-branch box, including:

[0080] According to the standardized meter box voltage time series data and the branch box voltage time series data, a second mean coefficient is constructed; based on the second mean coefficient, the standard deviation of the meter box voltage time series data and the standard deviation of the branch box voltage time series data, a second correlation coefficient between the meter box and the branch box is obtained; based on the second correlation coefficient, a membership factor between the meter box and the branch box is calculated;

[0081] S106: Based on the phase identification of the user, the topology identification of the user-meter box, and the topology identification of the meter box-branch box, the full-level topology identification of the low-voltage distribution station area to be identified is completed.

[0082] Specifically, in step S102 of the embodiment of the present invention, the step of obtaining the initial cluster center by clustering using the density peak method specifically includes:

[0083] S102-1: Obtain the voltage sampling data of each user at any sampling time in a preset sampling period, and use Gaussian Kernel to calculate the local density of each user, which is expressed as:

[0084]

[0085] Among them, ρ irepresents the local density of the i-th user; d ij Represents the distance between the i-th user and the j-th user, and the expression is: N represents the total number of users in the low-voltage distribution station area to be identified; voltage sampling data set represents the voltage sampling data of the i-th user at the T-th sampling time; d c Indicates the preset cutoff distance.

[0086] S102-2: For each user, obtain users whose Euclidean distance to the user exceeds a preset distance and whose local density is higher than the user's local density, and form a set of users to be clustered for the user;

[0087] S102-3: Determine whether the user set to be clustered is empty:

[0088] If the user set to be clustered is empty, the maximum distance between the user and the remaining users is obtained as the cluster center distance of the user, which is expressed as:

[0089] If the user set to be clustered is not empty, the minimum distance between the user and all users in the user set to be clustered is obtained as the cluster center distance of the user, which is expressed as:

[0090] S102-4: Obtain the density peak value γ of each user based on the product of the local density of each user and the distance to the cluster center i , expressed as: γ i =ρ i ·δ i ;

[0091] S102-5: Sort the density peaks of all users from large to small, and select the users corresponding to the first three density peaks as three initial clustering centers corresponding to the three-phase voltages respectively.

[0092] In this embodiment, the density peak method is used to construct a density peak according to the local density of the user and the distance from the cluster center for clustering. The spatial information and density information of the user voltage data are taken into consideration, and clustering is performed based on the density distribution difference. Since the voltage fluctuation curve has different distribution patterns in different distribution transformer areas, the density difference can accurately reflect the similarities and differences between them. Therefore, the similarities and differences of different voltage fluctuation curves can be accurately identified, ensuring that the initial cluster center obtained by clustering is highly representative, and avoiding the cluster offset and local optimal solution problems that are prone to occur in traditional clustering methods, thereby improving the accuracy and stability of voltage fluctuation curve clustering, thereby improving the accuracy of topology identification.

[0093] Specifically, in step S103, phase recognition is performed on the user, which specifically includes:

[0094] S103-1: Based on the voltage sampling data of all users at any sampling time and the three initial cluster centers, a clustering objective function is constructed, which is expressed as:

[0095] Among them, r ij is a hidden variable, whose value is α k represents the kth cluster center, 1≤k≤K, K represents the number of cluster centers; N represents the total number of users in the low-voltage distribution station area to be identified;

[0096] S103-2: Using K-means clustering, based on the three initial clusters, recalculate the centroid of each cluster as a new cluster center, and divide and generate new clusters;

[0097] New cluster centers:

[0098] C k represents the data set of the kth cluster; x i Represents the voltage sampling data of the i-th user.

[0099] S103-3: Based on the new clustering clusters, repeatedly select new cluster centers until the clustering objective function converges;

[0100] S103 - 4 : Obtain three clusters when the clustering objective function converges, and the phase of the user in each cluster is consistent with the phase of the cluster to which it belongs.

[0101] The method for identifying the full-level topology of a low-voltage distribution station area described in the present invention uses a density peak method for clustering, obtains three initial cluster centers, and then uses K-means clustering to update the cluster centers and divide the cluster clusters, thereby obtaining three cluster clusters corresponding to the three-phase voltages, respectively, and completing the phase identification of each user. The present invention combines the density peak method with the K-means clustering, and after obtaining the initial cluster centers, guides the K-means algorithm to search in the direction of the global optimal solution, further optimizes the clustering results, and greatly reduces the number of iterations of the K-means algorithm, thereby reducing the computational complexity, and avoiding the problem of local optimal solutions, thereby improving the accuracy and stability of clustering.

