Power distribution network topology and line parameter identification method based on intelligent electric meter

Abstract

The invention relates to the technical field of power distribution network data processing, and discloses a power distribution network topology and line parameter identification method based on an intelligent electric meter, which comprises the following steps of: configuring the intelligent electric meter at a node of a power distribution network system, obtaining injection active power, reactive power and voltage amplitude of all nodes in a set sampling time period, and performing denoising processing to obtain a power distribution network topology and line parameter identification result; obtaining a denoised measurement data set; establishing a voltage power relation equation between the denoised measurement data and the impedance information, converting the voltage power relation equation into a matrix form, and normalizing the data; reactance information matrixes and resistance information matrixes of all the nodes are obtained, an alternating least square regression method is adopted for iterative solution, and optimal reactance information matrixes and resistance information matrixes of all the nodes are obtained; performing clustering analysis based on the optimal resistance information matrix of all nodes to obtain a node level array; based on the node level array, constructing a graph topological structure of the nodes to obtain line parameters; according to the method, the calculation precision of the line parameters in the topological structure of the power distribution network is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of distribution network data processing, and in particular to a distribution network topology and line parameter identification method based on a smart meter. Background Art

[0002] With the continuous advancement of the intelligent construction of power grids, the operational complexity of basic power distribution networks has increased dramatically, and the accurate acquisition of their real-time topological structure has become increasingly difficult. This is mainly because the topology and line parameter identification of the distribution network are highly dependent on detailed and accurate voltage, power and phase angle data. However, in actual operation, the data collection of smart meters is limited by their own measurement capabilities, and problems such as channel blocking during data transmission further affect the integrity and real-time performance of the data. At the same time, the composition and operation mode of the distribution network are becoming increasingly complex and changeable, which puts forward new requirements for the safe and stable operation of the power grid. In addition, factors such as load switching caused by demand response, system failures, and maintenance operations lead to frequent changes in the switch status of connection points on distribution lines. After maintenance activities, the on-site update of topological information is often delayed, and the local topological changes are not fully recorded. In addition, due to the scarcity of topological monitoring equipment and the non-instantaneous nature of remote signal transmission, the exact grid topological status is often unknown.

[0003] Most of the existing distribution network topology identification methods are based on machine learning. Although the use of machine learning-based methods to process distribution network topology identification and line parameter identification can significantly improve processing speed and accuracy, it also has shortcomings. The main problem is that the data is highly dependent, requiring a large amount of high-quality and diverse training data to support the accuracy of the model; at the same time, the model has poor interpretability, and it is difficult to intuitively understand its decision-making process, which increases the difficulty of application in complex power grid environments; in addition, machine learning models may also be affected by overfitting or underfitting, resulting in unstable identification results. Summary of the invention

[0004] In view of the above-mentioned deficiencies in the prior art, the present invention provides a distribution network topology and line parameter identification method based on smart meters, which is used to solve the problems of low data accuracy and inaccurate identification results in the existing distribution network topology identification method.

[0005] In order to achieve the above-mentioned object of the invention, the technical solution adopted by the present invention is:

[0006] The method for identifying the topology and line parameters of a distribution network based on a smart meter includes the following steps:

[0007] S1. Smart meters are configured at all nodes of the distribution network system to obtain the injected active power, injected reactive power and voltage amplitude of all nodes within a set sampling period to obtain a measurement data set, and the measurement data set is denoised to obtain a denoised measurement data set;

[0008] S2. Based on the denoised measurement data set, a voltage-power relationship equation between the denoised measurement data and the impedance information is established, and the voltage-power relationship equation is converted into a matrix form and then the data is normalized to obtain normalized measurement data values ​​of all nodes;

[0009] S3. Based on the normalized measured data values ​​of all nodes, the reactance information matrix and the resistance information matrix of all nodes are obtained, and the alternating least squares regression method is used for iterative solution to obtain the optimal reactance information matrix and the optimal resistance information matrix of all nodes, and then the optimal reactance information matrix and the optimal resistance information matrix of all nodes are obtained and spliced ​​to obtain the impedance information matrix of all nodes;

[0010] S4, based on the optimal resistance information matrix of all nodes, the point order search clustering method is used to perform cluster analysis to obtain the cluster index cluster and core point information to which each node belongs;

[0011] S5. Obtain a node level array according to the cluster index cluster and core point information to which each node belongs, and construct a graph topology structure of the node based on the node level array;

[0012] S6. Based on the graph topology of the nodes, calculate the resistance, reactance, and impedance information of the lines between the nodes.

