Power distribution network low-voltage transformer area topology identification method based on tree diagram coding representation

Abstract

The invention provides a power distribution network low-voltage transformer area topology identification method based on tree diagram coding representation, and belongs to the technical field of power distribution networks, and the method comprises the steps: S1, carrying out the identification of a user phase of a low-voltage power distribution network based on the Kirchhoff's current law, and converting the three-phase topology identification into single-phase identification; s2, expressing the topological structure of the low-voltage power distribution network through the tree diagram, and coding based on the expressed tree diagram; s3, calculating an impedance matrix based on a gradient descent method; s4, defining optimization indexes of user node voltage and prediction deviation; and S5, optimizing in the potential topological space through a genetic algorithm, and outputting an optimal low-voltage power distribution network topological structure. According to the method, the potential topological structure of the low-voltage power distribution network is represented on the basis of tree diagram coding, the deviation between the predicted value and the measured value of the user node voltage is directly used as an optimization index, complete topological identification of the low-voltage power distribution network can be achieved, and the method has high accuracy.

Description

Technical Field

[0001] The present invention belongs to the technical field of distribution network, and in particular relates to a method for identifying low-voltage distribution network topology based on tree diagram coding representation. Background Art

[0002] The relationship between low-voltage distribution networks in substations is the basis for carrying out intelligent services of low-voltage distribution networks in other substations, which can help solve various power grid services such as accurate determination of power outage scope and line loss statistics. However, in the process of power grid operation and maintenance, there is a lack of detailed records of low-voltage distribution network topology information. In addition, after the overall construction of the distribution network, the topology of the power grid is checked, which is not only cumbersome but also inefficient, and it is difficult to meet the requirements of voltage prediction for real-time and accuracy. Only by accurately identifying the connection relationship of the power grid can the flow calculation and voltage prediction be effectively carried out.

[0003] Distribution network topology identification based on metering data analysis performs well in high-voltage distribution networks. However, in low-voltage distribution networks, due to incomplete metering facilities, low-voltage distribution network topology identification based on metering data is extremely challenging and has the following disadvantages: it is limited to partial topology identification of low-voltage distribution networks, that is, it focuses on household phase / household transformer identification and lacks further branch identification. The low-voltage distribution network topology obtained in this way is incomplete and has low accuracy. Summary of the invention

[0004] In order to solve the above problems existing in the prior art, a method for identifying low-voltage substation topology of a distribution network based on tree diagram coding representation is provided.

[0005] The technical solution adopted by the present invention to solve its technical problem is:

[0006] This technical solution proposes a method for identifying the topology of low-voltage distribution network areas based on tree diagram encoding representation, including the following steps:

[0007] S1: Based on Kirchhoff's current law, the user phase of the low-voltage distribution network is identified, and the three-phase topology identification is converted into single-phase identification;

[0008] S2: Represent the topology of the low voltage distribution network through a tree diagram, and perform encoding based on the represented tree diagram;

[0009] S3: Calculate the impedance matrix based on the gradient descent method;

[0010] S4: define the optimization index of user node voltage and prediction deviation;

[0011] S5: Use genetic algorithm to find the best solution in the potential topology space and output the optimal low-voltage distribution network topology.

[0012] Preferably, in S1, for the power users on each phase of the same low-voltage distribution network, based on formula (1) as the optimization target, the user phase is identified through 0-1 planning, and the formula is as follows:

[0013] min Z=|I p,A -∑I p,A,user |+|I p,B -∑I p,B,user |+|I p,C -∑I p,C,user | (1);

[0014] In the formula, I p,A ,I p,B ,I p,C They represent the current values ​​of the transformer output phase A, phase B, and phase C in the same direction as the voltage after the current is decomposed according to the voltage direction. p,A,user ,I p,B,user ,I p,C,user Respectively represent the current values ​​in the same direction as the voltage after the users of phase A, phase B and phase C are decomposed according to the voltage direction;

[0015] Based on formula (2), the optimization problem of formula (1) is transformed into a least squares problem, and the formula is as follows:

[0016]

[0017] In the formula, I p,user Represents the active current matrix of all users, and matrices A, B, and C represent the current I of each user respectively. p,user The proportion of each phase in the total current;

[0018] After phase identification of the low-voltage distribution network, the three-phase topology identification is converted into single-phase identification.