[0102] Specifically, in step S104, the membership factor between the user and the meter box is calculated, which specifically includes:

[0103] S104-1: Average voltage X based on user voltage time series data mean With standard deviation X std, standardize the user voltage time series data, and obtain the standardized user voltage time series data A t , expressed as:

[0104] S104-2: Mean voltage Y based on meter box voltage time series data mean and standard deviation Y std , standardize the meter box voltage time series data, and obtain the standardized meter box voltage time series data B t , expressed as:

[0105] S104-3: Using the standardized user voltage timing data A t With meter box voltage timing data B t , construct the first mean coefficient C, expressed as:

[0106] S104-4: Based on the first mean coefficient C and the standard deviation X of the user voltage time series data std The standard deviation Y of the meter box voltage time series data std , get the first correlation coefficient R between the user and the table box, expressed as:

[0107] S104-5: Based on the first correlation coefficient R and the minimum value R of the first correlation coefficient min The maximum value of the first correlation coefficient R max , construct the membership factor R between the user and the table box subordinate , expressed as:

[0108] Among them, X t Indicates the voltage sampling data of the user at the tth sampling time within the preset sampling period, Y t represents the voltage sampling data of the meter box at the tth sampling time within the preset sampling period, m represents the total number of sampling times within the preset sampling period; X mean It represents the mean voltage of the user voltage time series data within the preset sampling period, expressed as Y mean It represents the mean voltage of the meter box voltage time series data within the preset sampling period, expressed as X std It represents the standard deviation of the user voltage timing data within the preset sampling period, expressed as Y std It represents the standard deviation of the meter box voltage time series data within the preset sampling period, expressed as

[0109] Specifically, in step S105, the membership factor between the meter box and the branch box is calculated, which specifically includes:

[0110] S105-1: Average voltage Y based on meter box voltage time series data mean and standard deviation Y std , standardize the meter box voltage time series data, and obtain the standardized meter box voltage time series data B t , expressed as:

[0111] S105-2: Mean voltage Z based on branch box voltage time series data mean and standard deviation Z std , standardize the branch box voltage time series data, and obtain the standardized branch box voltage time series data D t , expressed as:

[0112] S105-3: Using the standardized meter box voltage time series data B t With branch box voltage timing data D t , construct the second mean coefficient C', expressed as:

[0113] S105-4: Based on the second mean coefficient C' and the standard deviation Y of the meter box voltage time series data std The standard deviation Z of the branch box voltage time series data std , obtain the second correlation coefficient R' between the meter box and the branch box, expressed as:

[0114] S105-5: Based on the second correlation coefficient R' and the minimum value R' of the second correlation coefficient min The maximum value of the second correlation coefficient R' max , construct the membership factor R' between the user and the table box subordinate , expressed as:

[0115] Among them, Z t It represents the voltage sampling data of the branch box at the tth sampling time within the preset sampling period, Y t represents the voltage sampling data of the meter box at the tth sampling time within the preset sampling period, m represents the total number of sampling times within the preset sampling period; Z mean It represents the mean voltage of the branch box voltage time series data within the preset sampling period, expressed as Y mean It represents the mean voltage of the meter box voltage time series data within the preset sampling period, expressed as Z stdIt represents the standard deviation of the branch box voltage time series data within the preset sampling period, expressed as Y std It represents the standard deviation of the meter box voltage time series data within the preset sampling period, expressed as

[0116] The present invention determines the physical topological connection relationship within the substation under similar voltage fluctuation amplitudes by constructing a membership factor based on the similarity of voltage curves; the higher the membership factor value between the user and the meter box, and between the meter box and the branch box, the stronger the correlation between the corresponding two parties, that is, the two parties have a topological connection relationship. The use of the membership factor is more helpful to intuitively determine the full-level structure of the low-voltage distribution substation; and based on the membership factor, through real-time monitoring and analysis of voltage changes, the physical topological connection relationship within the low-voltage distribution substation can be dynamically evaluated and determined, and the topological structure of the low-voltage distribution substation can be accurately identified in real time, thereby improving the adaptability and reliability of the identification method.

[0117] Specifically, in an embodiment of the present invention, after obtaining the voltage sampling data of all branch boxes, meter boxes and users in the low-voltage distribution station area to be identified at each sampling moment in a preset sampling period, it also includes: preprocessing all voltage sampling data using a Z-Score standardization method.

[0118] Based on the above embodiments, in the embodiments of the present invention, the full-level topology identification method of the low-voltage distribution substation area provided by the present invention is used to select the actual user voltage data of a distribution substation area on a certain day as an embodiment for analysis to identify the hierarchical topology in the distribution substation area; Figure 2 As shown in the figure, it is a schematic diagram of the structure of the low-voltage distribution station area; Figure 3 The figure shows a flowchart of the specific steps for topology identification of a distribution station area.

[0119] The low-voltage distribution area of ​​the power system refers to the low-voltage power supply area of ​​a distribution transformer. The intelligent fusion terminal collects voltage, current and power data every 15 minutes to monitor the status of the total meter in the area, summarizes and analyzes the data uploaded by downstream equipment, and can upload it to the distribution automation system or the power consumption information collection system. The branch box and meter box can collect three-phase voltage, current and power data. On the user side, residential electricity is usually single-phase. When monitoring the current data of the area, there are some users who have not used electricity for a long time, resulting in the current data cannot be used as a basis for identifying the topology. Therefore, collecting voltage data in the area is a better choice. When the low-voltage distribution network is connected to the load, as the load fluctuates, the voltage of the transformer and the user will change accordingly, and the electrical distance under the same area is close, and the voltage fluctuation curve is highly similar.

