Low-voltage distribution network topology identification method based on interval type-2 fuzzy clustering algorithm

Abstract

The present invention belongs to the technical field of power system automation. Disclosed is a low-voltage distribution network topology identification method based on an interval type-2 fuzzy clustering algorithm. The method comprises: on the basis of longitude and latitude geographic information of end-users and distribution transformers in a power GIS, and combined with radiation shapes of distribution transformer areas, preliminarily determining topological connection relationships between the end-users and the distribution transformer areas; for users with possible power resupply or load transfer and uncertain affiliation, performing anomaly factor detection of voltage data, so as to filter out user data not belonging to a transformer area to be analyzed; and introducing a correlation coefficient, forming an improved interval type-2 fuzzy clustering algorithm, and determining "transformer-user" topology identification. The present invention can realize accurate identification of the phase affiliation of users in low-voltage transformer areas, thereby effectively improving the accuracy of user identification; and the present invention can maintain relatively high stability in different data environments.

Description

A low voltage distribution network topology identification method based on interval type 2 fuzzy clustering algorithm Technical Field

[0001] The present invention belongs to the technical field of power system automation, and in particular relates to a low-voltage distribution network topology identification method based on an interval type-2 fuzzy clustering algorithm. Background Art

[0002] Low-voltage distribution network topology identification is fundamental to the development of smart substations, and the correctness of this topology is a prerequisite for other advanced applications. Automatic identification of the physical topology of "substation-branch-meter box-user" creates a "single grid map" and enables digital management of the entire multi-faceted process, including planning, construction, operation and maintenance, and service. This can effectively reduce the burden on grassroots teams and improve the overall efficiency of power supply companies.

[0003] The Technical Guidelines for Planning and Design of Distribution Networks stipulate in the rule design: For the power supply radius of 10kV lines, in principle, the power supply radius of Class A and Class B power supply areas should not exceed 3km, that of Class C should not exceed 5km, and that of Class D should not exceed 15km; for the power supply radius of low-voltage lines, in principle, the power supply radius of Class A power supply areas should not exceed 150m, that of Class B should not exceed 250m, that of Class C should not exceed 400m, and that of Class D should not exceed 500m.

[0004] As the terminal link connecting users, the intelligence level of the low-voltage distribution network directly affects the efficiency and difficulty of operations and maintenance personnel, and also affects users' electricity experience and satisfaction. Power companies use Geographic Information Systems (GIS) to record distribution network assets and their topological connections within the system. For a long time, due to the simplistic management of low-voltage distribution networks and the complex and frequently changing low-voltage lines, substation grid topology information has been missing, incomplete, and out of date. These errors have a negative impact on effective asset management, maintenance, fault response, system operation, and the personal safety of on-site personnel.

[0005] To address these issues, power companies typically establish strict processes to ensure that relevant personnel promptly update system records after operations. They also regularly organize dedicated human resources to investigate errors in an effort to reduce these errors. However, manual identification methods are expensive and rarely achieve the desired results. The main reason is that the distribution system is too large, and new operations are constantly occurring daily. Failure to update information in a timely manner results in user information that is inconsistent with actual conditions. User information in distribution network substations is confusing, lost, or inaccurate, impacting the operation and management of the distribution network. Existing low-voltage distribution network topology technologies address abnormal data through interpolation, feature extraction, and other methods, but these can lead to overcompensation and inaccurate data. Summary of the Invention

[0006] In order to solve the above technical problems, the present invention provides a low-voltage distribution network topology identification method based on the interval type 2 fuzzy clustering algorithm, which pre-processes the user's voltage meter data, and then classifies the samples to be identified by improving the interval type 2 fuzzy clustering algorithm to accurately realize the "substation-user" structure topology.

