A power grid partition equivalence method and device considering community structure and a medium

Abstract

The application discloses a power grid partition equivalence method and device considering community structure and a medium, and belongs to the technical field of long-distance power transmission. The method comprises the following steps: obtaining power grid data, and preprocessing the power grid data; for the preprocessed power grid data, first nodes related to electromagnetic loop networks are reserved, and radiating second nodes are merged into communities through an electromagnetic loop network automatic identification algorithm based on a power grid topology diagram; in view of the requirement of retaining the structure characteristics of the power grid, a spatial channel constraint and a connectivity constraint criterion of node partition are designed, and the Louvain algorithm process and the criterion are improved; after the improved Louvain algorithm is used to divide the community nodes, the first nodes and the second nodes are added to the communities to which the first nodes and the second nodes belong. The application effectively reduces the node scale that a subsequent community partition algorithm should deal with, and retains the topology structure characteristics of the power grid.

Description

Technical Field

[0001] This invention relates to the field of long-distance power transmission technology, and in particular to a method, apparatus and medium for power grid partitioning equivalence considering community structure. Background Technology

[0002] Long-distance inter-provincial power transmission is an important means for my country to address the geographical differences in power generation resources and load demand. With the western region increasing its efforts in developing new energy power generation, the planning and research of inter-regional long-distance power transmission networks are becoming increasingly important. Conducting network planning for long-term inter-provincial transmission of new energy power requires both simulation studies of the security risks of the sending and receiving end power grids and a careful consideration of local issues with weak relevance to inter-provincial transmission. It also necessitates the development of targeted power grid simplification methods.

[0003] Power grid equivalence simplification is an important method for studying the operating characteristics of large power grids. Traditionally, power grid equivalence typically involves retaining all nodes within the study area and directly treating the power grid outside the study area using a finite number of generating units and nodes for equivalent processing. This equivalence method cannot meet the equivalence modeling research requirements for the aforementioned inter-provincial transmission network planning. Summary of the Invention

[0004] In order to at least partially solve one of the technical problems existing in the prior art, the present invention aims to provide a method, apparatus and medium for power grid zoning equivalence that takes into account community structure.

[0005] The technical solution adopted in this invention is:

[0006] An equivalence method for power grid zoning considering community structure includes the following steps:

[0007] Acquire power grid data and preprocess the power grid data;

[0008] For the preprocessed power grid data, an automatic electromagnetic ring network identification algorithm based on the power grid topology is used to retain the first node related to the electromagnetic ring network and merge the radial second nodes into communities.

[0009] To address the requirement of preserving power grid structural features, spatial channel constraints and connectivity constraints for node partitioning are designed, and the Louvain algorithm for community identification process and criteria are improved.

[0010] After partitioning the first node using the improved Louvain algorithm, the second node is added to the community to which the first node belongs, forming the final community partitioning result.

[0011] Among them, an electromagnetic ring network refers to two sets of lines with different voltage levels operating in parallel through the connection of the magnetic circuits of transformers at both ends.

[0012] Furthermore, the preprocessing of the power grid data includes:

[0013] Identify the set of nodes (busbars) within the same substation according to the preset node naming rules;

[0014] If the impedance of the connecting branches between nodes is less than a preset value, identify the nodes (busbars) operating in parallel within the substation and merge these nodes.

[0015] After preprocessing, the topological relationships between nodes are represented as a graph. It is a node combination, where ε is the set of transmission lines connecting the nodes, called the edge set; the edge information is stored using {{node number and name, voltage level, and station number of one end}, {node number and name, voltage level, and station number of the other end}}.

[0016] Furthermore, the automatic electromagnetic ring network identification algorithm obtains the electromagnetic ring network subgraph in the following manner:

[0017] Let the set of nodes for high voltage levels be {V}. h The corresponding voltage level is u. h The set of nodes for low voltage levels is {V}. l The corresponding voltage level is represented as u. l ; Let {V} be a subset of the V subset. h} and {V l The complete power grid topology diagram of node G is shown in Figure G. hl ;

[0018] The automatic identification steps for electromagnetic ring networks are as follows:

[0019] A1. Delete image G hl All dangling nodes in the array; a dangling node is a node that has one and only one edge connected to it.