[0120] In this embodiment, the sampling interval of the smart meter is 15 minutes, and 96 voltage nodes can be collected in one day. According to the collected information, the low-voltage distribution station is divided into four levels: one distribution transformer, four branch boxes X1, X2, X3, X4; 20 meter boxes are numbered Y1, Y2, ... Y 20 ; 120 users are numbered Z1, Z2, ... Z 120 The three-phase voltage data of the meter box and the voltage data change trend curve of all users in one day are combined into a voltage sequence, and the voltage change trend curve is drawn, such as Figure 4 As shown, it can be seen that the voltages in the same area have more similar voltage characteristics.

[0121] S201: extracting and clustering the voltage fluctuation curve based on the multi-dimensional characteristics of the voltage fluctuation of each user using the DPK-means algorithm to identify the phase to which each user belongs;

[0122] The DPK-means algorithm is obtained by optimizing the K-means algorithm using the density peak method;

[0123] S201-1: Input voltage sample data in represents the voltage of the i-th sample at the T-th moment;

[0124] S201-2: Use Gaussian Kernel to calculate the density ρ of sample points i , expressed as:

[0125]

[0126] Among them, d ij Table Sample U i With sample U j The distance between c It represents the cutoff distance. Usually, the data distance of the first 2% position is selected as the cutoff distance.

[0127] S201-3: Get distance sample U i Within the preset distance range and with a density higher than the sample U i The density ρ i The other sample points constitute the set of sample points to be clustered;

[0128] Calculate the sample point U i The minimum distance δ between other sample points with higher density i ; If the sample point U i is the sample point with the largest density, and its set of sample points to be clustered is empty, then the sample point U i The maximum distance between the remaining sample points is set to δi ;

[0129]

[0130] Among them, δ i Represents the sample point U i The minimum distance from all the sample points to be clustered in the set of sample points to be clustered; j represents a point with a higher density than the sample point U i The numbers of the remaining sample points;

[0131] S201-4: Calculate the value of γ and select the cluster center. γ is expressed as: γ i =ρ i *δ i ;

[0132] Among them, the density peak method takes the sample points that meet both the local density and the distance as the cluster centers. The number of cluster centers K and the cluster centers α1, α2, ...α can be determined by the change of the value of γ. K ;

[0133] S201-5: Basis Recalculate the center of each cluster;

[0134] S201-6: Re-divide the clusters and determine the cluster centers;

[0135] S201-7: Repeat steps S201-5 and S201-6 until the cluster center does not change and the value of the objective function is minimized;

[0136] The objective function is expressed as:

[0137] in, α k Represents the cluster center.

[0138] S201-8: Clustering the meter box voltage and user voltage data according to the DPK-means algorithm. The obtained result type can distinguish which users belong to the same phase, and determine which phase the user belongs to according to the voltage data shown in the meter box.

[0139] S202: Calculate the membership factors between the voltage curves at each level in the substation area, and obtain the membership relationship between "branch box-meter box-user" in the distribution substation area, thereby realizing the full-level topology identification of the low-voltage distribution substation area.

[0140] When the voltage amplitude fluctuations of users in the substation area are similar, the introduction of membership factors to determine their topological connection relationship can help to intuitively distinguish the differences. The membership factor can measure the degree of correlation between two sets of data.

[0141] The data elements in the two sets of data X and Y to be compared are normalized respectively, expressed as:

[0142]

[0143] Based on the normalized coefficients, the membership factor is calculated and expressed as:

[0144]

[0145] Wherein, X and Y represent two voltage data groups used to calculate the membership relationship, and each data group contains m data; X mean With Y mean Respectively represent the data means in the two sets of data; X std With Y std Respectively represent the standard deviation of the two data sets; the larger the |R|, the higher the correlation between the two data sets; R subordinate represents the membership factor; R min represents the minimum characteristic value, R max Indicates the maximum value of a feature.

[0146] The variation between voltages in the same substation is smaller than that in different substations. Therefore, when calculating the membership factor, there are values ​​that are too close, which leads to incorrect judgment. In this case, normalizing the data to be compared can improve the visualization of the data. When judging the hierarchical relationship between single substations, the membership factors between the hierarchies are calculated using the normalized expression and arranged in descending order according to the correlation coefficient. The value with the greatest similarity indicates the highest degree of association between the two.