[0019] Preferably, in S2, the topology structure of the low-voltage distribution network is abstracted into a tree structure, with the transformer as the root node, the power users as leaf nodes, the number of leaf nodes being n, and each leaf node representing different information. A tree is generated with a root node degree of 2 and an internal node degree of 3, and the internal nodes are other nodes in the tree except the root node and the leaf nodes.

[0020] Preferably, the tree with n=2 is used as the starting tree, and the tree with n=3 is generated by adding leaf nodes to the tree with n=2. Every time a leaf node is added, an internal node is added simultaneously. The degree of the internal nodes of the tree with n=3 is 3. An internal node is added to the edge of the original tree, and the newly added leaf node is connected to the internal node.

[0021] Preferably, assuming that the root node of the original tree is an internal node of the new tree, the root node of the new tree is connected to the internal node to generate a virtual edge, the root node of the new tree is a virtual root node, and when the new leaf node is connected to the virtual root node, a virtual edge is generated.

[0022] When a new leaf node is added, a new intermediate node is also added. The new intermediate node eliminates the original edge and adds 3 new edges. The number of positions that can be placed after adding the fourth leaf node is 5. And so on, the number of trees with n different leaf nodes is m, and the formula is as follows:

[0023]

[0024] Where m is the possible number of trees and n is the number of different leaf nodes.

[0025] Preferably, the edges of the tree are actual power transmission lines, and the edge values ​​are line impedances, converting the original tree into a binary tree;

[0026] Connect nodes b, c, d to node a. The circuit information on nodes a, b, c, d is the same, and the formula is as follows:

[0027]

[0028] In the formula, Represents the voltage of nodes a, b, and c respectively, Z ab , Z ac They represent the impedance values ​​of the line between the node ab and the node ac respectively;

[0029] Combining equation (4) and equation (5) we can get equation (6), which is as follows:

[0030]

[0031] After connecting nodes b and c to node e, and then connecting nodes e and d to node a, the circuit information on nodes a, b, c, and d is the same. Similarly, the formula can be obtained as follows:

[0032]

[0033] Equations (6) and (7) are valid at any time when the circuit is running, and the solution is Z ac =Z ec , Z ab =Z eb , so the voltage at point e is equal to the voltage at point a;

[0034] The degree of the root node and the leaf node are both 1. The edge connected to the root node is the transformer outgoing line, and the edge connected to the leaf node is the household power supply line. The node representing the transformer in the tree is defined as the root node. The possibility of adding a new leaf node to generate a tree is equal to the number of edges in the original tree.

[0035] Preferably, after the encoding is completed, the binary tree is described and generated by edges connecting points. In a binary tree with n leaf nodes, number 0 is used as its root node, numbers 1 to n are used as its leaf nodes, and numbers n+1 to 2n-1 represent the internal nodes of the tree. The initial tree structure is represented in the form of [(0, n+1), (n+1, 1), (n+1, 2)]. The i-th edge in the structure is deleted, and the edge consisting of the starting node of the i-th edge and the new internal node, the edge consisting of the new internal node and the ending node of the i-th edge, and the edge consisting of the new internal node and the new leaf node are added in sequence.

[0036] Preferably, in S3, multiple time data are used to iteratively solve the circuit parameters, and ΔU(T), I p (T), I q (T), let A(T) = [I p (T) I q (T)], b(T) = ΔU(T), and we get equation (8), which is as follows:

[0037]

[0038] In the formula, I p (T), I q (I) represents the current value in the same direction as the voltage and perpendicular to the voltage after decomposition according to the voltage direction, ΔU(T) represents the voltage difference between the user and the transformer, and R and X represent the line impedance;

[0039] The solution of the linear equation system of formula (8) is transformed into an optimization problem as shown in formula (9), and the objective function F(x) is defined as follows:

[0040]

[0041] Iterative calculation is performed based on the gradient descent method. The formula is as follows:

[0042]

[0043] Among them, α is the learning rate, is the objective function F(R,X) in R (t) ,X (t) The gradient at , t is the number of iterations.