[0020] A2, Delete image G hl All {V} h} node; delete {V h After}, {V l The network composed of nodes will be divided into several non-connected subgraphs, denoted as .

[0021] A3, Delete image G hl All {V} l} nodes, resulting in a connected subgraph containing only high-voltage nodes, called G. h ;

[0022] A4. For each subgraph In Figure G hlSearch for the connection edges between the subgraph and nodes outside the subgraph; all nodes on the opposite side of the edge belong to {V}. h The nodes in the subgraph are called the high-voltage power supply nodes of that subgraph.

[0023] A5, Subgraph If the number of high-voltage power supply nodes is 1, then the subgraph node and its high-voltage power supply node are assigned to the same community, and the subsequent search of the subgraph is skipped.

[0024] A6. Subgraph If all high-voltage power supply nodes are connected to only one low-voltage node, then skip the subsequent search of this subgraph;

[0025] A7, Subgraph If the high-voltage power supply node does not meet the conditions in steps A5 and A6, then in the connected subgraph G... h Search subgraph Find the shortest path between all high-voltage power supply nodes, and add the found nodes and edges to the subgraph. The resulting new subgraph is called This is a sub-graph of an electromagnetic ring network.

[0026] Furthermore, low-voltage level nodes are grouped into communities in the following way:

[0027] B1. Identify electromagnetic loops in the network using an automatic electromagnetic loop identification algorithm;

[0028] B2, Retaining {V} h} Nodes form an electromagnetic ring network structure {V l}node;

[0029] B3. Using the cut point method, transform the original network topology G... hl Electromagnetic ring network {V l Other {V} nodes l The nodes are identified and grouped, and then merged with the corresponding cut points into a community.

[0030] Furthermore, the execution steps of the cut point method are as follows:

[0031] C1. Identify the original topology network G hl The cut points in G, and G hl All cut points are stored in set Ω1;

[0032] C2, for {V} in set Ω1 h The nodes are traversed sequentially, and each traversal is within the original network topology G. hl Execute on: For node V in set Ω1 h,k , will node V h,k From the original network topology Ghl Remove, and obtain a new network topology graph G. hl,k ;Statistical G hl,k The number of connected subgraphs, N; if N = 1, skip node V. h,k The loop; if N > 1, and the connected subgraph formed by the separated nodes is entirely composed of {V l If the nodes are composed of}, they are stored in the set Γ. k In the middle, Γ k Nodes and node V in h,k Belong to the same community; otherwise skip node V. h,k The loop continues until all nodes in set Ω1 have been traversed.

[0033] Furthermore, the improved Louvain algorithm works as follows:

[0034] D1. Obtain the power grid topology model;

[0035] D2. Initialization: Treat each node in the power grid topology model as an independent community;

[0036] D3. Clustering optimization: For each node, try to move it to the community of its neighboring nodes;

[0037] D4. Determine whether the spatial channel constraints are met. If they are met, proceed to step D5; otherwise, proceed to step D6.

[0038] D5. Determine whether the connectivity constraint is satisfied. If satisfied, proceed to step D7; otherwise, proceed to step D6.

[0039] D6. Set the constraint-based correction term to a negative value with a large absolute value;

[0040] D7. Calculate the improved modularity gain;

[0041] D8. Determine if there is a move that improves the module level. If there is an improvement, return to step D7; if there is no improvement, proceed to step D9.

[0042] D9. Community Aggregation: Using the current community as a super node to build a new network;

[0043] D10. On the new network, determine whether there are still moves that improve module degree. If there are, return to step D7; if not, proceed to step D11.

[0044] D11. Output the final community division results.