[0147] Specifically, the three-phase voltage data of any meter box and the data of 120 users are clustered using the density peak method. Figure 5 is the ρ-δ diagram obtained by clustering the phase relationship of the low-voltage distribution station area, Figure 6 is a graph of the user's density peak in descending order. Figure 5 There are 3 points with large changes in the density peak value. Figure 6 The three selected cluster centers correspond to the three phases to which the users belong. The phases of the remaining 120 households correspond to these three cluster centers. The clustering results are compared with the original topology results, and the clustering success rate is 100%.

[0148] According to the classification results obtained by the DPK-means clustering algorithm, it is determined that the users correspond to three phases, including 40 users in phase A, 40 users in phase B, and 40 users in phase C. Figure 7 The figure shows the voltage curve of phase A in the three-phase user voltage curve; refer to Figure 8The following is the B-phase user voltage curve in the three-phase user voltage curve; refer to Fig. 9 As shown, it is the C-phase user voltage curve in the three-phase user voltage curve.

[0149] The user set of phase A is {Z3, Z9, Z11, Z13, Z14, Z15, Z17, Z20, Z22, Z28, Z32, Z40, Z41, Z42, Z44, Z49, Z56, Z60, Z65, Z66, Z67, Z68, Z70, Z75, Z79, Z81, Z82, Z83, Z84, Z88, Z90, Z98, Z99, Z102, Z104, Z112, Z114, Z116, Z119, Z120};

[0150] The user set of phase B is {Z1, Z2, Z5, Z7, Z10, Z16, Z23, Z24, Z25, Z27, Z29, Z30, Z34, Z35, Z36, Z37, Z45, Z47, Z50, Z51, Z53, Z54, Z59, Z61, Z62, Z69, Z71, Z76, Z80, Z85, Z86, Z89, Z91, Z92, Z93, Z95, Z103, Z105, Z106, Z107};

[0151] The user set of phase C is {Z4, Z6, Z8, Z12, Z18, Z19, Z21, Z26, Z31, Z33, Z38, Z39, Z43, Z46, Z48, Z52, Z55, Z57, Z57, Z58, Z63, Z64, Z72, Z73, Z74, Z77, Z78, Z87, Z94, Z96, Z97, Z100, Z101, Z108, Z109, Z110, Z111, Z113, Z115, Z117, Z118}.

[0152] When judging the hierarchical relationship between single substations, the membership factors between the hierarchical levels are calculated and arranged in descending order according to the membership degree. The value with the greatest similarity indicates the highest degree of association between the two. Specifically, taking the relationship between the meter box and the branch box as an example, due to the large table, only the normalized membership factors between some meter boxes and branch boxes are listed in the example. The calculation results are shown in Table 1.

[0153] Table 1 Membership factors between some branch boxes and meter boxes

[0154]

[0155] Reference Fig.10As shown, it is a diagram of the affiliation between the branch box and the meter box; the affiliation between the meter boxes Y1, Y6, Y11, Y16 and the branch boxes X1, X2, X3, X4 is determined. The meter box Y1 belongs to the branch box X3, and the similarity coefficient between the two is 9.91. Similarly, it can be determined that Y6 belongs to X2, Y11 belongs to X2, and Y16 belongs to X4. The same method can be used to determine which branch boxes X1-X4 Y1-Y20 belong to.

[0156] When the same method is used to determine the affiliation between the meter box and the user, in the same phase, even if the user does not belong to the meter box to which it belongs, the similarity between them is still very high, while in different phases, even if the user belongs to the meter box, the similarity coefficient between them is very low. Therefore, based on the relationship between the user phases obtained in step 1, the normalized Pearson correlation coefficient between the user belonging to phase A and the meter box is calculated to further determine the affiliation between the user and the meter box.

[0157] In the embodiment, only the normalized Pearson coefficients between some meter boxes and users are still listed, as shown in Table 2.

[0158] Table 2 Membership factors between some meter boxes and users

[0159]

[0160]

[0161] According to the analysis in Table 2, the correlation coefficient between Z3 and Y2 is the highest, which is 9.99. In the normalized Pearson coefficient table between 20 meter boxes and 120 users, the correlation coefficient between Z3 and Y2 is also the highest. Therefore, it can be judged that Z3 belongs to the branch under the Y2 meter box; similarly, Z99 belongs to Y3, Z17 belongs to Y1, and Z67 belongs to Y5. The correlation coefficient between Z13 and Y2 is 9.69, which is the highest value in Table 2. However, in the normalized Pearson coefficient table, the meter box with the highest correlation coefficient with Z13 is Y14. Therefore, Z13 does not belong to Y2, but to Y14. It is only because Y2 and Y14 belong to the same branch box X1 that the correlation coefficient between Z13 and Y2 is higher than that of other meter boxes; similarly, Z20 does not belong to Y3, which has the highest similarity with it in Table 2, but belongs to Y20, Z90 belongs to Y12, and Z98 belongs to Y20.

[0162] The users of phase B and phase C divided in step S201 are calculated again according to the above steps to obtain the affiliation relationship between the remaining users and the meter boxes.