[0013] Furthermore, step S1 specifically includes:

[0014] S11, configure smart meters at all nodes of the distribution network system, set sampling periods and sampling days, and obtain a measurement data set by recording the injected active power, injected reactive power and voltage amplitude of all nodes in each sampling period within the sampling days;

[0015] S12, denoising the measurement data set to obtain a denoised measurement data set, namely:

[0016]

[0017] Among them, y smooth (t) represents the denoised measurement data set at the tth moment, W represents the radius of the filter, y(th) represents the measurement data set at the th moment, and g(h) represents the Gaussian function.

[0018] Furthermore, step S2 specifically includes:

[0019] S21. Based on the denoised measurement data set, a voltage-power relationship equation between the measurement data of injected reactive power, injected active power, and voltage amplitude including all nodes denoised and impedance information is established, namely:

[0020]

[0021] Among them, v i represents the voltage amplitude of node i after denoising, v r represents the denoised voltage amplitude of the root node r, k represents the set of downstream nodes of node i, l represents the branch, E i represents the path from the root node r to the node i, E j represents the path from root node r to node j, r l represents the resistance on branch l, v j represents the voltage amplitude of the downstream node j of node i after denoising, p j represents the injected active power of the downstream node j of node i after denoising, x l represents the reactance on branch l, q j represents the injected reactive power of the downstream node j of node i to be denoised;

[0022] S22. Convert the voltage-power relationship equation into a matrix form to obtain the voltage-power matrix equation of the measured data, namely:

[0023]

[0024] Among them, V i represents the voltage amplitude information matrix of node i, V0 represents the voltage amplitude information matrix of root node r, R represents the resistance information matrix, P j represents the injected active power of downstream node j, V j represents the voltage amplitude information matrix of downstream node i, X represents the reactance information matrix, Q j represents the injected reactive power of downstream node j;

[0025] S23. Based on the voltage power matrix equation of the measured data, the voltage amplitude information matrix, the injected active power matrix and the injected reactive power matrix of all nodes are obtained to obtain the matrix information set of all nodes, and the matrix information set is normalized to obtain the normalized measured data values ​​of all nodes.

[0026] Furthermore, step S3 specifically includes:

[0027] S31, based on the normalized measured data values ​​of all nodes, the reactance information matrix and the resistance information matrix of all nodes are obtained and initialized. After the initialization is completed, the first iteration is performed to calculate the reactance information matrix of the first iteration of all nodes, that is:

[0028] X (1) =(V′-R (0) P′)Q′ +

[0029] Among them, X (1)represents the first iteration reactance information matrix of all nodes, V′ represents the normalized voltage amplitude matrix of all nodes, R (0) represents the initialized resistance information matrix of all nodes, P′ represents the normalized injected active power of all nodes, Q′ + represents the generalized inverse of the normalized injected reactive power matrix of all nodes;

[0030] S32, based on the first iteration reactance information matrix of all nodes, calculate the first iteration resistance information matrix of all nodes, that is:

[0031] R (1) =(V′-X (1) Q′)P′ +

[0032] Among them, R (1) represents the resistance information matrix of the first iteration of all nodes, Q′ represents the normalized injected reactive power matrix of all nodes, P′ + represents the generalized inverse of the normalized injected active power matrix P′ of all nodes;

[0033] S33, repeating steps S31-S32 for multiple iterative calculations until the change in the Frobenius norm of the resistance information matrix of all nodes and the reactance information matrix of all nodes is less than a set threshold, then stopping the iteration, and obtaining the optimal reactance information matrix of all nodes and the optimal resistance information matrix of all nodes;

[0034] S34, concatenating the optimal reactance information matrix of all nodes with the optimal resistance information matrix of all nodes to obtain the impedance information matrix of all nodes.

[0035] Furthermore, step S4 specifically includes:

[0036] Initialize the minimum number of clustering points as MinPts and the minimum neighbor threshold distance as ε. Take the optimal resistance information matrix of all nodes as the input of the point order search clustering method, perform clustering operation on each column of the optimal resistance information matrix of all nodes, and obtain the clustering index cluster and core point information to which each node belongs.

[0037] The core point information is the centroid of each clustering index cluster, which is used to define and represent the central position and shape characteristics of each clustering index cluster.

[0038] Furthermore, the clustering operation is based on the reachable distance between nodes. If the reachable distances between nodes are the same, the nodes with the same reachable distances are divided into the same clustering index cluster.

[0039] Furthermore, the calculation formula for the reachable distance between nodes is:

[0040] rd(i,j)=max(core_distance(j),d(i,j))

[0041] Where rd(i, j) represents the reachable distance between node i and node j, core_distance(j) represents the core distance of node j, and d(i, j) represents the direct distance between node i and node j;

[0042] The core distance is the minimum neighborhood distance obtained from node j that can contain at least MinPts points.