[0045] Furthermore, the spatial channel constraint detection detects the spatial distance l between any two nodes i and j. i-jDoes it exceed the threshold L? The formula for calculating the node distance threshold L is as follows:

[0046]

[0047] In the formula, L represents the node distance threshold, l ew and l sn These represent the total east-west span and total north-south span of the power grid coverage area (unit: km); r is the set scaling factor;

[0048] Based on the detection results of spatial channel constraints, the spatial correction term for modularity is calculated, as shown in the following expression:

[0049]

[0050] K = 10 n ,n∈Z +

[0051] In the formula, P L-(i,j) This is a spatial correction term for modularity; i-j Let K be the spatial distance between node i and node j, where K is a positive value with a large absolute value, and n is a preset value.

[0052] The connectivity constraint detection node i originates from the original community c. i Assigned to the community c where node j belongs j Afterwards, will the two clubs lose their connection?

[0053] In each iteration, after node i is removed from one community and then added to another, the connectivity between the two communities before and after node i's move in the original topology is ensured. Based on the connectivity constraint detection results, the connectivity correction term for modularity is calculated, as shown in the following expression:

[0054]

[0055] K = 10 n ,n∈Z +

[0056] In the formula, β i-j This is a spatial correction term for modularity.

[0057] Furthermore, the formula for calculating the improved modularity gain is as follows:

[0058] ΔQ i-j * =ΔQ i-j +P L-(i,j) +β i-j

[0059] In the formula, ΔQ i-j *P represents the improved modularity gain. L-(i,j) β is the spatial correction term for modularity. i-j For connectivity correction terms related to modularity; ΔQ i-j This is the module gain in Louvain's algorithm.

[0060] Another technical solution adopted in this invention is:

[0061] A grid zoning equivalence device considering community structure, comprising:

[0062] At least one processor;

[0063] At least one memory for storing at least one program;

[0064] When the at least one program is executed by the at least one processor, the at least one processor performs the method as described above.

[0065] Another technical solution adopted in this invention is:

[0066] A computer-readable storage medium storing a processor-executable program, which, when executed by a processor, performs the method described above.

[0067] The beneficial effects of this invention are: this invention merges a large number of radial low-voltage level nodes into the community where the corresponding high-voltage level nodes are located, while preserving the independence of low-voltage level nodes and branches in the electromagnetic ring network; it effectively reduces the node scale that the subsequent community partitioning algorithm has to deal with, and completely preserves the characteristics of the power grid topology. Attached Figure Description

[0068] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following description is provided with accompanying drawings of the relevant technical solutions in the embodiments of the present invention or the prior art. It should be understood that the accompanying drawings described below are only for the purpose of clearly illustrating some embodiments of the technical solutions of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0069] Figure 1 This is a flowchart of the automatic electromagnetic ring network identification algorithm in an embodiment of the present invention;

[0070] Figure 2 This is a schematic diagram of a loop structure that does not meet the electromagnetic loop network conditions in an embodiment of the present invention;

[0071] Figure 3 This is a flowchart of the community merging algorithm for low-voltage level nodes in this embodiment of the invention;

[0072] Figure 4 This is an improved Louvain algorithm flow and a power grid node partitioning flowchart in the embodiments of the present invention;

[0073] Figure 5 This is a local power grid topology connection diagram of a certain area in an embodiment of the present invention;

[0074] Figure 6 This is a topology diagram of a local power grid in a certain area after removing suspended nodes in an embodiment of the present invention;

[0075] Figure 7 This is a schematic diagram illustrating the deletion of all 500kV nodes in the topology in an embodiment of the present invention;

[0076] Figure 8 This is a schematic diagram of supplementing adjacent 500kV nodes in an embodiment of the present invention;

[0077] Figure 9 This is a schematic diagram of the connection nodes between 500kV nodes in an embodiment of the present invention;

[0078] Figure 10 This is a diagram showing the effectiveness test results of the electromagnetic loop network of a local power grid in a certain area according to an embodiment of the present invention;