[0044] Preferably, the calculation formula of the gradient is as follows:

[0045]

[0046] In the formula, A(T)=[I p (T) I q (T)],b(T)=ΔU(T),I p (T), I q (T) represents the current value in the same direction as the voltage and the current value perpendicular to the voltage after decomposition according to the voltage direction at time T, ΔU(T) represents the voltage difference between the user and the transformer at time T, R and X represent the line impedance, and t is the number of iterations;

[0047] The solution steps are as follows:

[0048] Initialization: Choose an initial solution R (0) ,X (0) ;

[0049] Iterative update: In each iteration, the gradient of each set of data solutions is calculated one by one, and the solution is updated according to the gradient;

[0050] Convergence judgment: Set a tolerance ∈. When the difference between the solutions of two consecutive iterations is less than ∈, the algorithm is considered to have converged.

[0051] Preferably, in S4, the deviation between the line parameters and the actual voltage in the low-voltage distribution network is calculated, and the deviation is used as the objective function, and the formula is as follows:

[0052] minu loss =min∑|u real -u pred | (12);

[0053] In the formula, u real is the actual voltage amplitude, u pred To predict the voltage amplitude.

[0054] Compared with the prior art, the present invention has the following advantages:

[0055] 1. This application characterizes the potential low-voltage distribution network topology structure based on tree diagram coding, directly uses the deviation between the predicted value and the measured value of the user node voltage as the optimization index, and indirectly evaluates the recognition effect of the low-voltage distribution network topology structure by the accuracy of the prediction of the user node voltage of the low-voltage distribution network. It can realize the complete topology recognition of the low-voltage distribution network with high accuracy.

[0056] 2. This application achieves complete topology identification of low-voltage distribution networks with high accuracy, effectively enriches the topology identification methods and tools of low-voltage distribution networks, and can greatly promote the implementation of low-voltage distribution network topology identification. Through testing on different low-voltage distribution networks, the test results show that the prediction deviation of user node voltage is very small, which indirectly illustrates the rationality and correctness of topology structure identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the description of the embodiments in conjunction with the following drawings, in which:

[0058] Figure 1 It is the overall flow chart of the present invention;

[0059] Figure 2 It is a schematic diagram of feasible results of a binary tree with four leaf nodes in the present invention;

[0060] Figure 3 It is a schematic diagram of node splitting in the present invention;

[0061] Figure 4 It is a tree diagram of the medium and low voltage distribution network of the present invention. DETAILED DESCRIPTION

[0062] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be understood as limiting the present invention.

[0063] like Figure 1-Figure 4 As shown, this embodiment proposes a method for identifying the topology of a low-voltage distribution network area based on tree diagram encoding representation, comprising the following steps:

[0064] S1: Based on Kirchhoff's current law, the user phase of the low-voltage distribution network is identified, and the three-phase topology identification is converted into single-phase identification;

[0065] S2: Represent the topology of the low voltage distribution network through a tree diagram, and perform encoding based on the represented tree diagram;

[0066] S3: Calculate the impedance matrix based on the gradient descent method;

[0067] S4: define the optimization index of user node voltage and prediction deviation;

[0068] S5: Use genetic algorithm to find the best solution in the potential topology space and output the optimal low-voltage distribution network topology.

[0069] In S1, for each power user on the same low-voltage distribution network, based on formula (1) as the optimization target, the user phase is identified through 0-1 planning, and the formula is as follows:

[0070] min Z=|I p,A -∑I p,A,user |+|I p,B -∑I p,B,user |+|I p,C -∑I p,C,user | (1);

[0071] In the formula, I p,A ,I p,B ,I p,C They represent the current values ​​of the transformer output phase A, phase B, and phase C in the same direction as the voltage after the current is decomposed according to the voltage direction. p,A,user ,I p,B,user ,I p,C,user Respectively represent the current values ​​in the same direction as the voltage after the users of phase A, phase B and phase C are decomposed according to the voltage direction;

[0072] 0-1 programming is an NP-hard problem. To simplify the difficulty of solving the problem, based on formula (2), the optimization problem of formula (1) is transformed into a least squares problem. The formula is as follows:

[0073]

[0074] In the formula, I p,user Represents the active current matrix of all users, and matrices A, B, and C represent the current I of each user respectively. p,user The proportion of each phase in the total current;

[0075] For the solution matrix [ABC] T , select the phase corresponding to the element closest to 1 in the column vector as the phase of the user. For the equation group of formula (2), the number of measurement time data must be at least equal to the number of power users to obtain a well-posed equation group.