[0007] The present invention provides a low-voltage distribution network topology identification method based on an interval type-2 fuzzy clustering algorithm, comprising the following steps:

[0008] Step 1: Based on the latitude and longitude coordinates of the distribution transformer area and the end user in the GIS system and the radiation range limit of the area, preliminarily determine the topological connection between the area transformer and the end user;

[0009] Step 2: Collect transformer low-voltage side voltage and user voltage data, pre-process the collected user voltage data based on the local anomaly factor algorithm, and obtain user voltage data of the substation to be identified;

[0010] Step 3: Determine the initial data of the improved interval type 2 fuzzy clustering algorithm, cluster the user voltage data of the area to be identified based on the improved interval type 2 fuzzy clustering algorithm, and finally identify the topological connection relationship of the users in the area.

[0011] Furthermore, in step 1, the length of the low-voltage distribution line should meet the terminal voltage quality requirements. According to the distribution network rules and guidelines, the power supply radius of Class A, B, C, and D areas is defined, where Class A represents urban areas, Class B represents suburban areas, Class C represents towns, and Class D represents rural areas.

[0012] Verify based on the radiation shape of the distribution transformer area and the latitude and longitude information of the end user:

[0013] With the distribution transformer as the center and r as the radius, the area radiates. The electricity users covered by the radiation circle are defined as the potential users of the area.

[0014] If a user has one and only one subordinate distribution transformer area, the user is directly identified as belonging to the substation area where the radiation circle is located;

[0015] If there are several overlapping areas in the radiation circles of multiple distribution transformer areas, users in the area will receive power across the transformer areas.

[0016] Furthermore, in step 2, the transformer low-voltage side voltage and user voltage data are collected. Assuming that the voltage data set is U, where p is any sample in U, the collected user voltage data is preprocessed based on the local anomaly factor algorithm to obtain the user voltage data of the substation to be identified, specifically:

[0017] Step 2-1, the expression of k-dist(p) of sample p is as follows:

[0018] k-dist(p)=d(p,o) (1),

[0019] Where k-dist(p) represents the k-th distance from p, that is, the distance between p and its k-th neighbor; d(p,o) is the distance between sample p and sample o;

[0020] Formula (1) satisfies the following conditions: In the dataset U, in addition to sample p itself, there are at least k samples p′∈U, satisfying d(p,p′)≤k-dist(p);

[0021] Step 2-2: The kth neighborhood N of sample p k (p) is expressed as follows:

[0022] N k (p)={o∈U\{p}|d(p,o)≤k-dist(p)} (2),

[0023] Where N k (p) represents all samples in the sample set except sample p whose spatial distance to sample p is less than or equal to k-dist(p);

[0024] Step 2-3: The reachable distance r between sample p and other samples k The expression of (p, o) is as follows:

[0025] r k (p,o)=max{k-dist(p),d(p,o)} (3),

[0026] Where r k (p,o) represents the reachable distance between sample p and sample o. The k points closest to point p have the same reachable distance.

[0027] Step 2-4: The expression of the local reachability density of sample p is as follows:

[0028]

[0029] Where, lrd k (p) is the sample p and the sample N within the kth distance k The reciprocal of the average reachable distance of (p); |N k (p)| is the number of samples p;

[0030] Step 2-5: Local Outlier Factor LOF of Sample pk (p) is expressed as follows:

[0031]

[0032] If the local outlier factor is greater than 2, the user is determined not to belong to the area to be analyzed and the data is removed.

[0033] Furthermore, in step 3, the correlation coefficient C is introduced into the interval type 2 fuzzy clustering algorithm. ij To describe the consistency of sample data fluctuations, an improved interval type 2 fuzzy clustering algorithm is constructed; the correlation coefficient C ij The expression is as follows:

[0034]

[0035] Where, is a unit vector, m is the data dimension of the sample, m = sampling frequency * sampling duration; sample i and sample j represent different sample objects respectively, X i and X j are the data sets of sample i and sample j in the time dimension m; X id and X jd are the data of sample i and sample j in the dth dimension respectively; X′ represents the transposed matrix of matrix X.

[0036] Furthermore, in step 3, the necessary initial data for the improved interval type 2 fuzzy clustering algorithm are: the number of transformers in the substation area, that is, the number of clusters k; the initial centroid, the maximum number of iterations I max , distance threshold, fuzzy coefficients m1, m2.