[0079] Figure 11 This is a diagram of the local power grid structure where the remaining 220kV nodes are located in an embodiment of the present invention;

[0080] Figure 12 This is the connected subgraph after removing the YC node in this embodiment of the invention;

[0081] Figure 13 This is the connected subgraph after removing the WL node in this embodiment of the invention;

[0082] Figure 14 This is the connected subgraph after removing the LD node in this embodiment of the invention;

[0083] Figure 15 This is the topology of the input nodes of the partitioning algorithm in this embodiment of the invention;

[0084] Figure 16 This is a diagram showing the results of local power grid community division in an embodiment of the present invention (considering spatial channel constraints and connectivity constraints);

[0085] Figure 17 This is a diagram showing the final result of the local power grid community division in an embodiment of the present invention;

[0086] Figure 18 This is a diagram of the equivalent network structure of a local power grid in an embodiment of the present invention;

[0087] Figure 19This is a flowchart illustrating the steps of an equivalent method for power grid partitioning that considers community structure in an embodiment of the present invention. Detailed Implementation

[0088] The embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention. The step numbers in the following embodiments are set only for ease of explanation, and there is no limitation on the order between the steps. The execution order of each step in the embodiments can be adaptively adjusted according to the understanding of those skilled in the art.

[0089] In the description of this invention, it should be understood that the orientation descriptions, such as up, down, front, back, left, right, etc., are based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limiting this invention.

[0090] In the description of this invention, "several" means one or more, "more than" means two or more, "greater than," "less than," and "exceeding" are understood to exclude the stated number, while "above," "below," and "within" are understood to include the stated number. The use of "first" and "second" in the description is merely for distinguishing technical features and should not be construed as indicating or implying relative importance, or implicitly indicating the number of indicated technical features, or implicitly indicating the order of the indicated technical features.

[0091] Furthermore, in the description of this invention, unless otherwise stated, "multiple" means two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. The character " / " generally indicates that the preceding and following related objects have an "or" relationship.

[0092] In the description of this invention, unless otherwise explicitly defined, terms such as "setting," "installing," and "connecting" should be interpreted broadly, and those skilled in the art can reasonably determine the specific meaning of the above terms in this invention in conjunction with the specific content of the technical solution.

[0093] To address existing technical problems, this invention proposes an equivalence method that comprehensively preserves the structural characteristics and system-level security risks of large power grids by grouping and simplifying nodes in all regions of the network according to their communities. The proposed method employs a rapid approach to reduce network node size. An automatic electromagnetic ring network identification algorithm for the power grid topology is used to group radial network nodes into communities, thereby reducing the node size input to the community identification algorithm. Improved criteria and procedures for the community identification algorithm are designed, aiming to preserve the structural characteristics of the power grid. Spatial and connectivity constraints for node partitioning are designed, and the Louvain algorithm for community identification is improved, resulting in a node community partitioning and equivalence method suitable for power grid equivalence.

[0094] like Figure 19 As shown, this embodiment provides a power grid zoning equivalence method considering community structure, including the following steps:

[0095] S1. Obtain power grid data and preprocess the power grid data.

[0096] This embodiment performs equivalent modeling with structural preservation for high-voltage transmission networks, considering voltage levels of 220V and above. Preprocessing includes:

[0097] 1) Identify the set of nodes (busbars) within the same substation according to the node naming rules;

[0098] 2) Identify the nodes (busbars) operating in parallel within the substation based on the fact that the impedance of the connecting branches between nodes is less than a set small positive number, and then merge them.

[0099] After preprocessing, the topological relationships between nodes are represented as a graph. It is a node combination. ε is the set of transmission lines connecting the nodes, called the edge set. Edge information is stored using {{node number and name, voltage level, and site number of one end}, {node number and name, voltage level, and site number of the other end}}.

[0100] S2. For the preprocessed power grid data, the first node related to the electromagnetic ring network is retained and the radial second node is merged into a community using an automatic electromagnetic ring network identification algorithm based on the power grid topology.

