A method and device for identifying low voltage area topology

Abstract

The present invention relates to the technical field of substation topology identification, and discloses a low-voltage substation topology identification method and device, including: obtaining the original voltage data sequence of each terminal to be identified in the low-voltage substation, calculating and splicing the first-order differential coefficient sequence and the voltage peak coefficient sequence of each terminal to be identified, and obtaining a characteristic data sequence; based on the characteristic data sequence of each terminal to be identified in the low-voltage substation, presetting the differential weight and the peak weight, calculating the weighted Euclidean distance between any two terminals to be identified in the low-voltage substation, calculating the local density and sample distance of each terminal to be identified, performing DPC clustering on all terminals to be identified in the low-voltage substation, and obtaining multiple clustering clusters; based on the characteristic data sequence, using the Kendall correlation coefficient calculation method, constructing a correlation matrix for each clustering cluster; based on the correlation matrix corresponding to each clustering cluster, the affiliation of all terminals to be identified in the clustering cluster, identifying and obtaining the low-voltage substation topology structure.

Description

Technical Field

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

[0002] The low-voltage distribution network is at the end of the power grid and directly serves users. However, due to the complex wiring structure of the low-voltage distribution station, frequent line changes caused by new additions, cancellations, relocations, etc., and the lack of effective management and control measures and technical means, the update of the station topology records is slow or even erroneous.

[0003] At present, the methods for identifying the topology of low-voltage distribution networks can be divided into the injection signal method, the data label method, the data-driven method, etc. The first two methods require a large amount of equipment to be invested in the early stage to collect the required physical signals. Although they are highly accurate, they are often difficult to implement on a large scale in low-voltage distribution substations. The data-driven method only needs to collect data such as voltage, current, and power returned by smart meters, and analyze and process them through mathematical methods to realize phase relationship identification and affiliation identification. Among them, using voltage data as the main analysis object is more efficient and quick.

[0004] However, the existing low-voltage distribution station area topology structure identification method driven by voltage data cannot correctly obtain the relevant features in the voltage data because it directly uses voltage data for clustering prediction. As a result, the phase relationship and affiliation between each user and the meter box identified after clustering based on voltage data are incorrect, and the affiliation of the power topology level is unclear. Inaccurate area topology makes it impossible to accurately obtain the operating status of the area, and it is impossible to perform effective energy scheduling and optimization, which affects the rational use of energy. Summary of the invention

[0005] To this end, the technical problem to be solved by the present invention is to overcome the problem in the prior art that the data features of the voltage data cannot be correctly obtained, resulting in errors in the phase relationship and affiliation between each user and the meter box identified after clustering based on the voltage data.

[0006] In order to solve the above technical problems, the present invention provides a low voltage area topology identification method, comprising:

[0007] Obtaining the original voltage data of each terminal to be identified in the low-voltage area at each moment within a preset time period to form an original voltage data sequence of each terminal to be identified; the terminal to be identified includes a user to be identified, a meter box to be identified, and a branch box to be identified;

[0008] Calculate the difference between the original voltage data at each moment in the original voltage data sequence and the original voltage data at the previous moment as the first-order differential coefficient of the original voltage data at each moment; obtain the first-order differential coefficient sequence of each terminal to be identified based on the first-order differential coefficient of the original voltage data at each moment of each terminal to be identified;

[0009] The original voltage data sequence is divided into multiple subsequences, and the ratio of the voltage peak value to the root mean square in each subsequence is calculated as the voltage peak coefficient of each subsequence; the voltage peak coefficient sequence of the original voltage data sequence is composed of the voltage peak coefficients of each subsequence, and the voltage peak coefficient sequence of each terminal to be identified is obtained;

[0010] The first-order differential coefficient sequence of each terminal to be identified is concatenated with the voltage peak coefficient sequence to obtain a characteristic data sequence of each terminal to be identified;

[0011] Based on the characteristic data sequence of each terminal to be identified in the low-voltage area, the differential weight and the peak weight are preset, and the weighted Euclidean distance between any two terminals to be identified in the low-voltage area is calculated;

[0012] Calculate the local density of each terminal to be identified based on the weighted Euclidean distance between each terminal to be identified and all other terminals to be identified;

[0013] Based on the local density of each terminal to be identified and the weighted Euclidean distance between each terminal to be identified and other terminals to be identified, a sample distance of each terminal to be identified is obtained;

[0014] Based on the local density and sample distance of each terminal to be identified, DPC clustering is performed on all terminals to be identified in the low-voltage area, and terminals to be identified with the same phase are divided into the same cluster to obtain multiple clusters;

[0015] For each cluster, based on the characteristic data sequences of all terminals to be identified in the cluster, the Kendall correlation coefficient calculation method is used to calculate the correlation between any two terminals to be identified in the cluster, and a correlation matrix corresponding to the cluster is constructed;

[0016] Based on the correlation matrix corresponding to each cluster, the affiliation between each meter box to be identified and the branch box to be identified, as well as the affiliation between each user to be identified and the meter box to be identified are obtained; based on the affiliation of all terminals to be identified in all clusters, the low-voltage substation topology structure is identified and obtained.

[0017] Preferably, obtaining a first-order differential coefficient sequence of a terminal to be identified includes:

[0018] Get the The original voltage data sequence of the terminal to be identified is expressed as ;

[0019] Calculate the The original voltage data at the moment The original voltage data of the previous moment The difference between The first-order difference coefficient of the original voltage data at the moment , expressed as: , , Indicates the number of voltage value samples in the original voltage data sequence;

[0020] Based on The first-order differential coefficient of the original voltage data at each moment in the original voltage data sequence of the terminal to be identified is obtained. The first-order difference coefficient sequence corresponding to the original voltage data sequence of the terminal to be identified is expressed as: .

[0021] Preferably, obtaining a voltage peak coefficient sequence of a terminal to be identified includes:

[0022] The first The original voltage data sequence of the terminal to be identified Divide into raw voltage data sets, each of which contains The original voltage data, , Indicates the number of voltage value samples in the original voltage data sequence;

[0023] Get the first The peak voltage of the raw voltage data set , expressed as: , Indicates The first of the raw voltage data sets data, ; Get the voltage peak value of each original voltage data group to form a voltage peak sequence , expressed as: ;

[0024] Calculate the first The RMS value of the raw voltage data set , expressed as: ; Get the RMS of each raw voltage data group to form an RMS sequence , expressed as: ;

[0025] Calculate the first The ratio of the voltage peak value to the RMS value of the original voltage data group is obtained. The voltage peak factor of the original voltage data set is expressed as: ;

[0026] Based on The voltage peak sequence and RMS sequence of the terminal to be identified are obtained. The voltage peak coefficient sequence of the terminals to be identified is expressed as: .

[0027] Preferably, the first-order differential coefficient sequence of each terminal to be identified is concatenated with the voltage peak coefficient sequence to obtain the characteristic data sequence of each terminal to be identified, which is expressed as:

[0028] ;

[0029] in, Indicates The first-order difference coefficient sequence of the terminal to be identified, the sequence length is ; Indicates The voltage peak coefficient sequence of the terminals to be identified is ;No. The length of the characteristic data sequence of the terminal to be identified .