[0043] Furthermore, step S5 specifically includes:

[0044] S51, obtaining a node level array set N through clustering operation according to the cluster index cluster and core point information to which each node belongs;

[0045] Wherein, N={N{1}, ..., N{e}}, N{1} represents the first node level array, and N{e} represents the e-th node level array;

[0046] S52, taking the node in the first node level array N{1} in the node level array set as the initial root node of the graph, and constructing the graph topology structure of the nodes by traversing the node level array set, specifically:

[0047] Take the node in the first node level array N{1} as the initial root node of the graph, start from the initial root node and traverse each node in all node level arrays except the first node level array N{1} in turn, use the point order search clustering method to obtain the clustering index cluster of the current node, and determine whether the clustering index cluster of the current node contains the node of the previous node level array. If so, add the edge from the node in the previous node level array to the current node to obtain the graph topology of the node, otherwise, do not add the edge;

[0048] Among them, the edges are the lines between nodes in the distribution network system.

[0049] Furthermore, the graph topology of the nodes in step S52 is:

[0050] G′=G∪{(i′,j′)|i′∈(C(i′)∩N L-1 ), j′∈C(i′)∩N L}

[0051] Where G′ represents the graph topology of the node, G represents the initial graph, i′ represents the node in the previous node level array, j′ represents the current node, C(i′) represents the numbered set of cluster index clusters to which node i′ belongs, and N L-1Represents the node set of the L-1th node level array, N L A node set representing the Lth node level array.

[0052] Furthermore, step S6 specifically includes:

[0053] Based on the graph topology of the nodes, the resistance, reactance, and impedance information of the lines between the nodes are calculated, namely:

[0054]

[0055] Among them, r ij Represents the resistance information of the line between node i and node j, R′ jj Represents the resistance information from node j to node j in the optimal resistance information matrix of all nodes, R′ ii Represents the resistance information from node i to node i in the optimal resistance information matrix of all nodes, R′ ij represents the resistance information from node i to node j in the optimal resistance information matrix of all nodes, x ij Represents the reactance information of the line between node i and node j, X′ jj represents the reactance information from node j to node j in the optimal reactance information matrix of all nodes, X′ ii represents the reactance information from node i to node i in the optimal reactance information matrix of all nodes, X′ ij represents the reactance information from node i to node j in the optimal reactance information matrix of all nodes, z ij It represents the impedance information of the line between node i and node j, and a represents the imaginary part.

[0056] The present invention has the following beneficial effects:

[0057] 1. The distribution network topology and line parameter identification method based on smart meters proposed in the present invention configures smart meters at the nodes of the distribution network system to record the injected active power, injected reactive power and voltage amplitude of the nodes to obtain real-time distribution network status information, and denoises the recorded node data to improve data quality and ensure the accuracy and reliability of subsequent analysis and graph topology structure establishment;

[0058] 2. Establish the relationship equation between the measured data and the resistance information matrix and reactance information matrix, and solve the equation by alternating least squares method to obtain the impedance information matrix, which can more accurately solve the impedance information matrix and provide a more accurate data basis;

[0059] 3. Extract the resistance information matrix from the impedance information matrix, that is, use the point-order search clustering method to analyze the optimal resistance information matrix of all nodes column by column to determine the level of the node in the topology graph. At the same time, based on the level information, the edge information is gradually added in each layer through clustering operation again to construct a complete distribution network topology structure. Finally, combined with the optimal resistance and reactance information matrices of all nodes, the line parameters of all lines, namely resistance, reactance and impedance information, are calculated to realize the graph topology structure and line parameter identification, thereby improving the reliability, safety and economy of the distribution network system. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 A schematic diagram of the flow of the distribution network topology and line parameter identification method based on smart meters proposed in the present invention;

[0061] Figure 2 It is a schematic diagram of the solution results of the impedance information matrix of all nodes;

[0062] Figure 3 Schematic diagram of topology identification of distribution network. DETAILED DESCRIPTION

[0063] The specific implementation modes of the present invention are described below so that those skilled in the art can understand the present invention. However, it should be clear that the present invention is not limited to the scope of the specific implementation modes. For those of ordinary skill in the art, as long as various changes are within the spirit and scope of the present invention as defined and determined by the attached claims, these changes are obvious, and all inventions and creations utilizing the concept of the present invention are protected.

[0064] like Figure 1 As shown, the distribution network topology and line parameter identification method based on smart meters includes the following steps:

[0065] S1. Smart meters are configured at all nodes of the distribution network system to obtain the injected active power, injected reactive power and voltage amplitude of all nodes within a set sampling period to obtain a measurement data set, and the measurement data set is denoised to obtain a denoised measurement data set.