[0076] After phase identification of the low-voltage distribution network, the three-phase topology identification is converted into single-phase identification, thereby reducing the complexity of the topology identification algorithm.

[0077] In S2, before encoding the low-voltage distribution network topology, it is necessary to fully represent the potential solution space corresponding to the low-voltage distribution network topology. When only considering the low-voltage distribution network part, the operation mode of the power grid can be regarded as a form in which the transformer provides electricity and the power users consume electricity. Since the low-voltage distribution network itself is a radial network, that is, a transformer drives the power supply of several power users.

[0078] Therefore, the topological structure of the low-voltage distribution network can be abstracted into a tree structure, with the transformer as the root node, the power user as the leaf node, the number of leaf nodes n, and each leaf node representing different information. A tree with a root node degree of 2 and an internal node degree of 3 is generated. The internal nodes are the nodes in the tree except the root node and the leaf nodes.

[0079] Taking the tree with n=2 as the starting tree, when n=2 there is only one possibility, that is, two leaf nodes are connected to the root node. The tree with n=3 is generated by adding leaf nodes to the tree with n=2. Since the degree of the root node is 2, the newly added leaf nodes cannot be connected to the root node. Therefore, every time a leaf node is added, an internal node is added synchronously.

[0080] The degree of the internal nodes of the tree with n=3 is 3. An internal node is added to the edge of the original tree, and the newly added leaf node is connected to the internal node, so that the degree of the internal node obtained in each operation is 3.

[0081] The original tree itself has two edges, and internal nodes can be added. Assume that the root node of the original tree is an internal node of the new tree, and the root node of the new tree is connected to the internal node to generate a virtual edge. The root node of the new tree is the virtual root node, and when the new leaf node is connected to the virtual root node, a virtual edge is generated.

[0082] At this time, there are three possibilities for adding new leaf nodes to the original tree, namely, connecting to the virtual root node, connecting to the internal node added between the original tree root node and leaf node No. 1, and connecting to the internal node added between the original tree root node and leaf node No. 2. The operation examples are as follows: Figure 2 As shown, Figure 2 It is a schematic diagram of feasible results of a binary tree with four leaf nodes in the present invention.

[0083] When a new leaf node is added, a new intermediate node is also added. The new intermediate node eliminates the original edge and adds 3 new edges. The number of positions that can be placed after adding the fourth leaf node is 5. And so on, the number of trees with n different leaf nodes is m, and the formula is as follows:

[0084]

[0085] Where m is the possible number of trees and n is the number of different leaf nodes.

[0086] In actual operation, low-voltage distribution networks are not all binary trees with a root node degree of 2 and an internal node degree of 3. When the number of leaf nodes n is known and the leaf nodes are different, the possible number of complete binary trees is far less than that of ordinary trees. In order to reduce the solution space of the heuristic algorithm, if the low-voltage distribution network can be considered as a binary tree, the solution time can be reduced and the solution accuracy can be increased.

[0087] In the low-voltage distribution network topology described by a tree, the edges of the tree are real transmission lines, and the values ​​of the edges are line impedances, such as Figure 3 As shown, Figure 3 It is a schematic diagram of node splitting in the present invention, which converts the original tree into a binary tree.

[0088] Since the operation of converting a tree into a binary tree is limited to describing the abstract circuit structure rather than changing the circuit in actual operation, the circuit information on nodes a, b, c, and d is the same, and the new structure also satisfies Kirchhoff's law.

[0089] Connect nodes b, c, d to node a. The circuit information on nodes a, b, c, d is the same, and the formula is as follows:

[0090]

[0091] In the formula, Represents the voltage of nodes a, b, and c respectively, Z ab , Z ac They represent the impedance values ​​of the line between the node ab and the node ac respectively;

[0092] Combining equation (4) and equation (5) we can get equation (6), which is as follows:

[0093]

[0094] After connecting nodes b and c to node e, and then connecting nodes e and d to node a, the circuit information on nodes a, b, c, and d is the same. Similarly, the formula can be obtained as follows:

[0095]

[0096] Equations (6) and (7) are valid at any time when the circuit is running. Since the line impedance Z is a constant and the node circuit information is the same, the solution is Z ac =Z ec , Z ab =Z eb Therefore, the voltage at point e is equal to the voltage at point a. Points with the same voltage in the circuit can be regarded as one point, that is, the converted binary tree has the same circuit meaning as the original topological structure, and all low-voltage distribution networks can be regarded as binary trees.