[0030] Preferably, the method of calculating the weighted Euclidean distance between any two terminals to be identified in the low-voltage area based on the feature data sequence of each terminal to be identified in the low-voltage area, presetting the differential weight and the peak weight, and comprising:

[0031] Get the characteristic data sequence of each terminal to be identified in the low-voltage area, expressed as: ;

[0032] Preset differential weights With peak weight , using the weighted Euclidean distance calculation method, calculate the distance between any two terminals to be identified in the low-voltage area and The weighted Euclidean distance between , expressed as:

[0033] ;

[0034] in, Indicates the total number of elements in the characteristic data sequence of the terminal to be identified, ; ; When calculating the first-order difference coefficient sequence in the characteristic data sequence, that is, hour, , represents the difference weight, ; When calculating the voltage peak coefficient sequence in the characteristic data sequence, that is, hour, , represents the peak weight, .

[0035] Preferably, the local density of each terminal to be identified is calculated based on the weighted Euclidean distance between each terminal to be identified and all other terminals to be identified, which is expressed as:

[0036] ;

[0037] in, Indicates The local density of terminals to be identified, Indicates the preset cutoff distance.

[0038] Preferably, the acquiring the sample distance of each terminal to be identified based on the local density of each terminal to be identified and the weighted Euclidean distance between each terminal to be identified and other terminals to be identified includes:

[0039] Get local density Higher than the local density of the terminal to be identified All terminals to be identified , forming a candidate set;

[0040] If the set to be selected is empty, the maximum distance value among the weighted Euclidean distances from the current terminal to be identified to the other terminals to be identified in the low-voltage area is obtained as the sample distance of the current terminal to be identified. , expressed as:

[0041] ;

[0042] If the candidate set is not empty, then in the candidate set, select the distance value with the minimum weighted Euclidean distance to the current terminal to be identified as the sample distance of the current terminal to be identified. , expressed as:

[0043] .

[0044] Preferably, based on the local density and sample distance of each terminal to be identified, DPC clustering is performed on all terminals to be identified in the low-voltage area, and terminals to be identified with the same phase are divided into the same cluster to obtain multiple clusters, including:

[0045] Based on the local density and sample distance of each terminal to be identified, a clustering discriminant value is constructed, which is expressed as:

[0046] ;

[0047] Based on cluster discriminant value , draw a decision diagram and obtain the initial number of clusters With each cluster center ;

[0048] For each terminal to be identified in the low-voltage area, divide it into the cluster where the cluster center closest to it is located, and obtain Clusters;

[0049] The cluster to which each terminal to be identified belongs is expressed as: .

[0050] Preferably, the method of calculating the correlation between any two terminals to be identified in the cluster based on the characteristic data sequences of all the terminals to be identified in the cluster and using the Kendall correlation coefficient calculation method to construct a correlation matrix corresponding to the cluster includes:

[0051] Taking a characteristic data sequence of any terminal to be identified in the cluster as a first observation value sequence, and a characteristic data sequence of another terminal to be identified as a second observation value sequence;

[0052] Obtaining the total number of pairs of same order and the total number of pairs of different order in the first observation value sequence and the second observation value sequence, and calculating the Kendall correlation coefficient between the two terminals to be identified as the correlation between the two terminals to be identified;

[0053] Obtain the correlation between any two terminals to be identified in the cluster, and construct a correlation matrix containing the correlations between all terminals to be identified in the cluster, expressed as:

[0054] ;

[0055] in, Indicates the cluster The terminal to be identified and Kendall correlation coefficient between the terminals to be identified; {user to be identified, table box to be identified}, {meter box to be identified, branch box to be identified}; when When the user is to be identified, The meter box is to be identified. At this time, Represents the total number of users to be identified in the cluster. Represents the total number of boxes to be identified in the cluster; when When the meter box is to be identified, The branch box is to be identified. At this time, Represents the total number of bins to be identified in the cluster. Indicates the total number of branch boxes to be identified in the cluster.

[0056] Preferably, the obtaining of the total number of pairs of the same order and the total number of pairs of the different order in the first observation value sequence and the second observation value sequence, and calculating the Kendall correlation coefficient between the terminals to be identified includes:

[0057] Get the characteristic data sequence of any terminal to be identified As the first observation sequence, it is expressed as: , The length of the characteristic data sequence of the terminal to be identified;

[0058] Get the characteristic data sequence of another terminal to be identified , as the second observation sequence, is expressed as: ;

[0059] The total number of pairs of identical order in the first observation value sequence and the second observation value sequence is obtained, expressed as: ;

[0060] When the first observation sequence Elements With Elements , and the first Elements With Elements The size relationship is and , or and , then the element , , and Form an identical sequence pair of the first observation value sequence and the second observation value sequence;

[0061] The total number of pairs of different order in the first observation value sequence and the second observation value sequence is obtained, expressed as: ;

[0062] When the first observation sequence Elements With Elements , and the first Elements With Elements The size relationship is and , or and , then the element , , and A different sequence pair is formed between the first observation value sequence and the second observation value sequence;

[0063] in, represents a symbolic function, when When it is greater than 0, the sign function value is 1. When it is equal to 0, the sign function value is 0. When it is less than 0, the sign function value is -1;

[0064] When the first observation value sequence and the second observation value sequence do not exist or When , the Kendall correlation coefficient between the two terminals to be identified is expressed as: ;

[0065] When the first observation value sequence and the second observation value sequence exist or When , the Kendall correlation coefficient between the two terminals to be identified is expressed as: ;

[0066] in, , , ;

[0067] In the first observation sequence, all The same elements of the same element set form the same element set, and all the same element sets are constructed into the first total set. The total number of the same element sets in the first total set is ; In the second observation sequence, all The same elements of form the same element set, and all the same element sets are constructed into the second total set. The total number of the same element sets in the second total set is ; and Respectively represent the first total set and the second total set The number of elements contained in a set of identical elements.

[0068] This embodiment also provides a low-voltage area topology identification device, including:

[0069] A data acquisition module is used to acquire the original voltage data of each terminal to be identified in the low-voltage area at each moment within a preset time period to form an original voltage data sequence of each terminal to be identified; the terminal to be identified includes a user to be identified, a meter box to be identified, and a branch box to be identified;

[0070] The differential coefficient calculation module is used to calculate the difference between the original voltage data at each moment in the original voltage data sequence and the original voltage data at the previous moment as the first-order differential coefficient of the original voltage data at each moment; based on the first-order differential coefficient of the original voltage data of each terminal to be identified at each moment, obtain the first-order differential coefficient sequence of each terminal to be identified;

[0071] The peak coefficient calculation module is used to divide the original voltage data sequence into multiple subsequences, calculate the ratio of the voltage peak value to the root mean square in each subsequence, and use it as the voltage peak coefficient of each subsequence; the voltage peak coefficient sequence of the original voltage data sequence is composed of the voltage peak coefficients of each subsequence, and the voltage peak coefficient sequence of each terminal to be identified is obtained;

[0072] A feature data construction module is used to concatenate the first-order differential coefficient sequence of each terminal to be identified with the voltage peak coefficient sequence to obtain a feature data sequence of each terminal to be identified;

[0073] A clustering module is used to calculate the weighted Euclidean distance between any two terminals to be identified in the low-voltage area based on the feature data sequence of each terminal to be identified in the low-voltage area, preset differential weights and peak weights; calculate the local density of each terminal to be identified based on the weighted Euclidean distance between each terminal to be identified and all other terminals to be identified; obtain the sample distance of each terminal to be identified based on the local density of each terminal to be identified and the weighted Euclidean distance between each terminal to be identified and other terminals to be identified; perform DPC clustering on all terminals to be identified in the low-voltage area based on the local density and sample distance of each terminal to be identified, divide the terminals to be identified with the same phase into the same cluster cluster, and obtain multiple cluster clusters;

[0074] A correlation matrix construction module is used to calculate the correlation between any two terminals to be identified in each cluster based on the characteristic data sequences of all terminals to be identified in the cluster using the Kendall correlation coefficient calculation method, and to construct a correlation matrix corresponding to the cluster;

[0075] The topology identification module is used to obtain the affiliation between each meter box to be identified and the branch box to be identified, as well as the affiliation between each user to be identified and the meter box to be identified based on the correlation matrix corresponding to each cluster; based on the affiliation of all terminals to be identified in all clusters, identify and obtain the low-voltage substation topology structure.