[0037] Furthermore, in step 3, the voltage data of users in the area to be identified are clustered based on the improved interval type 2 fuzzy clustering algorithm, and the topological connection relationship of users in the area is finally identified, specifically:

[0038] Step 3-1: Assume that there are n users in the overlapping area, and the user's voltage vector is represented by x i ,i=1,2,…n;The voltage data of the user to be identified X={x i |i=1,2,…,n}; the number of clusters is k, and the voltage mean representative points of each transformer area are v j ,j=1,2,…k, i.e. the initial center of mass;

[0039] Step 3-2, calculate each sample data x i and cluster center v j The correlation coefficient C ij ;

[0040] Step 3-3, calculate the fuzzy membership function of each sample data to each cluster, is the upper membership function value, μ ij is the membership function value; where z represents the zth cluster, and its value range is 1 to k:

[0041]

[0042] Step 3-4, the membership of each sample data is an interval By μ ij and The calculated cluster center will also be an interval, and the lower boundary v of the cluster center interval is L and upper boundary v R The calculation method is as follows:

[0043]

[0044] Update the cluster center point v according to formula (8) and formula (9) j ;

[0045]

[0046] Steps 3-5: Determine whether the number of iterations has reached the maximum value, or whether the new centroid is the same as the original centroid; if so, end the algorithm to complete the classification; if not, repeat the above steps until it is satisfied; finally, assign each terminal user to be assigned to the most matching substation transformer to achieve the "transformer-user" topology identification of users in the overlapping area.

[0047] The beneficial effects described in the present invention are as follows: the method described in the present invention preliminarily determines the "transformer-user" relationship based on the latitude and longitude information of the distribution transformer substation and the terminal user in the geographic information system; for users in the overlapping area of ​​the substation, the user data of the non-analyzed substation is eliminated through the local anomaly factor algorithm, and the correct user set connected to the transformer of the analysis substation is obtained, and based on this, the substation users are clustered and analyzed to achieve accurate identification of the different users in the substation; in addition, the classic clustering algorithm uses Euclidean distance as the standard for measuring sample similarity, however, Euclidean distance can only represent the absolute distance between samples and cannot reflect the changing trend of the data; since the voltage data fluctuation patterns of users in the same substation are relatively similar, the present invention introduces a correlation coefficient to describe the consistency of the sample data fluctuation, so as to replace the Euclidean distance for cluster analysis, complete topology identification, accurately and automatically identify and correct the substation power grid topology information, and realize refined management of the distribution network. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] FIG1 is a flow chart of the method of the present invention;

[0049] Figure 2 is a schematic diagram of the spatial structure distribution of low-voltage stations;

[0050] Figure 3 is a flow chart of the algorithm for local anomaly factors of user pressure data;

[0051] FIG4 is a flow chart of the improved interval type 2 fuzzy clustering algorithm;

[0052] Figure 5 shows the household change relationship clustering results obtained using the improved interval type 2 fuzzy clustering algorithm;

[0053] FIG6 is a voltage curve diagram of selected users in the transformer 1 of the substation;

[0054] FIG7 is a voltage curve diagram of selected users in the transformer 2 in the substation. DETAILED DESCRIPTION

[0055] In order to make the contents of the present invention more clearly understood, the present invention is further described in detail below based on specific embodiments in conjunction with the accompanying drawings.

[0056] As shown in FIG1 , the low-voltage distribution network topology identification method based on the interval type 2 fuzzy clustering algorithm of the present invention includes the following steps:

[0057] Step 1: Based on the longitude and latitude coordinates of the distribution transformer area and the end user in the GIS system, and according to the radiation range limit of the area, preliminarily determine the "transformer-user" topological connection between the area transformer and the end user;

[0058] Step 2: Collect transformer low-voltage side voltage and user voltage data, pre-process the collected user voltage data based on the local anomaly factor algorithm, and obtain user voltage data of the substation to be identified;

[0059] Step 3: Determine the initial data of the improved interval type 2 fuzzy clustering algorithm, cluster the user voltage data of the area to be identified based on the improved interval type 2 fuzzy clustering algorithm, and finally identify the topological connection relationship of the users in the area.