[0097] Since the operation of adding internal nodes does not change the circuit meaning of the topological structure, according to actual needs, in the tree diagram used to represent the topological structure of the low-voltage distribution network, the degree of the root node and the leaf node are both 1, the edge connected to the root node is the transformer outlet line, and the edge connected to the leaf node is the household power supply line. This behavior ensures that the possibility of the transformer connecting to a single feeder is not ignored.

[0098] At this time, the tree structure of leaf node n=2 is as follows Figure 4 As shown, Figure 4 It is a tree diagram of the low-voltage distribution network in the present invention. It can be seen that the root node and the leaf nodes of the tree have the same properties in the tree. However, in practice, since the transformer is the power supply end of the low-voltage distribution network, the node representing the transformer in the tree is defined as the root node. At the same time, adding new leaf nodes based on the tree does not require the introduction of the concepts of virtual edges and virtual root nodes. The possibility of a tree generated by adding new leaf nodes is equal to the number of edges in the original tree.

[0099] In the process of building the tree, there are 3 selection methods in the first step, 5 selection methods in the second step, and 2k-3 selection methods in the k-2th step. Then the encoding of the binary tree is composed of the mth selection used in the kth step.

[0100] After the encoding is completed, the binary tree is described and generated through the edges connecting the points. In the binary tree with n leaf nodes, number 0 is used as its root node, numbers 1 to n are used as its leaf nodes, and numbers n+1 to 2n-1 represent the internal nodes of the tree. Since the degree of each node of the tree is fixed, the number of internal nodes can be guaranteed to be fixed.

[0101] The initial tree structure is represented in the form of [(0, n+1), (n+1, 1), (n+1, 2)]. The operation of adding a new leaf node to the i-th edge is to delete the i-th edge in the structure and add the edge consisting of the starting node of the i-th edge and the new internal node, the edge consisting of the new internal node and the ending node of the i-th edge, and the edge consisting of the new internal node and the new leaf node in sequence, to ensure that the trees generated under the same operation are the same.

[0102] In S3, since the metering data at each moment corresponds to the same low-voltage distribution network topology, the circuit parameters are iteratively solved using multiple moment data to calculate ΔU(T), I p (T), I q (T), let A(T) = [I p (T) I q (T)], b(T) = ΔU(T), and we get equation (8), which is as follows:

[0103]

[0104] In the formula, I p (T), I q (T) represents the current value in the same direction as the voltage and perpendicular to the voltage after decomposition according to the voltage direction, ΔU(T) represents the voltage difference between the user and the transformer, R and X represent the line impedance;

[0105] The solution of the linear equation system of formula (8) is transformed into an optimization problem as shown in formula (9), and the objective function F(x) is defined as follows:

[0106]

[0107] Iterative calculation is performed based on the gradient descent method. The formula is as follows:

[0108]

[0109] Among them, α is the learning rate, is the objective function F(R,X) in R (t) ,X (t) The gradient at , t is the number of iterations;

[0110] The gradient is calculated as follows:

[0111]

[0112] In the formula, A(T)=[I p (T) I q (T)],b(T)=ΔU(T),I p (T), I q (T) represents the current value in the same direction as the voltage and the current value perpendicular to the voltage after decomposition according to the voltage direction at time T, ΔU(T) represents the voltage difference between the user and the transformer at time T, R and X represent the line impedance, and t is the number of iterations;

[0113] The solution steps are as follows:

[0114] Initialization: Choose an initial solution R (0) ,X (0) ;

[0115] Iterative update: In each iteration, the gradient of each set of data solutions is calculated one by one, and the solution is updated according to the gradient;

[0116] Convergence judgment: Set a tolerance ∈. When the difference between the solutions of two consecutive iterations is less than ∈, the algorithm is considered to have converged.

[0117] In S4, the deviation between the line parameters and the actual voltage in the low-voltage distribution network is calculated, and the deviation is used as the objective function. The formula is as follows:

[0118] min u loss =minΣ|u real -u pred | (12);

[0119] In the formula, u real is the actual voltage amplitude, u pred To predict the voltage amplitude.