[0066] In this embodiment, a smart meter is configured at each node in the IEEE 33-node distribution network system, and the initial line parameters are set, the sampling period is set to 15 minutes, and the sampling days are set to 7 days, that is, the injected active power, injected reactive power and voltage amplitude of the smart meter are recorded every 15 minutes within 7 days, and the injected active power, injected reactive power and voltage amplitude of all nodes are obtained, and denoising is performed; wherein, the purpose of denoising is to improve data quality to ensure the accuracy and reliability of subsequent analysis and topology map establishment.

[0067] Specifically, step S1 includes S11-S12:

[0068] S11. Smart meters are configured at all nodes of the distribution network system, and sampling periods and sampling days are set. The measurement data set is obtained by recording the injected active power, injected reactive power and voltage amplitude of all nodes in each sampling period within the sampling days.

[0069] In this embodiment, the measurement data set consists of measurement data, and the measurement data consists of injected active power, injected reactive power and voltage amplitude of the node.

[0070] S12, denoising the measurement data set to obtain a denoised measurement data set, namely:

[0071]

[0072] Among them, y smooth (t) represents the denoised measurement data set at the tth moment, W represents the radius of the filter, y(th) represents the measurement data set at the th moment, and g(h) represents the Gaussian function.

[0073] S2. Based on the denoised measurement data set, a voltage-power relationship equation between the denoised measurement data and the impedance information is established, and the voltage-power relationship equation is converted into a matrix form and then the data is normalized to obtain the normalized measurement data values ​​of all nodes.

[0074] Specifically, step S2 specifically includes S21-S23:

[0075] S21. Based on the denoised measurement data set, a voltage-power relationship equation between the measurement data of injected reactive power, injected active power, and voltage amplitude including all nodes denoised and impedance information is established, namely:

[0076]

[0077] Among them, v i represents the voltage amplitude of node i after denoising, v r represents the denoised voltage amplitude of the root node r, k represents the set of downstream nodes of node i, l represents the branch, E i represents the path from the root node r to the node i, E j represents the path from root node r to node j, r l represents the resistance on branch l, v j represents the voltage amplitude of the downstream node j of node i after denoising, p j represents the injected active power of the downstream node j of node i after denoising, x l represents the reactance on branch l, q jrepresents the injected reactive power denoised by the downstream node j of node i.

[0078] In this embodiment, the purpose of establishing the voltage-power relationship equation is to facilitate solving the impedance information matrix in the subsequent steps.

[0079] S22. Convert the voltage-power relationship equation into a matrix form to obtain the voltage-power matrix equation of the measured data, namely:

[0080]

[0081] Among them, V i represents the voltage amplitude information matrix of node i, V0 represents the voltage amplitude information matrix of root node r, R represents the resistance information matrix, P j represents the injected active power of downstream node j, V j represents the voltage amplitude information matrix of downstream node i, X represents the reactance information matrix, Q j represents the injected reactive power of downstream node j.

[0082] In this embodiment, the purpose of converting the voltage-power relationship equation into a matrix form is: since the measured data has more than one time section, the measured data needs to be represented in a matrix form to facilitate subsequent operations.

[0083] S23. Based on the voltage power matrix equation of the measured data, the voltage amplitude information matrix, the injected active power matrix and the injected reactive power matrix of all nodes are obtained to obtain the matrix information set of all nodes, and the matrix information set is normalized to obtain the normalized measured data values ​​of all nodes.

[0084] In this embodiment, the matrix information set of all nodes includes the voltage amplitude information matrix, the injected active power matrix and the injected reactive power matrix of all nodes. At the same time, the purpose of data normalization is to unify the dimensions.

[0085] S3. Based on the normalized measured data values ​​of all nodes, the reactance information matrix and resistance information matrix of all nodes are obtained, and the alternating least squares regression method is used for iterative solution to obtain the optimal reactance information matrix and the optimal resistance information matrix of all nodes, and then they are spliced ​​to obtain the impedance information matrix of all nodes.

[0086] Specifically, step S3 specifically includes S31-S34:

[0087] S31, based on the normalized measured data values ​​of all nodes, the reactance information matrix and the resistance information matrix of all nodes are obtained and initialized. After the initialization is completed, the first iteration is performed to calculate the reactance information matrix of the first iteration of all nodes, that is:

[0088] X (1) =(V′-R (0) P′)Q′ +

[0089] Among them, X (1) represents the first iteration reactance information matrix of all nodes, V′ represents the normalized voltage amplitude matrix of all nodes, R (0) represents the initialized resistance information matrix of all nodes, P′ represents the normalized injected active power of all nodes, Q′ + represents the generalized inverse of the normalized injected reactive power matrix of all nodes.