[0060] In step 1, according to the guidelines of distribution network rules, the 220 / 380V line should have a clear power supply range. The power supply radius of Class A area should not exceed 150 meters, Class B should not exceed 250 meters, Class C should not exceed 400 meters, and Class D should not exceed 500 meters.

[0061] The verification is performed based on the radial shape of the distribution transformer in the substation and the longitude and latitude information of the end user. The specific shape of the substation is related to the regional topography, terrain, climate, construction difficulty, building and road layout, etc., so the radial shape may be fan-shaped or long. As shown in Figure 2, with the distribution transformer as the center and the radius r as the radius, the regional radiation circle is formed. The electricity users covered by the radiation circle are defined as potential users of the substation. If a user has one and only one subordinate distribution transformer area, the user can be directly identified as belonging to the substation where the radiation circle is located. If the radiation circles of multiple substation distribution transformers have several overlapping areas, users in this area may be connected to electricity across substations, which is the focus of topology verification.

[0062] In step 2, based on the collected user meter voltage data, the local anomaly factor is calculated. The size of the local outlier factor is used as the basis for evaluating the degree of anomaly of the sample relative to the local neighborhood. The local anomaly factor algorithm is used to detect anomalies and eliminate user data that does not belong to the substation to be analyzed. Assume that the voltage data set is U, where p is any sample in U. As shown in Figure 3, the algorithm flow is as follows:

[0063] Step 2-1, the expression of k-dist(p) of sample p is as follows:

[0064] k-dist(p)=d(p,o) (1),

[0065] Where k-dist(p) represents the k-th distance from p, that is, the distance between p and its k-th neighbor; d(p,o) is the distance between sample p and sample o;

[0066] Step 2-2: The kth neighborhood N of sample p k (p) is expressed as follows:

[0067] N k (p)={o∈U\{p}|d(p,o)≤k-dist(p)} (2),

[0068] Where N k (p) represents all samples in the sample set except sample p whose spatial distance to sample p is less than or equal to k-dist(p);

[0069] Step 2-3: The reachable distance r between sample p and other samples k The expression of (p, o) is as follows:

[0070] r k (p, o) = max {k-dist (p), d (p, o)} (3),

[0071] Where rk (p, o) represents the reachable distance between sample p and sample o. The k points closest to point p have the same reachable distance.

[0072] Step 2-4: The expression of the local reachability density of sample p is as follows:

[0073]

[0074] Where, lrd k (p) is the sample p and the sample N within the kth distance k The reciprocal of the average reachable distance of (p), the higher its value, the greater the possibility of belonging to the same category; |N k (p)| is the number of samples p;

[0075] Step 2-5: The local outlier factor expression of sample p is as follows:

[0076]

[0077] Where LOF k (p) is the average ratio of the local reachability density of sample p's kth neighboring samples to its own local reachability density. The higher the outlier degree of sample p, the lower the local reachability density. The lower the outlier degree of sample p's kth neighboring samples, the higher their local reachability density, and the larger the local outlier factor of sample p. For a sample deeply embedded in a consistent cluster, the local outlier factor is close to 1, so regardless of cluster density or sparseness, objects within the cluster will never be marked as outliers.

[0078] When the distribution network operates normally and there are no outliers, the LOF values ​​of most samples are approximately equal to 1; the greater the distance between sample p and other samples, the greater the reachable distance of sample p, and the smaller the local reachable density of sample p; when LOF>1, the larger the LOF, the greater the possibility that the sample is abnormal data. Therefore, to ensure the reliability of data preprocessing, the criterion for identifying abnormal data samples is set to LOF value greater than the set value. Considering the actual operating status of the distribution network and the actual needs of user identification in the substation, based on the principle of the LOF algorithm, when the set value is set to 2, the identification of abnormal data can be better achieved. Therefore, if the local outlier factor is greater than 2, it is determined that the user does not belong to the substation to be analyzed and the data is eliminated.