[0076] The above technical solution of the present invention has the following beneficial effects compared with the prior art:

[0077] The low-voltage substation topology identification method described in the present invention obtains the first-order differential coefficient sequence and the voltage peak coefficient sequence of each terminal to be identified based on the original voltage data sequence of each terminal to be identified to construct a feature data sequence; the first-order differential coefficient can reflect the sequence fluctuation, and the voltage peak coefficient can reflect the sequence morphology. By combining the two features together, a more comprehensive feature representation capability and a richer data representation capability are provided, thereby enhancing the performance and generalization capability of topology identification. Based on the feature data sequence of each terminal to be identified, the present invention calculates the weighted Euclidean distance to obtain the local density and sample distance of each terminal to be identified, uses the DPC clustering method to obtain the phase of each terminal to be identified, and uses the Kendall correlation coefficient to calculate the affiliation; when the original voltage data of the same substation is extremely similar, the feature data constructed by the present invention can also better reflect the differences between different voltage curves, and improve the efficiency and accuracy of low-voltage substation topology identification in phase identification using DPC clustering and affiliation identification using the Kendall correlation coefficient. BRIEF DESCRIPTION OF THE DRAWINGS

[0078] 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:

[0079] Figure 1 It is a flowchart of the steps of the low voltage area topology identification method provided by the present invention;

[0080] Figure 2 It is a data schematic diagram of the first-order difference and voltage peak value provided by the present invention;

[0081] Figure 3 It is a decision diagram calculated by using the DPC method provided by the present invention;

[0082] Figure 4 It is a γ descending arrangement diagram calculated based on the DPC method provided by the present invention;

[0083] Figure 5 The present invention provides a low-voltage distribution station area phase relationship clustering diagram based on raw data combined with a DPC clustering method;

[0084] Figure 6 The present invention provides a phase relationship cluster diagram of a low-voltage distribution station area based on voltage characteristic data combined with a k-means clustering method;

[0085] Figure 7 The present invention provides a phase relationship cluster diagram of a low-voltage distribution station area based on voltage characteristic data combined with a DPC clustering method;

[0086] Figure 8This is a topological structure diagram of the low-voltage area provided by the present invention. DETAILED DESCRIPTION

[0087] 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.

[0088] Reference Figure 1 As shown, the step flow chart of the low voltage area topology identification method of the present invention includes the following specific steps:

[0089] S101: Acquire the original voltage data of each terminal to be identified in the low-voltage area at each moment in a preset time period to form an original voltage data sequence of each terminal to be identified; the terminal to be identified includes a user to be identified, a meter box to be identified, and a branch box to be identified;

[0090] S102: Calculate the difference between the original voltage data at each moment in the original voltage data sequence and the original voltage data at the previous moment as the first-order differential coefficient of the original voltage data at each moment; obtain the first-order differential coefficient sequence of each terminal to be identified based on the first-order differential coefficient of the original voltage data at each moment of each terminal to be identified;

[0091] S103: Divide the original voltage data sequence into multiple subsequences, calculate the ratio of the voltage peak value to the root mean square in each subsequence, and use it as the voltage peak coefficient of each subsequence; compose the voltage peak coefficient sequence of the original voltage data sequence from the voltage peak coefficients of each subsequence, and obtain the voltage peak coefficient sequence of each terminal to be identified;

[0092] S104: concatenating the first-order differential coefficient sequence of each terminal to be identified with the voltage peak coefficient sequence to obtain a characteristic data sequence of each terminal to be identified;

[0093] S105: Based on the characteristic data sequence of each terminal to be identified in the low-voltage area, a differential weight and a peak weight are preset, and a weighted Euclidean distance between any two terminals to be identified in the low-voltage area is calculated;

[0094] S106: Calculating the local density of each terminal to be identified based on the weighted Euclidean distance between each terminal to be identified and all other terminals to be identified;

[0095] S107: acquiring a sample distance of each terminal to be identified based on the local density of each terminal to be identified and the weighted Euclidean distance between each terminal to be identified and other terminals to be identified;

[0096] S108: Based on the local density and sample distance of each terminal to be identified, DPC clustering is performed on all terminals to be identified in the low voltage area, and terminals to be identified with the same phase are divided into the same cluster to obtain multiple clusters;

[0097] S109: For each cluster, based on the feature data sequences of all the terminals to be identified in the cluster, using the Kendall correlation coefficient calculation method, calculate the correlation between any two terminals to be identified in the cluster, and construct a correlation matrix corresponding to the cluster;

[0098] S110: Based on the correlation matrix corresponding to each cluster, obtain the affiliation between each meter box to be identified and the branch box to be identified, as well as the affiliation between each user to be identified and the meter box to be identified; based on the affiliation of all terminals to be identified in all clusters, identify and obtain the low-voltage substation topology structure.

[0099] The low-voltage substation topology identification method described in the present invention obtains the first-order differential coefficient sequence and the voltage peak coefficient sequence of each terminal to be identified based on the original voltage data sequence of each terminal to be identified to construct a feature data sequence; the first-order differential coefficient can reflect the sequence fluctuation, and the voltage peak coefficient can reflect the sequence morphology. By combining the two features together, a more comprehensive feature representation capability and a richer data representation capability are provided, thereby enhancing the performance and generalization capability of topology identification. Based on the feature data sequence of each terminal to be identified, the present invention calculates the weighted Euclidean distance to obtain the local density and sample distance of each terminal to be identified, uses the DPC clustering method to obtain the phase of each terminal to be identified, and uses the Kendall correlation coefficient to calculate the affiliation; when the original voltage data of the same substation is extremely similar, the feature data constructed by the present invention can also better reflect the differences between different voltage curves, and improve the efficiency and accuracy of low-voltage substation topology identification in phase identification using DPC clustering and affiliation identification using the Kendall correlation coefficient.

[0100] In the embodiment of the present invention, the first The original voltage data sequence of the terminal to be identified is expressed as .

[0101] Specifically, in step S102, a first-order difference coefficient sequence is obtained, including:

[0102] S102-1: Calculate the The original voltage data at the moment The original voltage data of the previous moment The difference between The first-order difference coefficient of the original voltage data at the moment , expressed as: , , Indicates the number of voltage value samples in the original voltage data sequence;

[0103] S102-2: Based on The first-order differential coefficient of the original voltage data at each moment in the original voltage data sequence of the terminal to be identified is obtained. The first-order difference coefficient sequence corresponding to the original voltage data sequence of the terminal to be identified is expressed as: .

[0104] Specifically, in step S103, obtaining a voltage peak coefficient sequence includes:

[0105] S103-1: The original voltage data sequence of the terminal to be identified Divide into raw voltage data sets, each of which contains The original voltage data, , Indicates the number of voltage value samples in the original voltage data sequence;

[0106] S103-2: Obtain the first The peak voltage of the raw voltage data set , expressed as: , Indicates The first of the raw voltage data sets data, ; Get the voltage peak value of each original voltage data group to form a voltage peak sequence , expressed as: ;

[0107] S103-3: Calculate the first The RMS value of the raw voltage data set , expressed as: ; Get the RMS of each raw voltage data group to form an RMS sequence , expressed as: ;

[0108] S103-4: Calculate the first The ratio of the voltage peak value to the RMS value of the original voltage data group is obtained. The voltage peak factor of the original voltage data set is expressed as: ;

[0109] S103-5: Based on The voltage peak sequence and RMS sequence of the terminal to be identified are obtained. The voltage peak coefficient sequence of the terminals to be identified is expressed as: .