[0163] In summary, after the above steps, the normalized Pearson coefficients between the branch box and the meter box, and between the meter box and the user are calculated respectively, and the full-level topological relationship between "branch box-meter box-user-phase" can be correctly obtained, such as Fig.11 Comparing this topology with the original topology of the original area can obtain 100% accuracy.

[0164] The present invention extracts and utilizes the multidimensional characteristics of user voltage fluctuations, uses the DPK-Means algorithm to cluster the voltage fluctuation curves, and identifies the phase to which each user belongs. Secondly, the Pearson correlation coefficient between the voltage curves at each level in the substation area is calculated and normalized, and the affiliation between the "branch box-meter box-user" in the distribution station area is obtained, so as to realize the full-level topology identification of the low-voltage distribution station area, overcome the shortcomings of K-means that the number of clusters needs to be set in advance and the random selection of the initial cluster center may cause the result to fall into the local optimum, improve the stability of the clustering results, and realize the full-level topology identification of the low-voltage distribution station area.

[0165] Based on the above embodiments, in an embodiment of the present invention, a device for identifying a full-level topology of a low-voltage distribution station area is also provided, including:

[0166] The acquisition module 100 is used to acquire the voltage sampling data of all branch boxes, meter boxes and users in the low-voltage distribution station area to be identified at each sampling time in a preset sampling period, and form the voltage time series data corresponding to each branch box, each meter box and each user;

[0167] The phase identification module 200 is used to cluster the voltage sampling data of all users at any sampling time using the density peak method to obtain three initial cluster centers corresponding to the three-phase voltages respectively; based on the three initial cluster centers, K-means clustering is used to update the cluster centers and divide the cluster clusters until the clustering objective function converges, and all users are divided into three cluster clusters. According to the phase corresponding to the cluster cluster to which each user belongs, the phase identification of each user is completed;

[0168] The user-meter box topology identification module 300 is used to respectively calculate the membership factor between each user and all the meter boxes, obtain the meter box corresponding to the highest membership factor value as the meter box to which the user belongs, and complete the user-meter box topology identification, including: constructing a first mean coefficient according to the standardized user voltage time series data and the meter box voltage time series data; obtaining a first correlation coefficient between the user and the meter box based on the first mean coefficient, the standard deviation of the user voltage time series data and the standard deviation of the meter box voltage time series data; and calculating the membership factor between the user and the meter box based on the first correlation coefficient;

[0169] The meter box-branch box topology identification module 400 is used to respectively calculate the membership factors between each meter box and all branch boxes, obtain the branch box corresponding to the highest membership factor value as the branch box to which the meter box belongs, and complete the topology identification of the meter box-branch box, including: constructing a second mean coefficient according to the standardized meter box voltage time series data and the branch box voltage time series data; obtaining a second correlation coefficient between the meter box and the branch box based on the second mean coefficient, the standard deviation of the meter box voltage time series data and the standard deviation of the branch box voltage time series data; and calculating the membership factor between the meter box and the branch box based on the second correlation coefficient;

[0170] The full-level topology identification module 500 is used to complete the full-level topology identification of the low-voltage distribution station area to be identified based on the phase identification of the user, the topology identification of the user-meter box, and the topology identification of the meter box-branch box.

[0171] The low-voltage distribution station area full-level topology identification device of this embodiment is used to implement the aforementioned low-voltage distribution station area full-level topology identification method. Therefore, the specific implementation method of the low-voltage distribution station area full-level topology identification device can be seen in the implementation method part of the low-voltage distribution station area full-level topology identification method in the previous text, for example, the acquisition module 100 is used to implement step S101 in the above-mentioned low-voltage distribution station area full-level topology identification method; the phase identification module 200 is used to implement steps S102 and S103 in the above-mentioned low-voltage distribution station area full-level topology identification method; the user-meter box topology identification module 300, the meter box-branch box topology identification module 400, and the full-level topology identification module 500 are respectively used to implement steps S104, S105 and S106 in the above-mentioned low-voltage distribution station area full-level topology identification method. Therefore, its specific implementation method can refer to the description of the corresponding each part of the implementation method, which will not be repeated here.