[0120] In S5, the genetic algorithm is a heuristic search algorithm that simulates the principles of natural selection and genetics. It solves optimization problems by simulating the inheritance, crossover and mutation mechanisms of the biological evolution process. By using the genetic algorithm to search the topological structure space of the low-voltage distribution network, the search effect can be better improved.

[0121] The potential topology of the low-voltage distribution network is characterized based on tree diagram coding. The deviation between the predicted value and the measured value of the user node voltage is directly used as the optimization index. The accuracy of the prediction of the user node voltage of the low-voltage distribution network is used to indirectly evaluate the recognition effect of the low-voltage distribution network topology. The complete topology recognition of the low-voltage distribution network can be achieved with high accuracy.

[0122] The above method effectively enriches the topology identification methods and tools of low-voltage distribution networks, and can greatly promote the implementation of low-voltage distribution network topology identification. Tests are carried out on different low-voltage distribution networks. The test results show that the prediction deviation of user node voltage is very small, which indirectly illustrates the rationality and correctness of topology structure identification.

[0123] Although the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the claims and their equivalents.

Claims

1. A method for identifying low-voltage distribution network topology based on tree diagram coding representation, characterized in that: The following steps are involved: S1: Based on Kirchhoff's current law, the user phase of the low-voltage distribution network is identified, and the three-phase topology identification is converted into single-phase identification; S2: Represent the topology of the low voltage distribution network through a tree diagram, and perform encoding based on the represented tree diagram; S3: Calculate the impedance matrix based on the gradient descent method; S4: define the optimization index of user node voltage and prediction deviation; S5: Use genetic algorithm to find the best solution in the potential topology space and output the optimal low-voltage distribution network topology.

2. The method for identifying low-voltage distribution network topology based on tree diagram coding representation according to claim 1 is characterized in that: In S1, for the power users on each phase of the same low-voltage distribution network, based on formula (1) as the optimization target, the user phase is identified through 0-1 planning, and the formula is as follows: min Z=|I p,A -∑|I p,A,usr |+|And p,B -∑I p,B,user |+|And p,C -∑I p,C,user | (1); In the formula, I p,A ,I p,B ,I p,C They represent the current values ​​of the transformer output phase A, phase B, and phase C in the same direction as the voltage after the current is decomposed according to the voltage direction. p,A,user ,I p,B,user ,I p,C,user Respectively represent the current values ​​in the same direction as the voltage after the users of phase A, phase B and phase C are decomposed according to the voltage direction; Based on formula (2), the optimization problem of formula (1) is transformed into a least squares problem, and the formula is as follows: In the formula, I p,user Represents the active current matrix of all users, and matrices A, B, and C represent the current I of each user respectively. p,user The proportion of each phase in the total current; After phase identification of the low-voltage distribution network, the three-phase topology identification is converted into single-phase identification.

3. The method for identifying low-voltage distribution network topology based on tree diagram coding representation according to claim 1 is characterized in that: In S2, the topological structure of the low-voltage distribution network is abstracted into a tree structure, with the transformer as the root node and the power user as the leaf node. The number of leaf nodes is n, and each leaf node represents different information. A tree with a root node degree of 2 and an internal node degree of 3 is generated. The internal nodes are other nodes in the tree except the root node and the leaf nodes.

4. The method for identifying low-voltage distribution network topology based on tree diagram coding representation according to claim 3 is characterized in that: Taking the tree with n=2 as the starting tree, the tree with n=3 is generated by adding leaf nodes to the tree with n=2. Every time a leaf node is added, an internal node is added simultaneously. The degree of the internal nodes of the tree with n=3 is 3. An internal node is added to the edge of the original tree and the newly added leaf node is connected to the internal node.

5. The method for identifying low-voltage distribution network topology based on tree diagram coding representation according to claim 4 is characterized in that: Assume that the root node of the original tree is an internal node of the new tree, and the root node of the new tree is connected to the internal node to generate a virtual edge. The root node of the new tree is a virtual root node, and a virtual edge is generated when the new leaf node is connected to the virtual root node; When a new leaf node is added, a new intermediate node is also added. The new intermediate node eliminates the original edge and adds 3 new edges. The number of positions that can be placed after adding the fourth leaf node is 5. And so on, the number of trees with n different leaf nodes is m, and the formula is as follows: Where m is the possible number of trees and n is the number of different leaf nodes.