[0110] In step S104, the first-order differential coefficient sequence of each terminal to be identified is concatenated with the voltage peak coefficient sequence to obtain a characteristic data sequence of each terminal to be identified, which is expressed as:

[0111] ;

[0112] in, Indicates The first-order difference coefficient sequence of the terminal to be identified, the sequence length is ; Indicates The voltage peak coefficient sequence of the terminals to be identified is ;No. The length of the characteristic data sequence of the terminal to be identified .

[0113] In low-voltage distribution substations, since there is usually a very high similarity between voltage curves, traditional statistical methods are often not effective in amplifying the differences between voltage curves. Therefore, it is necessary to extract the first-order differential coefficient and peak coefficient of the voltage to highlight the similarity of the voltage curves and construct new characteristic data based on the two characteristic coefficients.

[0114] Based on the above embodiment, in step S105, the weighted Euclidean distance between any two terminals to be identified in the low-voltage area is calculated, including:

[0115] Get the characteristic data sequence of each terminal to be identified in the low-voltage area, expressed as: ;

[0116] Preset differential weights With peak weight , using the weighted Euclidean distance calculation method, calculate the distance between any two terminals to be identified in the low-voltage area and The weighted Euclidean distance between , expressed as:

[0117] ;

[0118] in, Indicates the total number of elements in the characteristic data sequence of the terminal to be identified, ; ; When calculating the first-order difference coefficient sequence in the characteristic data sequence, that is, hour, , represents the difference weight, ; When calculating the voltage peak coefficient sequence in the characteristic data sequence, that is, hour, , represents the peak weight, .

[0119] In step S106, the local density of the current terminal to be identified is calculated based on the weighted Euclidean distances from the remaining terminals to be identified to the current terminal to be identified. , expressed as: ;in, Indicates the preset cutoff distance.

[0120] In step S107, the acquisition of the sample distance of each terminal to be identified includes:

[0121] S107-1: Get local density Higher than the local density of the terminal to be identified All terminals to be identified , forming a candidate set;

[0122] S107-2: If the set to be selected is empty, obtain the maximum distance value among the weighted Euclidean distances between the current terminal to be identified and the other terminals to be identified in the low-voltage area as the sample distance of the current terminal to be identified , expressed as: ;

[0123] S107-3: If the candidate set is not empty, then select the distance value with the minimum weighted Euclidean distance to the current terminal to be identified in the candidate set as the sample distance of the current terminal to be identified , expressed as: .

[0124] In step S108, DPC clustering is performed on all terminals to be identified in the low voltage area, including:

[0125] S108-1: Based on the local density and sample distance of each terminal to be identified, a clustering discrimination value is constructed, which is expressed as: ;

[0126] S108-2: Based on cluster discriminant value , draw a decision diagram and obtain the initial number of clusters With each cluster center ;

[0127] S108-3: For each terminal to be identified in the low-voltage area, divide it into the cluster where the cluster center closest to it is located, and obtain Clusters;

[0128] The cluster to which each terminal to be identified belongs is expressed as: .

[0129] Specifically, in step S109, a correlation matrix of clusters is constructed, including:

[0130] S109-1: taking a feature data sequence of any terminal to be identified in the cluster as a first observation value sequence, and a feature data sequence of another terminal to be identified as a second observation value sequence;

[0131] S109-2: Obtain the total number of pairs of the same order and the total number of pairs of the different order in the first observation value sequence and the second observation value sequence, and calculate the Kendall correlation coefficient between the two terminals to be identified as the correlation between the two terminals to be identified;

[0132] S109-3: Obtain the correlation between any two terminals to be identified in the cluster, and construct a correlation matrix containing the correlations between all terminals to be identified in the cluster, expressed as:

[0133] ;

[0134] in, Indicates the cluster The terminal to be identified and Kendall correlation coefficient between the terminals to be identified; {user to be identified, table box to be identified}, {meter box to be identified, branch box to be identified}; when When the user is to be identified, The meter box is to be identified. At this time, Represents the total number of users to be identified in the cluster. Represents the total number of boxes to be identified in the cluster; when When the meter box is to be identified, The branch box is to be identified. At this time, Represents the total number of bins to be identified in the cluster. Indicates the total number of branch boxes to be identified in the cluster.

[0135] Specifically, in step S109-2, the specific steps of calculating the correlation between any two terminals to be identified include:

[0136] Get the characteristic data sequence of any terminal to be identified As the first observation sequence, it is expressed as: , The length of the characteristic data sequence of the terminal to be identified;

[0137] Get the characteristic data sequence of another terminal to be identified , as the second observation sequence, is expressed as: ;

[0138] The total number of pairs of identical order in the first observation value sequence and the second observation value sequence is obtained, expressed as: ;

[0139] When the first observation sequence Elements With Elements , and the first Elements With Elements The size relationship is and , or and , then the element , , and Form an identical sequence pair of the first observation value sequence and the second observation value sequence;

[0140] The total number of pairs of different order in the first observation value sequence and the second observation value sequence is obtained, expressed as: ;

[0141] When the first observation sequence Elements With Elements , and the first Elements With Elements The size relationship is and , or and , then the element , , and A different sequence pair is formed between the first observation value sequence and the second observation value sequence;

[0142] in, represents a symbolic function, when When it is greater than 0, the sign function value is 1. When it is equal to 0, the sign function value is 0. When it is less than 0, the sign function value is -1;

[0143] When the first observation value sequence and the second observation value sequence do not exist or When , the Kendall correlation coefficient between the two terminals to be identified is expressed as: ;

[0144] When the first observation value sequence and the second observation value sequence exist or When , the Kendall correlation coefficient between the two terminals to be identified is expressed as: ;

[0145] in, , , ;

[0146] In the first observation sequence, all The same elements of the same element set form the same element set, and all the same element sets are constructed into the first total set. The total number of the same element sets in the first total set is ; In the second observation sequence, all The same elements of form the same element set, and all the same element sets are constructed into the second total set. The total number of the same element sets in the second total set is ; and Respectively represent the first total set and the second total set The number of elements contained in a set of identical elements.

[0147] The low-voltage substation topology identification method provided by the present invention is a full-level topology identification method for low-voltage distribution substations based on feature analysis of voltage raw data. Feature data is constructed by acquiring the first-order differential coefficient and voltage peak coefficient of the terminal to be identified. DPC clustering is performed based on the characteristic data to obtain the local density and sample distance of each terminal to be identified, and the terminals to be identified with the same phase are divided. The Kendall correlation coefficient calculation method is used to obtain the correlation between any two terminals to be identified, and a correlation matrix is ​​constructed to complete the low-voltage substation topology identification.