[0079] In step 3, the user voltage meter data collection has a high-dimensional attribute with time as the scale. The Euclidean distance in the traditional interval type 2 fuzzy clustering algorithm cannot reflect the change trend of the data. Therefore, the correlation coefficient C is introduced. ij , used to describe the consistency of sample data fluctuations, the correlation coefficient C ij The expression is as follows:

[0080]

[0081] Where, is a unit vector, m is the data dimension of the sample, m = sampling frequency * sampling duration; sample i and sample j represent different sample objects respectively, X i and X j are the data sets of sample i and sample j in the time dimension m; X id and X jd are the data of sample i and sample j in the dth dimension respectively; X′ represents the transposed matrix of matrix X.

[0082] Determine the necessary initial data for the improved interval type 2 fuzzy clustering algorithm: the number of transformers in the substation area, that is, the number of clusters k; the initial centroid, the maximum number of iterations I max , distance threshold, fuzzy coefficients m1, m2, etc.

[0083] When end users switch loads or power supplies, their voltage waveforms will change. Since users in overlapping areas may belong to distribution transformers in different substations, an improved interval-type 2 fuzzy clustering algorithm is used to verify the topological affiliation of users in overlapping areas based on the voltage fluctuation characteristics of the distribution transformers in the substations, ultimately identifying the topological connectivity of users in the substations.

[0084] The sampled voltage data for each user and distribution transformer at each time point exists according to a specific topology, so the sampled voltage values ​​at each node contain hidden topological correlations. The voltage time series characterized by the sampled voltage constitutes the fluctuation characteristic sequence of that node. By using the voltage of each distribution transformer in each substation as the initial centroid, the voltage time series formed by multiple sampling times as the cluster characteristics, and the end users to be analyzed as cluster objects, the correlation coefficient between each object and each cluster center is calculated, and each object is assigned to the cluster center corresponding to the minimum correlation coefficient. The centroids are continuously updated iteratively to ultimately obtain a stable classification set.

[0085] Assume that there are n users in the overlapping area, and the user's voltage vector is represented by x i (i=1,2,…n); the voltage mean representative points of each transformer area are v j (j=1,2,…k), these points are cluster centers, and the initial point of the cluster center is defined as the characteristic voltage vector of the distribution transformer in the substation.

[0086] As shown in Figure 4, the specific process of the algorithm is as follows:

[0087] (1) Input: voltage data of the user to be identified X = {x i |i=1,2,...,n}; number of clusters k; initial centroid v j(j=1,2,...,k);

[0088] Output: k clusters;

[0089] (2) Calculate the correlation coefficient C between each sample data and the cluster center using formula (1) ij ;

[0090] (3) In order to deal with more uncertainties, the interval type 2 fuzzy clustering algorithm uses two fuzzy parameters. The fuzzy membership function of each sample data belonging to each cluster can be calculated by formula (7). is the upper membership function value, μ ij is the membership function value; where z represents the z-th cluster, and its value range is 1 to k:

[0091]

[0092] (4) The membership degree of each sample data is an interval By μ ij and The calculated cluster center will also be an interval, and the lower boundary v of the cluster center interval is L and upper boundary v R The calculation method is as follows:

[0093]

[0094] Update the cluster center point v according to formula (8) and formula (9) j ;

[0095]

[0096] (5) Convergence judgment: judge whether the number of iterations reaches the maximum value or whether the new centroid is the same as the original centroid; if so, end the algorithm and complete the classification; if not, repeat the above steps until it is satisfied.

[0097] Finally, the interval type 2 fuzzy clustering algorithm is used to divide each terminal user to be assigned to the most matching transformer in the substation, thereby realizing the "transformer-user" topology identification of users in the overlapping area.

[0098] Taking two low-voltage substations under the jurisdiction of a certain region as an example, this paper conducts analysis and verification, and uses the historical voltage data collected by the power grid company's metering automation system to perform Matlab simulation verification.