[0090] In this embodiment, after data normalization, the voltage power matrix equation of the measured data becomes: V′-R (0) P′=X (1) Q′, so by fixing R, we can solve for X.

[0091] S32, based on the first iteration reactance information matrix of all nodes, calculate the first iteration resistance information matrix of all nodes, that is:

[0092] R (1) =(V′-X (1) Q′)P′ +

[0093] Among them, R (1) represents the resistance information matrix of the first iteration of all nodes, Q′ represents the normalized injected reactive power matrix of all nodes, P′ + represents the generalized inverse of the normalized injected active power matrix P′ of all nodes.

[0094] In this embodiment, after the first iteration, based on R (1) , the voltage power matrix equation becomes: V′-X (1) Q′=R (1) P′, so based on this equation, by fixing X, we can solve for R.

[0095] S33, repeat steps S31-S32 for multiple iterative calculations until the change in the Frobenius norm of the resistance information matrix of all nodes and the reactance information matrix of all nodes is less than the set threshold, then stop the iteration and obtain the optimal reactance information matrix of all nodes and the optimal resistance information matrix of all nodes.

[0096] In this embodiment, the threshold is set to 10 -6 .

[0097] S34, concatenating the optimal reactance information matrix of all nodes with the optimal resistance information matrix of all nodes to obtain the impedance information matrix of all nodes.

[0098] In this embodiment, the impedance information matrix of all nodes records the topological connection of the distribution network system and the parameter values ​​of the line parameters, as shown in FIG. Figure 2 As shown, Figure 2 The horizontal and vertical axes represent the row and column numbers of the matrix, which actually represent the node numbers. For example, the value of the second row and third column actually represents the sum of the resistances on the common path of the path E2 from node 2 to the root node and the path E3 from node 3 to the root node. In addition, Figure 2 The depth of the color represents the numerical value of the elements in the information matrix.

[0099] S4. Based on the optimal resistance information matrix of all nodes, the point order search clustering method is used to perform clustering analysis to obtain the cluster index cluster and core point information to which each node belongs.

[0100] Specifically, step S4 includes:

[0101] The minimum number of points for initializing clustering is MinPts, and the minimum neighbor threshold distance is ε. The optimal resistance information matrix of all nodes is used as the input of the point-sequential search clustering method. Clustering operation is performed on each column of the optimal resistance information matrix of all nodes to obtain the clustering index cluster and core point information of each node.

[0102] The core point information is the centroid of each clustering index cluster, which is used to define and represent the central position and shape characteristics of each clustering index cluster.

[0103] Specifically, the clustering operation is based on the reachable distance between nodes. If the reachable distances between nodes are the same, the nodes with the same reachable distances are divided into the same clustering index cluster.

[0104] Specifically, the calculation formula for the reachable distance between nodes is:

[0105] rd(i,j)=max(core_distance(j),d(i,j))

[0106] Among them, rd(i,j) represents the reachable distance between node i and node j, core_distance(j) represents the core distance of node j, and d(i,j) represents the direct distance between node i and node j;

[0107] The core distance is the minimum neighborhood distance obtained from node j that can contain at least minPts points.

[0108] S5. According to the cluster index cluster and core point information to which each node belongs, a node level array is obtained, and based on the node level array, a graph topology structure of the node is constructed.

[0109] Specifically, step S5 includes S51-S52:

[0110] S51. According to the cluster index cluster and core point information to which each node belongs, a node level array set N is obtained through clustering operation; wherein, N = {N{1}, ..., N{e}}, N{1} represents the first node level array, and N{e} represents the e-th node level array.

[0111] In this embodiment, the above steps are used to perform clustering operations on each column in the optimal resistance information matrix of all nodes, and the level of the node in the graph is determined by traversing each column. The clustering results are stored in the node level data set N, and each cell in the node level data set N corresponds to a level and contains the indexes of all nodes with that level; for example: the first node level array N{1} contains the indexes of all nodes with level 1, and the second node level array N{2} contains the indexes of all nodes with level 2, and traverses in sequence until the last node level array N{e}; in this embodiment, e is 17, that is, there are 17 node level arrays, and the traversal execution results are shown in Table 1:

[0112] Table 1 Node level array result table

[0113]

[0114]

[0115] Among them, Table 1 shows the 32 nodes and the node level array to which they belong.