[0148] Based on the above embodiments, in the embodiments of the present invention, the low-voltage substation topology identification method provided by the present invention is used to select the actual user voltage data of a distribution substation on a certain day as an example for analysis. The sampling time 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 substation is divided into four levels, including: 1 distribution transformer, 4 branch boxes, 20 meter boxes and 120 single-phase users; among them, there are 40 users in phases A, B and C respectively; the 4 branch boxes are numbered F1, F2, F3, and F4 respectively; the 20 meter boxes are numbered B1, B2, ... B 20; 120 users are numbered Y1, Y2, ...Y 120 The specific identification process includes:

[0149] S201: Reference Figure 2 The figure shows a schematic diagram of the first-order difference and the voltage peak value. The first-order difference and the voltage peak value are extracted to highlight the differences between the voltage curves and form a new feature data set, including:

[0150] ① Calculate the first-order differential coefficient of voltage;

[0151] Since the positive and negative and the magnitude of the first-order difference can just describe the fluctuation trend of the voltage, by extracting this feature, we can reflect the fluctuation change of the voltage curve and refine the fluctuation characteristics of the voltage sequence change trend. , the voltage differential coefficient , whose expression is:

[0152] ;

[0153] ;

[0154] , ;

[0155] Where: is the voltage data at a certain moment, is the first-order difference coefficient between two moments, Indicates the number of voltage value samples in the original voltage data sequence.

[0156] ②Calculate the voltage peak factor;

[0157] The voltage peak coefficient is formed by dividing the peak value of the voltage data by the root mean square value. The voltage data is divided into m segments, each segment contains k voltage values, and the peak coefficient elements on each segment are calculated to form the peak coefficient. , whose expression is: ;

[0158] in, ; ; ;

[0159] Voltage peak: ;

[0160] RMS: ;

[0161] Voltage crest factor: , ;

[0162] In the formula, is the set of peak values ​​in each segment, is the set of RMS values ​​of each segment, For the The maximum value in the segment, For the The RMS value of the segment, For the Piecewise peak factor elements.

[0163] ③ Characteristic data;

[0164] Combine the extracted voltage feature data to form the feature data of a certain user , whose expression is: ;

[0165] Specifically, in this embodiment, the original voltage data collected by the smart voltage meter , calculate the first-order difference coefficient , voltage crest factor , get voltage characteristic data .

[0166] Reference Figure 3 The decision diagram obtained by using the DPC method in this embodiment is shown in FIG.

[0167] Voltage characteristic data set It contains the first-order difference that can reflect the sequence fluctuation and the peak coefficient that reflects the sequence morphology. By combining the two features, it provides a more comprehensive feature representation capability and richer data representation capability, thereby enhancing the performance and generalization ability of the model. When the original voltage data in the same substation has a very high similarity, the voltage feature data can better reflect the differences between different voltage curves and improve the recognition efficiency and accuracy in phase recognition and affiliation recognition.

[0168] S202: Based on the voltage characteristic data, DPC clustering is improved by calculating the weighted Euclidean distance to achieve phase identification, including:

[0169] ① Input the original voltage sample data ,in Represents a total of The voltage characteristic data set is obtained by the voltage characteristic analysis method: ; In the formula, express The first-order difference set of users, express The peak factor set of each user.

[0170] ②Use weighted Euclidean distance to calculate the distance between samples. , Assign weights to each , . ; In the formula, , Indicates the ID of a user. The weight sequence number when finding the difference of corresponding elements.

[0171] ③Calculate the density of sample points :

[0172] ; In the formula, Represents the cutoff distance. The distances between all data objects are arranged in ascending order. Usually, the data distances in the first 2% are selected as the cutoff distance.

[0173] ④Find distance sample points closest and with density greater than Field sample points .

[0174] Get all terminals to be identified whose local density is higher than the local density of the current terminal to be identified , forming a candidate set

[0175] If the set to be selected is empty, the maximum distance value among the weighted Euclidean distances from the current terminal to be identified to the other terminals to be identified in the low-voltage area is obtained as the sample distance of the current terminal to be identified. , expressed as: ;

[0176] If the candidate set is not empty, then in the candidate set, select the distance value with the minimum weighted Euclidean distance to the current terminal to be identified as the sample distance of the current terminal to be identified. , expressed as: ;

[0177] Where: For sample points With density higher than The minimum distance between samples if there is no density higher than The sample of The maximum distance between the samples and the rest is set to ; For density higher than The sample number.

[0178] ⑤Select cluster centers. The value of selects the cluster center, which usually has a larger density and distance , The expression is: ;

[0179] The results were sampled Decision diagram to determine the initial number of clusters K and cluster centers .

[0180] ⑥ Divide the sample categories. For the sample set to be clustered , calculate its partition clusters, and assign the samples to be clustered to the cluster closest to the cluster center: ;

[0181] In the formula, , ; Indicates The collection of class sample clusters, .

[0182] In the phase identification of low-voltage distribution station area, DPC clustering is performed on the meter box voltage with consistent phase information and the user voltage at the phase information position. The result is defined as a cluster label according to the meter box phase information. Under this cluster, the user and the meter box belong to the same phase.

[0183] Specifically, this embodiment uses the DPC clustering method to perform cluster analysis on 120 users based on the voltage feature data set. Figure 4 As shown, it is a descending order diagram of γ calculated based on the DPC method; Figure 4 It is obtained by calculating the γ value and arranging it in descending order, among which there are three larger points, respectively Figure 3 Therefore, three cluster centers can be identified, representing phases A, B, and C respectively, and the remaining 117 users are classified into phases A, B, and C respectively.

[0184] Reference Figure 5 As shown, the phase relationship clustering diagram of the low-voltage distribution station area is obtained based on the original data combined with the DPC clustering method; refer to Figure 6 As shown in the figure, the phase relationship clustering diagram of low-voltage distribution station area is obtained based on voltage characteristic data combined with k_means clustering method; refer to Figure 7 As shown in the figure, the phase relationship clustering diagram of the low-voltage distribution station area is obtained based on the voltage characteristic data combined with the DPC clustering method. Figure 5 , Figure 6 and Figure 7 It can be seen that the voltage characteristic data constructed based on the present embodiment and adopted in the embodiment of the present invention has the best clustering effect when clustered using the DCP clustering method.

[0185] S203: Calculate the Kendall correlation coefficient of the voltage characteristic data, and perform an arrangement test to identify the maximum correlation correspondence relationship to identify the affiliation relationship, thereby realizing the topology identification of the low-voltage distribution station area. The specific process includes:

[0186] ① Take the characteristic data sequence of any terminal to be identified in the cluster as the first observation value sequence, and the characteristic data sequence of another terminal to be identified as the second observation value sequence, recorded as: ; ;

[0187] The calculation method of Kendall correlation coefficient is: suppose the number of elements in both sets P and Q is L, for the elements in P and , and the corresponding elements in the permutation Q and , compare their order. When and ,or and When , these two groups of numbers are recorded as a pair of the same order; when and ,or and When , these two sets of numbers are recorded as a pair with different order. or , the two elements are neither in the same order nor in different order.

[0188] 1) Count the number of identical order pairs S and different order pairs D.

[0189] ;

[0190] ;

[0191] In the formula, is a symbolic function, when When it is greater than 0, the sign function value is 1. When it is equal to 0, the sign function value is 0. When it is less than 0, the sign function value is -1.

[0192] 2) Calculate the Kendall correlation coefficient.

[0193] a. When the collection does not exist or In case of: ;

[0194] b. When there is a or In case of: ;

[0195] in, , , ;

[0196] Combine the same elements in P and Q into small sets, where and are the number of small sets consisting of the same elements in sets P and Q respectively, and Indicates the first The number of elements contained in the small set. The value range of is between [-1, 1]. When it is 1, it means that the two random variables have consistent rank correlation; when When it is -1, it means that the two random variables have completely opposite rank correlation; when When it is 0, it means that the two random variables are independent of each other.

[0197] 3) Calculate the correlation matrix of the voltage characteristic data set.