[0172] The method for identifying the full-level topology of a low-voltage distribution station area described in the present invention uses a density peak method for clustering. After obtaining three initial cluster centers, K-means clustering is used to update the cluster centers and divide the cluster clusters, and three cluster clusters corresponding to three-phase voltages are obtained to complete the phase identification of each user; the density peak method is used to construct a density peak according to the local density of the user and the distance from the cluster center for clustering, and the spatial information and density information of the user's voltage data are taken into account, and the similarities and differences of different voltage fluctuation curves can be accurately identified, thereby improving the accuracy of topology identification; the present invention combines the density peak method with K-means clustering, and after obtaining the initial cluster center, guides the K-means algorithm to search in the direction of the global optimal solution, further optimizes the clustering results, and greatly reduces the number of iterations of the K-means algorithm, thereby reducing the computational complexity, and avoiding the problem of local optimal solutions, thereby improving the accuracy and stability of clustering. The present invention determines the physical topological connection relationship within the substation under similar voltage fluctuation amplitudes by constructing a membership factor based on the similarity of voltage curves; the higher the membership factor value between the user and the meter box, and between the meter box and the branch box, the stronger the correlation between the corresponding two parties, that is, the two parties have a topological connection relationship. The use of the membership factor is more helpful to intuitively determine the full-level structure of the low-voltage distribution substation; and based on the membership factor, through real-time monitoring and analysis of voltage changes, the physical topological connection relationship within the low-voltage distribution substation can be dynamically evaluated and determined, and the topological structure of the low-voltage distribution substation can be accurately identified in real time, realizing the full-level topological recognition of "branch box-meter box-user-phase", thereby improving the adaptability and reliability of the recognition method.

[0173] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0174] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0175] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0176] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0177] Obviously, the above embodiments are merely examples for the purpose of clear explanation and are not intended to limit the implementation methods. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the implementation methods here. The obvious changes or modifications derived therefrom are still within the scope of protection of the present invention.

Claims

1. A method for identifying the full-level topology of a low-voltage distribution station area, characterized in that: include: Obtain the voltage sampling data of all branch boxes, meter boxes and users in the low-voltage distribution station area to be identified at each sampling time in the preset sampling period, and form the voltage time series data corresponding to each branch box, each meter box and each user; Based on the voltage sampling data of all users at any sampling time, the density peak method is used for clustering to obtain three initial cluster centers corresponding to the three-phase voltages respectively; Based on the three initial cluster centers, K-means clustering is used to update the cluster centers and divide the clusters until the clustering objective function converges. All users are divided into three clusters, and the phase of each user is identified according to the phase corresponding to the cluster to which each user belongs. Calculate the membership factors between each user and all meter boxes respectively, obtain the meter box corresponding to the highest membership factor value as the meter box to which the user belongs, and complete the user-meter box topology identification, including: constructing a first mean coefficient according to the standardized user voltage time series data and the meter box voltage time series data; obtaining a first correlation coefficient between the user and the meter box based on the first mean coefficient, the standard deviation of the user voltage time series data and the standard deviation of the meter box voltage time series data; and calculating the membership factor between the user and the meter box based on the first correlation coefficient; specifically including: The average voltage X based on the user's voltage time series data mean With standard deviation X std , standardize the user voltage time series data, and obtain the standardized user voltage time series data A t , expressed as: The mean voltage Y based on the voltage time series data of the meter box mean and standard deviation Y std , standardize the meter box voltage time series data, and obtain the standardized meter box voltage time series data B t , expressed as: Using the standardized user voltage timing data A t With meter box voltage timing data B t , construct the first mean coefficient C, expressed as: Based on the first mean coefficient C, the standard deviation X of the user voltage time series data std The standard deviation Y of the meter box voltage time series data std , get the first correlation coefficient R between the user and the table box, expressed as: Based on the first correlation coefficient R, the minimum value R of the first correlation coefficient min The maximum value of the first correlation coefficient R max , construct the membership factor R between the user and the table box subordinate , expressed as: Among them, X t Indicates the voltage sampling data of the user at the tth sampling time within the preset sampling period, Y t represents the voltage sampling data of the meter box at the tth sampling time within the preset sampling period, m represents the total number of sampling times within the preset sampling period; X mean It represents the mean voltage of the user voltage time series data within the preset sampling period, expressed as Y mean It represents the mean voltage of the meter box voltage time series data within the preset sampling period, expressed as X std It represents the standard deviation of the user voltage timing data within the preset sampling period, expressed as Y std It represents the standard deviation of the meter box voltage time series data within the preset sampling period, expressed as Calculate the membership factors between each meter box and all branch boxes respectively, obtain the branch box corresponding to the highest membership factor value as the branch box to which the meter box belongs, and complete the topology identification of the meter box-branch box, including: constructing a second mean coefficient according to the standardized meter box voltage time series data and the branch box voltage time series data; obtaining a second correlation coefficient between the meter box and the branch box based on the second mean coefficient, the standard deviation of the meter box voltage time series data and the standard deviation of the branch box voltage time series data; calculating the membership factor between the meter box and the branch box based on the second correlation coefficient; specifically including: The mean voltage Y based on the voltage time series data of the meter box mean and standard deviation Y std , standardize the meter box voltage time series data, and obtain the standardized meter box voltage time series data B t , expressed as: Mean voltage Z based on branch box voltage time series data mean and standard deviation Z std , standardize the branch box voltage time series data, and obtain the standardized branch box voltage time series data D t , expressed as: Using the standardized meter box voltage time series data B t With branch box voltage timing data D t , construct the second mean coefficient C', expressed as: Based on the second mean coefficient C', the standard deviation Y of the meter box voltage time series data std The standard deviation Z of the branch box voltage time series data std , obtain the second correlation coefficient R' between the meter box and the branch box, expressed as: Based on the second correlation coefficient R', the minimum value R' of the second correlation coefficient min The maximum value of the second correlation coefficient R' max , construct the membership factor R' between the table box and the branch box subordinate , expressed as: Among them, Z t It represents the voltage sampling data of the branch box at the tth sampling time within the preset sampling period, Y t represents the voltage sampling data of the meter box at the tth sampling time within the preset sampling period, m represents the total number of sampling times within the preset sampling period; Z mean It represents the mean voltage of the branch box voltage time series data within the preset sampling period, expressed as Y mean It represents the mean voltage of the meter box voltage time series data within the preset sampling period, expressed as Z std It represents the standard deviation of the branch box voltage time series data within the preset sampling period, expressed as Y std It represents the standard deviation of the meter box voltage time series data within the preset sampling period, expressed as Based on the user's phase identification, user-meter box topology identification, and meter box-branch box topology identification, the full-level topology identification of the low-voltage distribution station area to be identified is completed.