[0116] S52, taking the node in the first node level array N{1} in the node level array set as the initial root node of the graph, and constructing the graph topology structure of the nodes by traversing the node level array set, specifically:

[0117] Take the node in the first node level array N{1} as the initial root node of the graph, start from the initial root node and traverse each node in all node level arrays except the first node level array N{1} in turn, use the point order search clustering method to obtain the clustering index cluster of the current node, and determine whether the clustering index cluster of the current node contains the node of the previous node level array. If so, add the edge from the node in the previous node level array to the current node to obtain the graph topology of the node, otherwise, do not add the edge;

[0118] Among them, the edges are the lines between nodes in the distribution network system.

[0119] Specifically, the graph topology of the nodes in step S52 is:

[0120] G′=G∪{(i′,j′)|i′∈(C(i′)∩N L-1 ), j′∈C(i′)∩N L}

[0121] Where G′ represents the graph topology of the node, G represents the initial graph, i′ represents the node in the previous node level array, j′ represents the current node, C(i′) represents the numbered set of cluster index clusters to which node i′ belongs, and N L-1 Represents the node set of the L-1th node level array, N L A node set representing the Lth node level array.

[0122] In this embodiment, a directed graph G is initialized, and the nodes in the first node level array N{1} are taken as the root nodes of the graph G. Then, N{2}-N{e} are traversed in sequence from the first node level array N{1}, and the resistance information matrix numbers corresponding to the values ​​contained in each node level array are clustered layer by layer using the point search clustering method to build the graph topology structure of the nodes layer by layer, that is, the distribution network topology structure.

[0123] S6. Based on the graph topology of the nodes, calculate the resistance, reactance, and impedance information of the lines between the nodes.

[0124] In this embodiment, the resistance, reactance and impedance information of the lines between the nodes are obtained based on the graph topology of the nodes and the optimal resistance information matrix and reactance information matrix of all nodes. The topology identification result is as follows: Figure 3 As shown, Figure 3 It is a topological graph consisting of points and edges, where the points are nodes in the distribution network system and the edges are lines in the distribution network system.

[0125] Specifically, step S6 includes:

[0126] Based on the graph topology of the nodes, the resistance, reactance, and impedance information of the lines between the nodes are calculated, namely:

[0127]

[0128] Among them, r ij Represents the resistance information of the line between node i and node j, R′ jj Represents the resistance information from node j to node j in the optimal resistance information matrix of all nodes, R′ ii Represents the resistance information from node i to node i in the optimal resistance information matrix of all nodes, R′ ijrepresents the resistance information from node i to node j in the optimal resistance information matrix of all nodes, x ij Represents the reactance information of the line between node i and node j, X′ jj represents the reactance information from node j to node j in the optimal reactance information matrix of all nodes, X′ ii represents the reactance information from node i to node i in the optimal reactance information matrix of all nodes, X′ ij represents the reactance information from node i to node j in the optimal reactance information matrix of all nodes, z ij It represents the impedance information of the line between node i and node j, and a represents the imaginary part.

[0129] In this embodiment, the error between the calculated impedance information and the true value is used to verify the effectiveness of the distribution network topology and line parameter identification method based on the smart meter proposed in the present invention, that is:

[0130]

[0131]

[0132] In Table 1, the real value data is derived from the initial line parameters set in the embodiment. It can be seen from Table 1 that the error between the impedance information calculated by the present invention and the real value is small, so the method proposed by the present invention is effective.

[0133] In summary, the distribution network topology and line parameter identification method based on smart meters proposed in the present invention first configures smart meters at all nodes of the distribution network, records the active power injected by the nodes, the reactive power injected by the nodes, and the voltage amplitude of the nodes, and performs denoising on the data after collecting them. By recording the active power, reactive power, and voltage amplitude of the nodes, real-time distribution network status information is obtained, and the data quality is improved by denoising to ensure the accuracy and reliability of subsequent analysis and establishment of graph topology structure; secondly, establishes the relationship equation between the measured data and the resistance information matrix and the reactance information matrix, and solves the equation by alternating least squares method to obtain to the impedance information matrix in order to more accurately solve the impedance information matrix and provide a more precise data basis; finally, the resistance information matrix is ​​extracted from the impedance information matrix, and the matrix is ​​analyzed column by column using the point-order search clustering algorithm to determine the level of the node in the topology diagram. Based on the level information, the nodes and edges added in each layer are gradually determined through clustering operations again, and then a complete distribution network topology is constructed. Combined with the optimal resistance and reactance information matrix of all nodes, the parameters of all lines, namely resistance, reactance and impedance information, are calculated to realize the identification of topological structure and line parameters, thereby improving the reliability, safety and economy of the distribution network.