[0101] (1) Automatic identification algorithm for electromagnetic ring networks

[0102] An electromagnetic loop network refers to two sets of lines with different voltage levels connected in parallel through the magnetic circuits of transformers at both ends. The automatic search algorithm for electromagnetic loop networks in this invention is described below.

[0103] Let the set of nodes for high voltage levels be {V}. h The corresponding voltage level is u.h The set of nodes for low voltage levels is {V}. l The corresponding voltage level is represented as u. l Let {V} be a subset of {V}. h} and {V l The complete power grid topology diagram of node G is shown in Figure G. hl The automatic search process for electromagnetic ring networks is as follows: Figure 1 As shown:

[0104] The automatic identification steps for electromagnetic ring networks are as follows:

[0105] A1. Delete image G hl All dangling nodes in the array; a dangling node is a node that has one and only one edge connected to it.

[0106] A2, Delete image G hl All {V} h} node; delete {V h After}, {V l The network composed of nodes will be divided into several non-connected subgraphs, denoted as .

[0107] A3, Delete image G hl All {V} l} nodes, resulting in a connected subgraph containing only high-voltage nodes, called G. h ;

[0108] A4. For each subgraph In Figure G hl Search for the connection edges between the subgraph and nodes outside the subgraph; all nodes on the opposite side of the edge belong to {V}. h The nodes in the subgraph are called the high-voltage power supply nodes of that subgraph.

[0109] A5, Subgraph If the number of high-voltage power supply nodes is 1, then the subgraph node and its high-voltage power supply node are assigned to the same community, and the subsequent search of the subgraph is skipped.

[0110] A6. Subgraph All high-voltage power supply nodes are connected to only one low-voltage node, such as Figure 2 As shown, the subsequent search of that subgraph is skipped;

[0111] A7, Subgraph If the high-voltage power supply node does not meet the conditions in steps A5 and A6, then in the connected subgraph G... h Search subgraph Find the shortest path between all high-voltage power supply nodes, and add the found nodes and edges to the subgraph. The resulting new subgraph is called This is a sub-graph of an electromagnetic ring network.

[0112] (2) Community merging method for low voltage level nodes

[0113] The process for merging communities of low-voltage level nodes is as follows: Figure 3 As shown, the community merging steps for low-voltage level nodes are as follows:

[0114] B1. Identify electromagnetic loops in the network using an automatic electromagnetic loop identification algorithm;

[0115] B2, Retaining {V} h} Nodes form an electromagnetic ring network structure {V l}node;

[0116] B3. Using the cut point method, transform the original network topology G... hl Electromagnetic ring network {V l Other {V} nodes l The nodes are identified and grouped, and then merged with the corresponding cut points into a community.

[0117] In graph theory, an articulation point is a node that divides a graph into two or more disjoint subgraphs. The steps of the articulation method are as follows:

[0118] C1. Identify the original topology network G hl The cut points in G, and G hl All cut points are stored in set Ω1;

[0119] C2, for {V} in set Ω1 h The nodes are traversed sequentially, and each traversal is within the original network topology G. hl Execute on: For node V in set Ω1 h,k , will node V h,k From the original network topology G hl Remove, and obtain a new network topology graph G. hl,k ;Statistical G hl,k The number of connected subgraphs, N; if N = 1, skip node V. h,k The loop; if N > 1, and the connected subgraph formed by the separated nodes is entirely composed of {V l If the nodes are composed of}, they are stored in the set Γ. k In the middle, Γ k Nodes and node V in h,k Belong to the same community; otherwise skip node V. h,k The loop continues until all nodes in set Ω1 have been traversed.

[0120] S3. To meet the requirement of preserving the structural features of the power grid, spatial channel constraints and connectivity constraints for node partitioning are designed, and the Louvain algorithm process and criteria for community identification are improved.

[0121] S4. After dividing the community nodes using the improved Louvain algorithm, the first node and the second node are added to their respective communities.