[0099] For this embodiment, as shown in Table 1, the success rate of collecting the voltage of the substation total meter and user meter in this set of data is above 95%. A data point is collected once per household per hour, and each user has a total of 24 data points. The two low-voltage substations each include 28 single-phase users, the imbalance ratio of the cluster is close to 1:2, and a few voltage series have outlier abnormal data values. Based on the longitude and latitude coordinates in the GIS system, the topological connection relationship between the two distribution substations and the end users is preliminarily determined, among which substation transformer 1 contains 14 users; substation transformer 2 contains 37 users.

[0100] Table 1 Voltage data of users in two low-voltage areas in a certain region (unit: V / volt)

[0101]

[0102] The present invention first preprocesses the data to eliminate data from non-analyzed substations, thereby obtaining clustered basic analysis data. The local outlier factors of some users are shown in Table 2. As shown in Table 2, there are 6 users with a local outlier factor greater than 2, of which 51 still belong to the substation. Manual verification of this substation shows that the users in the non-analyzed substations are consistent with the data preprocessing results. From the above analysis, it can be seen that the proportion of user data in the non-analyzed substations reaches 10.52%. The number of substations in the distribution network is large, and the actual proportion of abnormal data caused by untimely file updates cannot be underestimated.

[0103] Table 2 User local outlier factors

[0104]

[0105] As shown in Figure 5, after removing abnormal outliers, noise and complex information were removed. To fully and three-dimensionally present the experimental results, the voltage data was reduced to three-dimensional space using principal component analysis. Matlab simulation tools were used to display the coordinate values ​​in Figure 5, which represent the data eigenvectors of the data in Table 1 after dimensionality reduction. Figure 5 shows the clustering results of the experimental analysis of the reduced data using the interval type 2 fuzzy clustering algorithm. The blue and green patterns in Figure 5 represent the users of substation transformer 1 and substation transformer 2, respectively. The results of the cluster analysis determined the topological connection relationship between these two distribution transformers and end users. Substation transformer 1 contains 17 users; substation transformer 2 contains 34 users. This corrected the topological connection of the five users misidentified by GIS, completing the precise topology between "transformer-user".

[0106] Because users on the same transformer or branch line are often affected by similar voltage fluctuations, their voltage curves will exhibit similar trends at certain moments. In the above example, after determining the substation to which the user belongs, based on the voltage data of each user within one day, the voltage curves of 10 users in substation transformer 1 and substation transformer 2 were randomly selected, as shown in Figures 6 and 7, respectively. This visualization method can intuitively observe the degree of similarity between different voltage curves. It can be clearly seen from Figures 6 and 7 that the voltage curves of users in the same substation are similar, while the voltage curves of users in different substations are quite different. A comparison of the results of field verification shows that this method successfully achieves accurate identification of household-transformer relationships, verifying the effectiveness and reliability of the algorithm in power system topology identification.

[0107] The above description is only a preferred embodiment of the present invention and is not intended to further limit the present invention. All equivalent changes made using the contents of the present invention description and drawings are within the scope of protection of the present invention.