2. The method for identifying the full-level topology of a low-voltage distribution station area according to claim 1 is characterized in that: The voltage sampling data of all users at any sampling time is clustered using the density peak method to obtain three initial cluster centers corresponding to the three-phase voltages, including: Obtain the voltage sampling data of each user at any sampling time in a preset sampling period, and use Gaussian Kernel to calculate the local density of each user; For each user, obtain users whose Euclidean distance exceeds a preset distance and whose local density is higher than that of the user, and form a set of users to be clustered for the user; If the user set to be clustered of the user is empty, the maximum distance between the user and the remaining users is obtained as the cluster center distance of the user; if the user set to be clustered of the user is not empty, the minimum distance between the user and all users in the user set to be clustered is obtained as the cluster center distance of the user; Based on the product of each user's local density and the distance to the cluster center, the density peak of each user is obtained; The density peaks of all users are sorted from large to small, and the users corresponding to the first three density peaks are selected as the three initial clustering centers corresponding to the three-phase voltages respectively.

3. The method for identifying the full-level topology of a low-voltage distribution station area according to claim 2 is characterized in that: The local density is expressed as: Among them, ρ i represents the local density of the i-th user; d ij Represents the distance between the i-th user and the j-th user, and the expression is: N represents the total number of users in the low-voltage distribution station area to be identified; voltage sampling data set represents the voltage sampling data of the i-th user at the T-th sampling time; d c Indicates the preset cutoff distance.

4. The method for identifying the full-level topology of a low-voltage distribution station area according to claim 3 is characterized in that: The cluster center distance includes: If the set of users to be clustered is empty, the maximum distance between the user and the other users is obtained as the cluster center distance δ of the user i , expressed as: If the set of users to be clustered is not empty, the minimum distance between the user and all users in the set of users to be clustered is obtained as the cluster center distance δ of the user i , expressed as:

5. The method for identifying the full-level topology of a low-voltage distribution station area according to claim 2 is characterized in that: Based on the three initial cluster centers, K-means clustering is used to update the cluster centers and divide the clusters until the clustering objective function converges, and all users are divided into three clusters. According to the phase corresponding to the cluster to which each user belongs, the phase identification of each user is completed, including: Based on the voltage sampling data of all users at any sampling time and three initial cluster centers, a clustering objective function is constructed; Using K-means clustering, based on the three initial clusters, recalculate the centroid of each cluster as the new cluster center, and divide and generate new clusters; Based on the new clusters, new cluster centers are repeatedly selected until the clustering objective function converges; Three clusters are obtained when the clustering objective function converges, and the phase of the users in each cluster is consistent with the phase of the cluster to which it belongs.

6. The method for identifying the full-level topology of a low-voltage distribution station area according to claim 5 is characterized in that: The method constructs a clustering objective function based on the voltage sampling data of all users at any sampling time and the three initial clustering centers, and uses K-means clustering to recalculate the centroid of each cluster as a new clustering center based on the three initial clustering clusters, including: The clustering objective function is expressed as: Among them, r ij is a hidden variable, whose value is α k represents the kth cluster center, 1≤k≤K, K represents the number of cluster centers; N represents the total number of users in the low-voltage distribution station area to be identified; Using K-means clustering, based on the three initial clusters, recalculate the centroid of each cluster as the new cluster center: C k represents the data set of the kth cluster; x i Represents the voltage sampling data of the i-th user.

7. The method for identifying the full-level topology of a low-voltage distribution station area according to claim 1 is characterized in that: After obtaining the voltage sampling data of all branch boxes, meter boxes and users in the low-voltage distribution station area to be identified at each sampling moment in the preset sampling period, the method also includes: preprocessing all voltage sampling data using a Z-Score standardization method.