[0134] The present invention uses specific embodiments to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

[0135] Those skilled in the art will appreciate that the embodiments described herein are intended to help readers understand the principles of the present invention, and should be understood that the protection scope of the present invention is not limited to such specific statements and embodiments. Those skilled in the art can make various other specific variations and combinations that do not deviate from the essence of the present invention based on the technical revelations disclosed by the present invention, and these variations and combinations are still within the protection scope of the present invention.

Claims

1. A distribution network topology and line parameter identification method based on smart meters, characterized in that: The following steps are involved: S1. Smart meters are configured at all nodes of the distribution network system to obtain the injected active power, injected reactive power and voltage amplitude of all nodes within a set sampling period to obtain a measurement data set, and the measurement data set is denoised to obtain a denoised measurement data set; S2. Based on the denoised measurement data set, a voltage-power relationship equation between the denoised measurement data and the impedance information is established, and the voltage-power relationship equation is converted into a matrix form and then the data is normalized to obtain normalized measurement data values ​​of all nodes; S3. Based on the normalized measured data values ​​of all nodes, the reactance information matrix and the resistance information matrix of all nodes are obtained, and the alternating least squares regression method is used for iterative solution to obtain the optimal reactance information matrix and the optimal resistance information matrix of all nodes, and then the optimal reactance information matrix and the optimal resistance information matrix of all nodes are obtained and spliced ​​to obtain the impedance information matrix of all nodes; S4, based on the optimal resistance information matrix of all nodes, the point order search clustering method is used to perform cluster analysis to obtain the cluster index cluster and core point information to which each node belongs; S5. Obtain a node level array according to the cluster index cluster and core point information to which each node belongs, and construct a graph topology structure of the node based on the node level array; S6. Based on the graph topology of the nodes, calculate the resistance, reactance, and impedance information of the lines between the nodes.

2. The method for identifying distribution network topology and line parameters based on smart meters according to claim 1, characterized in that: Step S1 specifically includes: S11, configure smart meters at all nodes of the distribution network system, set sampling periods and sampling days, and obtain a measurement data set by recording the injected active power, injected reactive power and voltage amplitude of all nodes in each sampling period within the sampling days; S12, denoising the measurement data set to obtain a denoised measurement data set, namely: Among them, y smooth (t) represents the denoised measurement data set at the tth moment, W represents the radius of the filter, y(th) represents the measurement data set at the th moment, and g(k) represents the Gaussian function.

3. The method for identifying distribution network topology and line parameters based on smart meters according to claim 2, characterized in that: Step S2 specifically includes: S21. Based on the denoised measurement data set, a voltage-power relationship equation between the measurement data of injected reactive power, injected active power, and voltage amplitude including all nodes denoised and impedance information is established, namely: Among them, v i represents the voltage amplitude of node i after denoising, v r represents the denoised voltage amplitude of the root node r, k represents the set of downstream nodes of node i, l represents the branch, E i represents the path from the root node r to the node i, E j represents the path from root node r to node j, r l represents the resistance on branch l, v j represents the voltage amplitude of the downstream node j of node i after denoising, p j represents the injected active power of the downstream node j of node i after denoising, x l represents the reactance on branch l, q j represents the injected reactive power of the downstream node j of node i to be denoised; S22. Convert the voltage-power relationship equation into a matrix form to obtain the voltage-power matrix equation of the measured data, namely: Among them, V i represents the voltage amplitude information matrix of node i, V0 represents the voltage amplitude information matrix of root node r, R represents the resistance information matrix, P j represents the injected active power of downstream node j, V j represents the voltage amplitude information matrix of downstream node i, X represents the reactance information matrix, Q j represents the injected reactive power of downstream node j; S23. Based on the voltage power matrix equation of the measured data, the voltage amplitude information matrix, the injected active power matrix and the injected reactive power matrix of all nodes are obtained to obtain the matrix information set of all nodes, and the matrix information set is normalized to obtain the normalized measured data values ​​of all nodes.