6. The method for identifying low-voltage distribution network topology based on tree diagram coding representation according to claim 5 is characterized in that: The edges of the tree are real power transmission lines, and the edge values ​​are line impedances, converting the original tree into a binary tree; Connect nodes b, c, d to node a. The circuit information on nodes a, b, c, d is the same, and the formula is as follows: In the formula, Represents the voltage of nodes a, b, and c respectively, Z ab , Z ac They represent the impedance values ​​of the line between the node ab and the node ac respectively; Combining equation (4) and equation (5) we can get equation (6), which is as follows: After connecting nodes b and c to node e, and then connecting nodes e and d to node a, the circuit information on nodes a, b, c, and d is the same. Similarly, the formula can be obtained as follows: Equations (6) and (7) are valid at any time when the circuit is running, and the solution is Z ac =Z ec , Z ab =Z eb , so the voltage at point e is equal to the voltage at point a; The degree of the root node and the leaf node are both 1. The edge connected to the root node is the transformer outgoing line, and the edge connected to the leaf node is the household power supply line. The node representing the transformer in the tree is defined as the root node. The possibility of adding a new leaf node to generate a tree is equal to the number of edges in the original tree.

7. The method for identifying low-voltage distribution network topology based on tree diagram coding representation according to claim 6 is characterized in that: After the encoding is completed, the binary tree is described and generated by the edges connecting the points. In the binary tree with n leaf nodes, number 0 is used as its root node, numbers 1 to n are used as its leaf nodes, and numbers n+1 to 2n-1 represent the internal nodes of the tree. The initial tree structure is represented in the form of [(0, n+1), (n+1, 1), (n+1, 2)]. The i-th edge in the structure is deleted, and the edge composed of the starting node of the i-th edge and the new internal node, the edge composed of the new internal node and the ending node of the i-th edge, and the edge composed of the new internal node and the new leaf node are added in sequence.

8. The method for identifying low-voltage distribution network topology based on tree diagram coding representation according to claim 1 is characterized in that: In S3, multiple time data are used to iteratively solve the circuit parameters, and ΔU(T), I p (T), I q (T), let A(T) = [I p (T) I q (T)], b(T) = ΔU(T), and we get equation (8), which is as follows: In the formula, I p (T), I q (T) represents the current value in the same direction as the voltage and perpendicular to the voltage after decomposition according to the voltage direction, ΔU(T) represents the voltage difference between the user and the transformer, R and X represent the line impedance; The solution of the linear equation system of formula (8) is transformed into an optimization problem as shown in formula (9), and the objective function F(x) is defined as follows: Iterative calculation is performed based on the gradient descent method. The formula is as follows: Among them, α is the learning rate, is the objective function F(R,X) in R (t) ,X (t) The gradient at , t is the number of iterations.

9. The method for identifying low-voltage distribution network topology based on tree diagram coding representation according to claim 8 is characterized in that: The gradient is calculated as follows: In the formula, A(T)=[I p (T) I q (T)],b(T)=ΔU(T),I p (T), I q (T) represents the current value in the same direction as the voltage and the current value perpendicular to the voltage after decomposition according to the voltage direction at time T, ΔU(T) represents the voltage difference between the user and the transformer at time T, R and X represent the line impedance, and t is the number of iterations; The solution steps are as follows: Initialization: Choose an initial solution R (0) ,X (0) ; Iterative update: In each iteration, the gradient of each set of data solutions is calculated one by one, and the solution is updated according to the gradient; Convergence judgment: Set a tolerance ∈. When the difference between the solutions of two consecutive iterations is less than ∈, the algorithm is considered to have converged.

10. The method for identifying low-voltage distribution network topology based on tree diagram coding representation according to claim 1, characterized in that: In S4, the deviation between the line parameters and the actual voltage in the low-voltage distribution network is calculated, and the deviation is used as the objective function. The formula is as follows: minutes in loss =min∑|u real -in pred | (12); In the formula, u real is the actual voltage amplitude, u pred To predict the voltage amplitude.

Images