Claims

1. A low voltage distribution network topology identification method based on interval type 2 fuzzy clustering algorithm, characterized in that: The following steps are involved: Step 1: Based on the longitude and latitude coordinates of the distribution transformer area and the end user in the GIS system and the radiation range limit of the area, preliminarily determine the topological connection between the area transformer and the end user; Step 2: Collect transformer low-voltage side voltage and user voltage data, pre-process the collected user voltage data based on the local abnormal factor algorithm, and obtain user voltage data of the substation to be identified; Step 3: Determine the initial data of the improved interval type 2 fuzzy clustering algorithm, cluster the user voltage data of the area to be identified based on the improved interval type 2 fuzzy clustering algorithm, and finally identify the topological connection relationship of the users in the area, specifically: For the interval type 2 fuzzy clustering algorithm, the correlation coefficient C is introduced ij To describe the consistency of sample data fluctuations, an improved interval type 2 fuzzy clustering algorithm is constructed; the correlation coefficient C ij The expression is as follows: In the formula, is a unit vector, m is the data dimension of the sample, m = sampling frequency * sampling duration; sample i and sample j represent different sample objects respectively, X i and X j are the data sets of sample i and sample j in time dimension m respectively; X id and X jd are the data of sample i and sample j in the dth dimension respectively; X′ represents the transposed matrix of matrix X; Step 3-1: Assume that there are n users in the overlapping area, and the user's voltage vector is represented by x i ,i=1,2,...n;The voltage data of the user to be identified X={x i |i=1,2,...,n}; the number of clusters is k, and the representative points of the mean voltage of transformers in each substation are v j ,j=1,2,...k, i.e. the initial centroid; Step 3-2: Calculate each sample data x i With cluster center v j The correlation coefficient C ij ; Step 3-3: Calculate the fuzzy membership function of each sample data to each cluster. is the upper membership function value, μ ij is the membership function value; z represents the zth cluster, ranging from 1 to k; m1 and m2 are fuzzy coefficients: Step 3-4: The degree of membership of each sample data is an interval By μ ij and The calculated cluster center will also be an interval, and the lower boundary v of the cluster center interval is L and the upper boundary v R The calculation method of is as follows: Update the center point v of the cluster according to formula (8) and formula (9) j ; Step 3-5, determine whether the number of iterations has reached the maximum value, or whether the new centroid is the same as the original centroid; if satisfied, end the algorithm to complete the classification; if not satisfied, repeat the above steps until satisfied; finally, assign each terminal user to be assigned to the most matching transformer in the substation to achieve topological identification of users in the overlapping area.

2. A low voltage distribution network topology identification method based on interval type 2 fuzzy clustering algorithm according to claim 1, characterized in that: Step 1 is as follows: According to the guiding principles of distribution network rules, four types of regional power supply radius are defined: A, B, C, and D, where A represents urban areas, B represents suburban areas, C represents towns, and D represents rural areas; Verify based on the radiation shape of the distribution transformer area and the latitude and longitude information of the end user: With the distribution transformer as the center and r as the radius, the area radiates, and the electricity users covered in the radiation circle are defined as potential users of the area; If a user has one and only one subordinate distribution transformer area, the user is directly identified as belonging to the area where the radiation circle is located; If there are several overlapping areas in the radiation circles of multiple distribution transformer areas, users in the area will be connected to electricity across the areas.

3. A low voltage distribution network topology identification method based on interval type 2 fuzzy clustering algorithm according to claim 2, characterized in that: In step 2, assuming that the voltage data set is U, where p is any sample in U, the collected user voltage data is preprocessed based on the local anomaly factor algorithm to obtain the user voltage data of the area to be identified, specifically: Step 2-1, the expression of k-dist(p) of sample p is as follows: k-dist(p)=d(p,o) (1), Where k-dist(p) represents the k-th distance from p, that is, the distance from p to its k-th neighbor; d(p,o) is the distance between sample p and sample o; Formula (1) satisfies the following conditions: In the data set U, in addition to sample p itself, there are at least k samples p′∈U, satisfying d(p,p′)≤k-dist(p); Step 2-2: The kth neighborhood N of sample p k (p) is expressed as follows: N k (p)={o∈U\{p}|d(p,o)≤k-dist(p)} (2), Where N k (p) represents all samples in the sample set except sample p whose spatial distance to sample p is less than or equal to k-dist(p); Step 2-3: The reachable distance r between sample p and other samples k The expression of (p,o) is as follows: r k (p,o)=max{k-dist(p),d(p,o)} (3), In the formula, r k (p,o) represents the reachable distance between sample p and sample o. The k points closest to point p have the same reachable distance. Step 2-4: Locally accessible density lrd of sample p k The expression of (p) is as follows: Where lrd k (p) is the sample p and the sample N within the kth distance k The inverse of the average reachable distance of (p); |N k (p)| is the number of samples p; Step 2-5: Local outlier factor LOF of sample p k (p) is expressed as follows: If the local outlier factor is greater than 2, it is determined that the user does not belong to the area to be analyzed and the data is removed.

Images