[0122] This embodiment uses a community partitioning approach to divide all network nodes into community partitions. By merging nodes within the same partition into an equivalent node and retaining the transmission lines between communities as equivalent edges, the network is simplified and made equivalent.

[0123] The improved Louvain algorithm performs node partitioning as follows: Figure 4 As shown in the figure. The dashed box represents the classic Louvain algorithm flow. The bold black box contains the newly added constraint criteria and flow of this patent. The improved Louvain algorithm works as follows:

[0124] D1. Obtain the power grid topology model;

[0125] D2. Initialization: Treat each node in the power grid topology model as an independent community;

[0126] D3. Clustering optimization: For each node, try to move it to the community of its neighboring nodes;

[0127] D4. Determine whether the spatial channel constraints are met. If they are met, proceed to step D5; otherwise, proceed to step D6.

[0128] D5. Determine whether the connectivity constraint is satisfied. If satisfied, proceed to step D7; otherwise, proceed to step D6.

[0129] D6. Set the constraint-based correction term to a negative value with a large absolute value;

[0130] D7. Calculate the improved modularity gain;

[0131] D8. Determine if there is a move that improves the module level. If there is an improvement, return to step D7; if there is no improvement, proceed to step D9.

[0132] D9. Community Aggregation: Using the current community as a super node to build a new network;

[0133] D10. On the new network, determine whether there are still moves that improve module degree. If there are, return to step D7; if not, proceed to step D11.

[0134] D11. Output the final community division results.

[0135] (1) Spatial channel constraints

[0136] Spatial channel constraint detection of the spatial distance l between any two nodes i and j i-j Does it exceed the threshold L? The formula for calculating the node distance threshold L is as follows:

[0137]

[0138] In the formula, L represents the node distance threshold, l ew and l sn These represent the total east-west span and total north-south span of the power grid coverage area (unit: km); r is the set scaling factor.

[0139] Based on the detection results of spatial channel constraints, the spatial correction term for modularity is calculated, as shown in the following expression:

[0140]

[0141] K = 10 n ,n∈Z +

[0142] Among them, P L-(i,j) This is a spatial correction term for modularity. i-j Let n be the spatial distance between node i and node j, and K be a positive value with a large absolute value. The appropriate value of n should be selected according to the actual situation.

[0143] (2) Connectivity constraints

[0144] Connectivity constraint detection node i from the original community c i Assigned to the community c where node j belongs j Will the two organizations lose their connection afterward?

[0145] In each iteration, after node i is removed from one community and then added to another, the connectivity between the two communities before and after the move is ensured in the original topology. Based on the connectivity constraint detection results, the connectivity correction term for modularity is calculated, as shown in the following expression:

[0146]

[0147] K = 10 n ,n∈Z +

[0148] Where, β i-j This is a spatial correction term for modularity. K is a positive value with a large absolute value, and n is chosen appropriately based on the actual situation.

[0149] (3) Improved formula for calculating modularity gain

[0150] The two new constraints mentioned above are reflected in the calculation of modularity gain. The improved formula for calculating modularity gain is as follows:

[0151] ΔQ i-j * =ΔQ i-j +P L-(i,j) +β i-j

[0152] Where, ΔQ i-j * P represents the improved modularity gain. L-(i,j) β is the spatial correction term for modularity. i-j This is a connectivity correction term for modularity. ΔQ i-j This is the modularity gain in the classic Louvain algorithm. It represents moving node i from its community c. i Move to the community c where node j is located. j The obtained modularity gain is calculated as follows:

[0153]

[0154]

[0155]

[0156]

[0157] Where A is a weighted adjacency matrix, A ij The weight of the connection between node i and node j is represented by s. i and s j are the sum of the degree weights of nodes i and j, respectively, representing the total weight of their connections to other nodes; w is the sum of the weights of all edges in the network; c i and node c j δ(c) represents the communities to which node i and node j belong, respectively; i ,c j ) is a function used to determine whether two nodes belong to the same module.