[0198] Obtain the correlation between any two terminals to be identified in the cluster, and construct a correlation matrix containing the correlations between all terminals to be identified in the cluster, expressed as:

[0199] ;

[0200] In the formula, u and v are the user to be assigned, meter box or meter box, and branch number respectively. It represents the The user (box) and The degree of correlation between the boxes (branches), The larger the value, the greater the similarity. Normalize the elements so that their values ​​are between [0,1], and sort the rows by Since the user (table box) level is lower than the table box (branch), one user (table box) can only belong to one table box (branch), but not necessarily vice versa. Therefore, the correct affiliation can be found by arranging them by rows and seeking the maximum value.

[0201] Reference Figure 8As shown, it is a low-voltage substation topology diagram identified by this embodiment. The purpose of affiliation identification is to determine that the user belongs to a specific meter box, and the meter box belongs to a specific branch. First, the Kendall correlation coefficients between the user and the meter box, and between the meter box and the branch box are calculated using the data processed by the voltage characteristic analysis method. Since a user can only belong to one meter box, and a meter box can only belong to one branch, it is only necessary to sort the correlation coefficients between a certain user and all meter boxes, and the user belongs to the meter box with the largest correlation coefficient. The identification of the affiliation relationship between the meter box and the branch is the same. Refer to Table 1, which shows the Kendall correlation coefficients between some meter boxes and branch boxes; refer to Table 2, which shows the Kendall correlation coefficients between some users and meter boxes.

[0202] Table 1 Kendall correlation coefficients between some meter boxes and branch boxes

[0203] F1 F2 F3 F4 B1 B2 B3 B4 F1 1.00 0.50 0.55 0.56 0.57 0.58 0.47 0.40 F2 0.50 1.00 0.66 0.64 0.13 0.12 0.03 0.00 F3 0.55 0.66 1.00 0.66 0.17 0.18 0.09 0.05 F4 0.56 0.64 0.66 1.00 0.18 0.18 0.11 0.08 B1 0.57 0.13 0.17 0.18 1.00 0.88 0.81 0.74 B2 0.58 0.12 0.18 0.18 0.88 1.00 0.84 0.77 B3 0.47 0.03 0.09 0.11 0.81 0.84 1.00 0.88 B4 0.40 0.00 0.05 0.08 0.74 0.77 0.88 1.00

[0204] Table 2 Normalized Kendall correlation coefficients between some users and meter boxes

[0205] B1 B2 B3 B4 B5 B6 Y1 Y7 Y13 Y19 Y25 Y31 B1 1.00 0.85 0.75 0.62 0.67 0.14 0.94 0.85 0.72 0.60 0.66 0.14 B2 0.85 1.00 0.74 0.62 0.66 0.14 0.80 0.97 0.71 0.60 0.64 0.13 B3 0.75 0.74 1.00 0.81 0.71 0.06 0.77 0.76 0.96 0.79 0.70 0.08 B4 0.62 0.62 0.81 1.00 0.71 0.05 0.65 0.64 0.83 0.98 0.71 0.11 B5 0.67 0.66 0.71 0.71 1.00 0.15 0.70 0.68 0.73 0.71 0.98 0.21 B6 0.14 0.14 0.06 0.05 0.15 1.00 0.13 0.13 0.05 0.05 0.15 0.91 Y1 0.94 0.80 0.77 0.65 0.70 0.13 1.00 0.81 0.74 0.64 0.68 0.15 Y7 0.85 0.97 0.76 0.64 0.68 0.13 0.81 1.00 0.73 0.62 0.66 0.13 Y13 0.72 0.71 0.96 0.83 0.73 0.05 0.74 0.73 1.00 0.81 0.71 0.09 Y19 0.60 0.60 0.79 0.98 0.71 0.05 0.64 0.62 0.81 1.00 0.70 0.11 Y25 0.66 0.64 0.70 0.71 0.98 0.15 0.68 0.66 0.71 0.70 1.00 0.22 Y31 0.14 0.13 0.08 0.11 0.21 0.91 0.15 0.13 0.09 0.11 0.22 1.00

[0206] The hierarchical affiliation can be determined from the data in the table. For example, in Table 1, the correlation coefficient between B1 and F1 is 0.57, which is the largest among the correlation coefficients with all branch boxes. It can be considered that meter box B1 belongs to branch box F1. In Table 2, the correlation coefficient between Y1 and B1 is 0.94, which is the largest among the correlation coefficients with all meter boxes. It can be considered that user Y1 belongs to meter box B1. Similarly, all user-(meter box)-meter box (branch box) affiliations can be obtained, and finally the complete topological structure of the low-voltage distribution station area can be obtained.

[0207] In order to verify the effectiveness of the voltage feature analysis method and the Kendall correlation coefficient in the identification of affiliation, the Pearson correlation coefficient was selected as a comparison, and the original data combined with the Pearson correlation coefficient and the voltage feature data combined with the Pearson correlation coefficient were set as a control group. Referring to Table 3, Table 4 and Table 5, they are the correlation calculated based on the original data combined with the Pearson correlation coefficient, the correlation calculated based on the voltage feature data combined with the Pearson correlation coefficient, and the correlation calculated based on the voltage feature data combined with the Kendall correlation coefficient provided in this embodiment.

[0208] Table 3 Correlations calculated based on original data combined with Pearson correlation coefficient

[0209]

[0210] Table 4 Correlations calculated based on voltage characteristic data combined with Pearson correlation coefficient

[0211]

[0212] Table 5 Correlations calculated based on voltage characteristic data combined with Kendall correlation coefficient

[0213]

[0214] From the overall numerical value, after the voltage characteristic analysis method is used to process the data, the maximum value can be more easily identified from the correlation coefficient table, while the Pearson correlation coefficient calculated using the original voltage data shows a phenomenon of concentration above 0.9, which is not conducive to the identification of the affiliation; the values ​​in Table 3 can more clearly reflect the position of the maximum value as a whole compared to Table 4, but some values ​​are still similar and difficult to identify. As can be seen from Table 3, the correlation coefficients of users Y1 and Y2 with meter boxes B1 and B2 are both 0.96, indicating that it cannot be determined whether users Y1 and user Y2 belong to meter box B1 or meter box B2; the correlation coefficients of user Y3 with meter boxes B2 and B3 are both 0.96, indicating that it cannot be determined whether user Y3 belongs to meter box B2 or meter box B3, and the Pearson correlation coefficient method is invalid at this time.

[0215] The Pearson correlation coefficient after processing the data using the voltage characteristic analysis method in this embodiment has been significantly improved in the identification of the affiliation relationship compared to the Pearson correlation coefficient calculated only using the original data. The value under the non-affiliation relationship is no longer concentrated between 0.9-1, which is conducive to distinguishing the correct corresponding relationship in the identification of the affiliation relationship. This shows that the voltage characteristic analysis method has obvious advantages in processing voltage data and can highlight the same / different features between voltage data. However, there are still cases where the affiliation relationship cannot be identified. For example, if the correlation coefficients of users Y1 and Y2 with meter boxes B1 and B2 are both 0.99, it cannot be determined whether users Y1 and Y2 belong to meter box B1 or meter box B2.

[0216] Compared with Table 4, the values ​​on the diagonal of Table 5 are higher than the values ​​of other non-corresponding affiliations, and there is no situation that cannot be identified. While retaining the correct identification value, the remaining values ​​are lower, which can better highlight the correct affiliation and is more conducive to the identification of the affiliation. Therefore, the Kendall correlation coefficient used in this embodiment is better in identifying the affiliation.