4. The method for identifying distribution network topology and line parameters based on smart meters according to claim 3 is characterized in that: Step S3 specifically includes: S31, based on the normalized measured data values ​​of all nodes, the reactance information matrix and the resistance information matrix of all nodes are obtained and initialized. After the initialization is completed, the first iteration is performed to calculate the reactance information matrix of the first iteration of all nodes, that is: X (1) =(V ′ -R (0) P ′ )Q ′+ Among them, X (1) represents the first iteration reactance information matrix of all nodes, V ′ represents the normalized voltage magnitude matrix of all nodes, R (0) Represents the resistance information matrix of all nodes initialized, P ′ represents the normalized injected active power of all nodes, Q ′+ represents the generalized inverse of the normalized injected reactive power matrix of all nodes; S32, based on the first iteration reactance information matrix of all nodes, calculate the first iteration resistance information matrix of all nodes, that is: R (1) =(V ′ -X (1) Q ′ )P ′+ Among them, R (1) represents the resistance information matrix of all nodes in the first iteration, Q ′ represents the normalized injected reactive power matrix of all nodes, P ′+ represents the normalized active power matrix P of all nodes injected ′ The generalized inverse of S33, repeating steps S31-S32 for multiple iterative calculations until the change in the Frobenius norm of the resistance information matrix of all nodes and the reactance information matrix of all nodes is less than a set threshold, then stopping the iteration, and obtaining the optimal reactance information matrix of all nodes and the optimal resistance information matrix of all nodes; S34, concatenating the optimal reactance information matrix of all nodes with the optimal resistance information matrix of all nodes to obtain the impedance information matrix of all nodes.

5. The method for identifying distribution network topology and line parameters based on smart meters according to claim 4 is characterized in that: Step S4 specifically includes: Initialize the minimum number of clustering points as MinPts and the minimum neighbor threshold distance as ε. Take the optimal resistance information matrix of all nodes as the input of the point order search clustering method, perform clustering operation on each column of the optimal resistance information matrix of all nodes, and obtain the clustering index cluster and core point information to which each node belongs. The core point information is the centroid of each clustering index cluster, which is used to define and represent the central position and shape characteristics of each clustering index cluster.

6. The method for identifying distribution network topology and line parameters based on smart meters according to claim 5, characterized in that: The basis of clustering operation is the reachable distance between nodes. If the reachable distance between nodes is the same, the nodes with the same reachable distance will be divided into the same clustering index cluster.

7. The method for identifying distribution network topology and line parameters based on smart meters according to claim 6, characterized in that: The calculation formula for the reachable distance between nodes is: rd(i,j)=max(core_distance(j),d(i,j)) Among them, rd(i,j) represents the reachable distance between node i and node j, core_distance(j) represents the core distance of node j, and d(i,j) represents the direct distance between node i and node j; The core distance is the minimum neighborhood distance obtained from node j that can contain at least MinPts points.

8. The method for identifying distribution network topology and line parameters based on smart meters according to claim 7, characterized in that: Step S5 specifically includes: S51, obtaining a node level array set N through clustering operation according to the cluster index cluster and core point information to which each node belongs; Wherein, N={N{1},…,N{e}}, N{1} represents the first node level array, and N{e} represents the e-th node level array; S52, taking the node in the first node level array N{1} in the node level array set as the initial root node of the graph, and constructing the graph topology structure of the nodes by traversing the node level array set, specifically: Take the node in the first node level array N{1} as the initial root node of the graph, start from the initial root node and traverse each node in all node level arrays except the first node level array N{1} in turn, use the point order search clustering method to obtain the clustering index cluster of the current node, and determine whether the clustering index cluster of the current node contains the node of the previous node level array. If so, add the edge from the node in the previous node level array to the current node to obtain the graph topology of the node, otherwise, do not add the edge; Among them, the edges are the lines between nodes in the distribution network system.

9. The method for identifying distribution network topology and line parameters based on smart meters according to claim 8, characterized in that: The graph topology of the nodes in step S52 is: G′=G∪{(i′,j′)|i′∈(C(i′)∩N L-1 ),j′∈C(i′)∩N L } Where G′ represents the graph topology of the node, G represents the initial graph, i′ represents the node in the previous node level array, j′ represents the current node, C(i′) represents the numbered set of cluster index clusters to which node i′ belongs, and N L-1 Represents the node set of the L-1th node level array, N L A node set representing the Lth node level array.

10. The method for identifying distribution network topology and line parameters based on smart meters according to claim 9, characterized in that: Step S6 specifically includes: Based on the graph topology of the nodes, the resistance, reactance, and impedance information of the lines between the nodes are calculated, namely: Among them, r ij Represents the resistance information of the line between node i and node j, R j ′ j Represents the resistance information from node j to node j in the optimal resistance information matrix of all nodes, R i ′ i Represents the resistance information from node i to node i in the optimal resistance information matrix of all nodes, R i ′ j represents the resistance information from node i to node j in the optimal resistance information matrix of all nodes, x ij represents the reactance information of the line between node i and node j, X j ′ j represents the reactance information from node j to node j in the optimal reactance information matrix of all nodes, X i ′ i represents the reactance information from node i to node i in the optimal reactance information matrix of all nodes, X i ′ j represents the reactance information from node i to node j in the optimal reactance information matrix of all nodes, z ij It represents the impedance information of the line between node i and node j, and a represents the imaginary part.

Images