[0158] The above method will be explained in detail below with reference to the accompanying drawings and specific examples.

[0159] (I) Introduction to the Calculation Examples

[0160] This example is based on a simplified construction of a regional power grid consisting of actual 500kV and 220kV nodes. It aims to illustrate the application of the community merging method for low-voltage level nodes and the improved Louvain algorithm proposed in this invention for node partitioning.

[0161] The node and edge information for the example is shown in Tables 1 and 2.

[0162] Table 1. Node Information for the Case Study

[0163] 1 YD 500kV 2 YC 500kV 3 LD 500kV 4 WL 500kV 5 QN 220kV 6 ZS 220kV 7 YCH 220kV 8 JY 220kV 9 HP 220kV 10 LT 220kV 11 WT 220kV 12 ZH 220kV 13 YH 220kV

[0164] Table 2. Transmission line information for the example.

[0165] [YD]-[YC] 0.0049 500kV-500kV [WL]-[YC] 0.0064 500kV-500kV [WL]-[LD] 0.0130 500kV-500kV [LD]-[LT] 0.0004 500kV-220kV [HP]-[LT] 0.0011 220kV-220kV [WT]-[HP] 0.0051 220kV-220kV [YC]-[YH] 0.0052 500kV-220kV [YC]-[ZH] 0.0052 500kV-220kV [YH]-[ZH] 0.0037 220kV-220kV [YD]-[QN] 0.0005 500kV-220kV [QN]-[ZS] 0.0026 220kV-220kV [ZS]-[YCH] 0.0166 220kV-220kV [YCH]-[JY] 0.0157 220kV-220kV [JY]-[WL] 0.0162 500kV-220kV

[0166] After node preprocessing, the topological connections of a certain regional power grid are as follows: Figure 5 As shown in the diagram. Solid gray lines represent 500kV lines, dashed gray lines represent 220kV lines, and dotted gray lines represent lines bridging 500kV and 220kV.

[0167] (ii) Community merging method for low-voltage level nodes

[0168] After power grid node preprocessing, Figure 5 This includes power plants and substations in the region with voltage levels of 500kV and 220kV. First, taking the power grid topology of this region as an example, a quick search is performed on the 500-220kV electromagnetic ring network, corresponding to... Figure 1 Then G hl This is a topology diagram showing the nodes of 500kV and 220kV power plants. h For a voltage level of 500kV, V l It is a 220kV voltage level.

[0169] To search for electromagnetic loop networks, first delete suspended nodes from the power grid topology diagram of that area. Figure 4 After necessary simplification, the result is as follows: Figure 6 As shown.

[0170] G hl For the topology diagram of 500kV and 220kV substation nodes, delete G. hl After all 500kV nodes are connected, G hl It is divided into several non-connected subgraphs. like Figure 7 As shown. Among them, Figure 7 (a) is a subgraph after deleting all 500kV nodes. Figure 7 (b) is another subgraph after deleting all 500kV nodes.

[0171] search Find the adjacent nodes and draw a subgraph containing the adjacent nodes, such as... Figure 8 As shown. Among them, Figure 8 (a) is Figure 7(a) The corresponding subgraph with adjacent nodes. Figure 8 (b) is Figure 7 (b) The corresponding subgraph with adjacent nodes.

[0172] Add connection nodes between 500kV nodes to the subgraph. This forms several electromagnetic ring network subgraphs, denoted as... like Figure 9 As shown. Among them, Figure 9 (a) is Figure 8 (a) Sub-diagram following the supplementary 500kV connection node. Figure 9 (b) is Figure 8 (b) Sub-diagram following the supplementary 500kV connection node.

[0173] The validity of some subgraphs of the electromagnetic ring network is checked and merged. In this example, the number of 500kV nodes among adjacent nodes is determined, and subgraphs with more than 2 500kV nodes are retained. Then, electromagnetic ring network subsets containing the same set of 500kV nodes are merged, resulting in the following: Figure 10 The electromagnetic loop network results are shown in the figure. The solid gray lines represent lines with a voltage level of 500kV, the dashed gray lines represent lines with a voltage level of 220kV, and the dotted gray lines represent lines that are connected between 500kV and 220kV.