[0217] The embodiment of the present invention also provides a low voltage area topology identification device; the specific device includes:

[0218] A data acquisition module is used to acquire the original voltage data of each terminal to be identified in the low-voltage area at each moment within a preset time period to form an original voltage data sequence of each terminal to be identified; the terminal to be identified includes a user to be identified, a meter box to be identified, and a branch box to be identified;

[0219] The differential coefficient calculation module is used to calculate the difference between the original voltage data at each moment in the original voltage data sequence and the original voltage data at the previous moment as the first-order differential coefficient of the original voltage data at each moment; based on the first-order differential coefficient of the original voltage data of each terminal to be identified at each moment, obtain the first-order differential coefficient sequence of each terminal to be identified;

[0220] The peak coefficient calculation module is used to divide the original voltage data sequence into multiple subsequences, calculate the ratio of the voltage peak value to the root mean square in each subsequence, and use it as the voltage peak coefficient of each subsequence; the voltage peak coefficient sequence of the original voltage data sequence is composed of the voltage peak coefficients of each subsequence, and the voltage peak coefficient sequence of each terminal to be identified is obtained;

[0221] A feature data construction module is used to concatenate the first-order differential coefficient sequence of each terminal to be identified with the voltage peak coefficient sequence to obtain a feature data sequence of each terminal to be identified;

[0222] A clustering module is used to calculate the weighted Euclidean distance between any two terminals to be identified in the low-voltage area based on the feature data sequence of each terminal to be identified in the low-voltage area, preset differential weights and peak weights; calculate the local density of each terminal to be identified based on the weighted Euclidean distance between each terminal to be identified and all other terminals to be identified; obtain the sample distance of each terminal to be identified based on the local density of each terminal to be identified and the weighted Euclidean distance between each terminal to be identified and other terminals to be identified; perform DPC clustering on all terminals to be identified in the low-voltage area based on the local density and sample distance of each terminal to be identified, divide the terminals to be identified with the same phase into the same cluster cluster, and obtain multiple cluster clusters;

[0223] A correlation matrix construction module is used to calculate the correlation between any two terminals to be identified in each cluster based on the characteristic data sequences of all terminals to be identified in the cluster using the Kendall correlation coefficient calculation method, and to construct a correlation matrix corresponding to the cluster;

[0224] The topology identification module is used to obtain the affiliation between each meter box to be identified and the branch box to be identified, as well as the affiliation between each user to be identified and the meter box to be identified based on the correlation matrix corresponding to each cluster; based on the affiliation of all terminals to be identified in all clusters, identify and obtain the low-voltage substation topology structure.

[0225] The low-voltage substation topology identification device of this embodiment is used to implement the aforementioned low-voltage substation topology identification method. Therefore, the specific implementation method of the low-voltage substation topology identification device can be seen in the implementation method part of the low-voltage substation topology identification method in the previous text. For example, the data acquisition module, the differential coefficient calculation module, the peak coefficient calculation module, and the characteristic data construction module are respectively used to implement steps S101, S102, S103 and S104 in the above-mentioned low-voltage substation topology identification method; the clustering module is used to implement steps S105, S106, S107 and S108 in the above-mentioned low-voltage substation topology identification method; the correlation matrix construction module and the topology identification module are respectively used to implement steps S109 and S110 in the above-mentioned low-voltage substation topology identification method; therefore, its specific implementation method can refer to the description of the corresponding embodiments of each part, which will not be repeated here.

[0226] The low-voltage substation topology identification method described in the present invention obtains the first-order differential coefficient sequence and the voltage peak coefficient sequence of each terminal to be identified based on the original voltage data sequence of each terminal to be identified to construct a feature data sequence; the first-order differential coefficient can reflect the sequence fluctuation, and the voltage peak coefficient can reflect the sequence morphology. By combining the two features together, a more comprehensive feature representation capability and a richer data representation capability are provided, thereby enhancing the performance and generalization capability of topology identification. Based on the feature data sequence of each terminal to be identified, the present invention calculates the weighted Euclidean distance to obtain the local density and sample distance of each terminal to be identified, uses the DPC clustering method to obtain the phase of each terminal to be identified, and uses the Kendall correlation coefficient to calculate the affiliation; when the original voltage data of the same substation is extremely similar, the feature data constructed by the present invention can also better reflect the differences between different voltage curves, and improve the efficiency and accuracy of low-voltage substation topology identification in phase identification using DPC clustering and affiliation identification using the Kendall correlation coefficient.

[0227] 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 combining software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0228] 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.

[0229] 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.