8. A device for identifying the full-level topology of a low-voltage distribution station area, characterized in that: include: An acquisition module is used to acquire the voltage sampling data of all branch boxes, meter boxes and users in the low-voltage distribution station area to be identified at each sampling moment in a preset sampling period, and to form voltage time series data corresponding to each branch box, each meter box and each user; The phase identification module is used to cluster the voltage sampling data of all users at any sampling time using the density peak method to obtain three initial cluster centers corresponding to the three-phase voltages respectively; based on the three initial cluster centers, K-means clustering is used to update the cluster centers and divide the cluster clusters until the clustering objective function converges, and all users are divided into three cluster clusters. According to the phase corresponding to the cluster cluster to which each user belongs, the phase identification of each user is completed; specifically, it includes: The average voltage X based on the user's voltage time series data mean With standard deviation X std , standardize the user voltage time series data, and obtain the standardized user voltage time series data A t , expressed as: The mean voltage Y based on the voltage time series data of the meter box mean and standard deviation Y std , standardize the meter box voltage time series data, and obtain the standardized meter box voltage time series data B t , expressed as: Using the standardized user voltage timing data A t With meter box voltage timing data B t , construct the first mean coefficient C, expressed as: Based on the first mean coefficient C, the standard deviation X of the user voltage time series data std The standard deviation Y of the meter box voltage time series data std , get the first correlation coefficient R between the user and the table box, expressed as: Based on the first correlation coefficient R, the minimum value R of the first correlation coefficient min The maximum value of the first correlation coefficient R max , construct the membership factor R between the user and the table box subordinate , expressed as: Among them, X t Indicates the voltage sampling data of the user at the tth sampling time within the preset sampling period, Y t represents the voltage sampling data of the meter box at the tth sampling time within the preset sampling period, m represents the total number of sampling times within the preset sampling period; X mean It represents the mean voltage of the user voltage time series data within the preset sampling period, expressed as Y mean It represents the mean voltage of the meter box voltage time series data within the preset sampling period, expressed as X std It represents the standard deviation of the user voltage timing data within the preset sampling period, expressed as Y std It represents the standard deviation of the meter box voltage time series data within the preset sampling period, expressed as A user-meter box topology identification module is used to respectively calculate the membership factor between each user and all meter boxes, obtain the meter box corresponding to the highest membership factor value as the meter box to which the user belongs, and complete the user-meter box topology identification, including: constructing a first mean coefficient according to the standardized user voltage time series data and the meter box voltage time series data; obtaining a first correlation coefficient between the user and the meter box based on the first mean coefficient, the standard deviation of the user voltage time series data and the standard deviation of the meter box voltage time series data; and calculating the membership factor between the user and the meter box based on the first correlation coefficient; The meter box-branch box topology identification module is used to respectively calculate the membership factors between each meter box and all branch boxes, obtain the branch box corresponding to the highest membership factor value as the branch box to which the meter box belongs, and complete the topology identification of the meter box-branch box, including: constructing a second mean coefficient according to the standardized meter box voltage time series data and the branch box voltage time series data; obtaining a second correlation coefficient between the meter box and the branch box based on the second mean coefficient, the standard deviation of the meter box voltage time series data and the standard deviation of the branch box voltage time series data; calculating the membership factor between the meter box and the branch box based on the second correlation coefficient; specifically including: The mean voltage Y based on the voltage time series data of the meter box mean and standard deviation Y std , standardize the meter box voltage time series data, and obtain the standardized meter box voltage time series data B t , expressed as: Mean voltage Z based on branch box voltage time series data mean and standard deviation Z std , standardize the branch box voltage time series data, and obtain the standardized branch box voltage time series data D t , expressed as: Using the standardized meter box voltage time series data B t With branch box voltage timing data D t , construct the second mean coefficient C', expressed as: Based on the second mean coefficient C', the standard deviation Y of the meter box voltage time series data std The standard deviation Z of the branch box voltage time series data std , obtain the second correlation coefficient R' between the meter box and the branch box, expressed as: Based on the second correlation coefficient R', the minimum value R' of the second correlation coefficient min The maximum value of the second correlation coefficient R' max , construct the membership factor R' between the table box and the branch box subordinate , expressed as: Among them, Z t It represents the voltage sampling data of the branch box at the tth sampling time within the preset sampling period, Y t represents the voltage sampling data of the meter box at the tth sampling time within the preset sampling period, m represents the total number of sampling times within the preset sampling period; Z mean It represents the mean voltage of the branch box voltage time series data within the preset sampling period, expressed as Y mean It represents the mean voltage of the meter box voltage time series data within the preset sampling period, expressed as Z std It represents the standard deviation of the branch box voltage time series data within the preset sampling period, expressed as Y std It represents the standard deviation of the meter box voltage time series data within the preset sampling period, expressed as The full-level topology identification module is used to complete the full-level topology identification of the low-voltage distribution station area to be identified based on the user's phase identification, the user-meter box topology identification, and the meter box-branch box topology identification.

Images