[0230] 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 in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0231] 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 low voltage area topology, characterized in that: include: Obtaining the original voltage data of each terminal to be identified in the low-voltage area at each moment in a preset time period to form an original voltage data sequence of each terminal to be identified; The terminals to be identified include users to be identified, meter boxes to be identified, and branch boxes to be identified; Calculate the difference between the original voltage data at each moment in the original voltage data sequence and the original voltage data at the previous moment as the first-order difference coefficient of the original voltage data at each moment; Based on the first-order differential coefficients of the original voltage data of each terminal to be identified at each moment, a first-order differential coefficient sequence of each terminal to be identified is obtained, including: obtaining the first The original voltage data sequence of the terminal to be identified is expressed as ; Calculate the The original voltage data at the moment The original voltage data of the previous moment The difference between The first-order difference coefficient of the original voltage data at the moment , expressed as: , , Indicates the number of voltage value samples in the original voltage data sequence; based on the The first-order differential coefficient of the original voltage data at each moment in the original voltage data sequence of the terminal to be identified is obtained. The first-order difference coefficient sequence corresponding to the original voltage data sequence of the terminal to be identified is expressed as: ; The original voltage data sequence is divided into multiple subsequences, and the ratio of the voltage peak value to the root mean square in each subsequence is calculated as the voltage peak coefficient of each subsequence; the voltage peak coefficient sequence of the original voltage data sequence is formed by the voltage peak coefficients of each subsequence, and the voltage peak coefficient sequence of each terminal to be identified is obtained, including: The original voltage data sequence of the terminal to be identified Divide into raw voltage data sets, each of which contains The original voltage data, , Indicates the number of voltage value samples in the original voltage data sequence; obtains the number of voltage value samples in the original voltage data sequence. The peak voltage of the raw voltage data set , expressed as: , Indicates The first of the raw voltage data sets data, ; Get the voltage peak value of each original voltage data group to form a voltage peak sequence , expressed as: ; Calculate the first The RMS value of the raw voltage data set , expressed as: ; Get the RMS of each raw voltage data group to form an RMS sequence , expressed as: ; Calculate the first The ratio of the voltage peak value to the RMS value of the original voltage data group is obtained. The voltage peak factor of the original voltage data set is expressed as: Based on The voltage peak sequence and RMS sequence of the terminal to be identified are obtained. The voltage peak coefficient sequence of the terminals to be identified is expressed as: ; The first-order differential coefficient sequence of each terminal to be identified is concatenated with the voltage peak coefficient sequence to obtain the characteristic data sequence of each terminal to be identified, which is expressed as: ;in, Indicates The first-order difference coefficient sequence of the terminal to be identified, the sequence length is ; Indicates The voltage peak coefficient sequence of the terminals to be identified is ;No. The length of the characteristic data sequence of the terminal to be identified ; Based on the characteristic data sequence of each terminal to be identified in the low-voltage area, the differential weight and the peak weight are preset, and the weighted Euclidean distance between any two terminals to be identified in the low-voltage area is calculated, including: obtaining the characteristic data sequence of each terminal to be identified in the low-voltage area, expressed as: ; Preset differential weights With peak weight , using the weighted Euclidean distance calculation method, calculate the distance between any two terminals to be identified in the low-voltage area and The weighted Euclidean distance between , expressed as: ;in, Indicates the total number of elements in the characteristic data sequence of the terminal to be identified, ; ; When calculating the first-order difference coefficient sequence in the characteristic data sequence, that is, hour, , represents the difference weight, ; When calculating the voltage peak coefficient sequence in the characteristic data sequence, that is, hour, , represents the peak weight, ; Based on the weighted Euclidean distance between each terminal to be identified and all other terminals to be identified, the local density of each terminal to be identified is calculated, which is expressed as: ;in, Indicates The local density of terminals to be identified, Indicates the preset cutoff distance; Based on the local density of each terminal to be identified and the weighted Euclidean distance between each terminal to be identified and other terminals to be identified, the sample distance of each terminal to be identified is obtained, including: obtaining the local density Higher than the local density of the terminal to be identified All terminals to be identified , forming a candidate set; if the candidate set is empty, the maximum distance value among the weighted Euclidean distances from the current terminal to be identified to the other terminals to be identified in the low-voltage area is obtained as the sample distance of the current terminal to be identified , expressed as: ; If the candidate set is not empty, then in the candidate set, select the distance value with the minimum weighted Euclidean distance to the current terminal to be identified as the sample distance of the current terminal to be identified , expressed as: ; Based on the local density and sample distance of each terminal to be identified, DPC clustering is performed on all terminals to be identified in the low-voltage area, and the terminals to be identified with the same phase are divided into the same cluster cluster to obtain multiple cluster clusters, including: Based on the local density and sample distance of each terminal to be identified, a cluster discrimination value is constructed, which is expressed as: ; Based on cluster discriminant value , draw a decision diagram and obtain the initial number of clusters With each cluster center ; For each terminal to be identified in the low-voltage area, divide it into the cluster with the closest cluster center, and obtain clusters; the cluster to which each terminal to be identified belongs is expressed as: ; For each cluster, based on the characteristic data sequences of all terminals to be identified in the cluster, the Kendall correlation coefficient calculation method is used to calculate the correlation between any two terminals to be identified in the cluster, and the correlation matrix corresponding to the cluster is constructed, including: Taking a characteristic data sequence of any terminal to be identified in the cluster as a first observation value sequence, and a characteristic data sequence of another terminal to be identified as a second observation value sequence; Obtaining the total number of pairs of same order and the total number of pairs of different order in the first observation value sequence and the second observation value sequence, and calculating the Kendall correlation coefficient between the two terminals to be identified as the correlation between the two terminals to be identified, including: Get the characteristic data sequence of any terminal to be identified As the first observation sequence, it is expressed as: , The length of the characteristic data sequence of the terminal to be identified; Get the characteristic data sequence of another terminal to be identified , as the second observation sequence, is expressed as: ; The total number of pairs of identical order in the first observation value sequence and the second observation value sequence is obtained, expressed as: ; When the first observation sequence Elements With Elements , and the first Elements With Elements The size relationship is and , or and , then the element , , and Form an identical sequence pair of the first observation value sequence and the second observation value sequence; The total number of pairs of different order in the first observation value sequence and the second observation value sequence is obtained, expressed as: ; When the first observation sequence Elements With Elements , and the first Elements With Elements The size relationship is and , or and , then the element , , and A different sequence pair is formed between the first observation value sequence and the second observation value sequence; in, represents a symbolic function, when When it is greater than 0, the sign function value is 1. When it is equal to 0, the sign function value is 0. When it is less than 0, the sign function value is -1; When the first observation value sequence and the second observation value sequence do not exist or When , the Kendall correlation coefficient between the two terminals to be identified is expressed as: ; When the first observation value sequence and the second observation value sequence exist or When , the Kendall correlation coefficient between the two terminals to be identified is expressed as: ; in, , , ; In the first observation sequence, all The same elements of the same element set form the same element set, and all the same element sets are constructed into the first total set. The total number of the same element sets in the first total set is ; In the second observation sequence, all The same elements of form the same element set, and all the same element sets are constructed into the second total set. The total number of the same element sets in the second total set is ; and Respectively represent the first total set and the second total set The number of elements contained in a set of identical elements; Obtain the correlation between any two terminals to be identified in the cluster, and construct a correlation matrix containing the correlations between all terminals to be identified in the cluster, expressed as: ； in, Indicates the cluster The terminal to be identified and Kendall correlation coefficient between the terminals to be identified; {user to be identified, table box to be identified}, {meter box to be identified, branch box to be identified}; when When the user is to be identified, The meter box is to be identified. At this time, Represents the total number of users to be identified in the cluster. Represents the total number of boxes to be identified in the cluster; when When the meter box is to be identified, The branch box is to be identified. At this time, Represents the total number of bins to be identified in the cluster. Indicates the total number of branch boxes to be identified in the cluster; Based on the correlation matrix corresponding to each cluster, the affiliation between each meter box to be identified and the branch box to be identified, as well as the affiliation between each user to be identified and the meter box to be identified are obtained; based on the affiliation of all terminals to be identified in all clusters, the low-voltage substation topology structure is identified and obtained.

2. A low voltage area topology identification device based on the low voltage area topology identification method according to claim 1, characterized in that: include: A data acquisition module is used to acquire the original voltage data of each terminal to be identified in the low-voltage area at each moment within a preset time period to form an original voltage data sequence of each terminal to be identified; the terminal to be identified includes a user to be identified, a meter box to be identified, and a branch box to be identified; The differential coefficient calculation module is used to calculate the difference between the original voltage data at each moment in the original voltage data sequence and the original voltage data at the previous moment as the first-order differential coefficient of the original voltage data at each moment; Based on the first-order differential coefficients of the original voltage data of each terminal to be identified at each moment, a first-order differential coefficient sequence of each terminal to be identified is obtained; The peak coefficient calculation module is used to divide the original voltage data sequence into multiple subsequences, calculate the ratio of the voltage peak value to the root mean square in each subsequence, and use it as the voltage peak coefficient of each subsequence; the voltage peak coefficient sequence of the original voltage data sequence is composed of the voltage peak coefficients of each subsequence, and the voltage peak coefficient sequence of each terminal to be identified is obtained; A feature data construction module is used to concatenate the first-order differential coefficient sequence of each terminal to be identified with the voltage peak coefficient sequence to obtain a feature data sequence of each terminal to be identified; A clustering module is used to calculate the weighted Euclidean distance between any two terminals to be identified in the low-voltage area based on the feature data sequence of each terminal to be identified in the low-voltage area, preset differential weights and peak weights; calculate the local density of each terminal to be identified based on the weighted Euclidean distance between each terminal to be identified and all other terminals to be identified; obtain the sample distance of each terminal to be identified based on the local density of each terminal to be identified and the weighted Euclidean distance between each terminal to be identified and other terminals to be identified; perform DPC clustering on all terminals to be identified in the low-voltage area based on the local density and sample distance of each terminal to be identified, divide the terminals to be identified with the same phase into the same cluster cluster, and obtain multiple cluster clusters; A correlation matrix construction module is used to calculate the correlation between any two terminals to be identified in each cluster based on the characteristic data sequences of all terminals to be identified in the cluster using the Kendall correlation coefficient calculation method, and to construct a correlation matrix corresponding to the cluster; The topology identification module is used to obtain the affiliation between each meter box to be identified and the branch box to be identified, as well as the affiliation between each user to be identified and the meter box to be identified based on the correlation matrix corresponding to each cluster; based on the affiliation of all terminals to be identified in all clusters, identify and obtain the low-voltage substation topology structure.

Images