[0174] Finish Figure 5 After identifying the electromagnetic loop network in the topology diagram, the 220kV nodes that form an electromagnetic loop network with the 500kV nodes are retained. The local power grid structure of the remaining 220kV nodes is as follows. Figure 11 As shown. Among them, Figure 11 (a) shows the local power grid structure where the remaining 220kV nodes are located. Figure 11 (b) is another local power grid structure where the remaining 220kV nodes are located.

[0175] Using the cut point method, the five nodes LT, HP, WT, YH, and ZH are identified and grouped:

[0176] The original network topology has six cut points: YC, WL, LD, LT, HP, and WT.

[0177] Ω1={YC, WL, LD, LT, HP, WT}

[0178] Traverse the three 500kV cut points: YC, WL, and LD. First, remove node YC from the original network topology, resulting in two disconnected subgraphs as follows: Figure 12 As shown, Figure 12 The subgraph formed by the nodes in (b) is entirely composed of 220kV nodes, therefore Γ YC={YH,ZH}, in the final result of community partitioning, nodes YH and ZH are merged into the same community as node YC. Next, nodes WL are identified and removed from the original network topology, resulting in two disconnected subgraphs as follows: Figure 13 As shown, Figure 13 (a) and Figure 13 (b) The subgraphs formed by the nodes in the two graphs are not entirely composed of 220kV nodes; therefore, the WL node does not involve the merging of related nodes. Finally, the LD node is removed from the original network topology, resulting in two disconnected subgraphs as shown below. Figure 14 As shown, Figure 14 The subgraph formed by the nodes in (b) is entirely composed of 220kV nodes, therefore Γ LD ={LT,HP,WT}, in the final result of community partitioning, the LT, HP, WT nodes and the LD node will be merged into the same community.

[0179] Therefore, the nodes {LT, HP, WT} and {YH, ZH} are removed from the original network, thereby reducing the size of the input nodes for the partitioning algorithm. After the algorithm is completed, the nodes {LT, HP, WT} and {YH, ZH} are then merged into the same community as the nodes LD and YC, respectively.

[0180] In this example, the number of nodes input to the algorithm was reduced from 13 in the original topology to 8 in the existing network, a reduction of 38.46% in node size.

[0181] (III) Power Grid Node Partitioning Algorithm Based on Louvain's Algorithm

[0182] This section uses a node partitioning method based on the Louvain algorithm to adapt to the needs of power grid partitioning. The original network node topology is then partitioned into communities. After the node size reduction described in Section (2), the network topology G0 analyzed in this section is as follows: Figure 15 As shown.

[0183] The region spans 796 km east to west and 698 km north to south. Based on the actual situation of the power grid in this region, the spatial proportion coefficient r is set to 0.13. The pairs of nodes that cannot be located in the same community are shown in Table 3. The control group uses a spatial proportion coefficient of 1.0, meaning that spatial channel constraints and connectivity constraints are completely disregarded. The reactance value of the edges connecting nodes is used as the weight.

[0184] Table 3. Node pairs that do not satisfy the spatial length constraint.

[0185] 1 WL-YC 2 WL-LD

[0186] Initialize the communities to which all nodes in graph G0 belong. Iterate through all nodes in graph G0 and assign each node to an independent community. The initialization results are shown in Table 4. The current community modularity is Q. 0 = -0.1788.

[0187] Table 4 Initial Community Numbers for Local Power Grid Stations in a Certain Province

[0188] 1 YD 1 2 YC 2 3 LD 3 4 WL 4 5 QN 5 6 ZS 6 7 YCH 7 8 JY 8

[0189] Divide nodes into communities using the following 5 steps:

[0190] (1) First, set the optimal module gain to 0;

[0191] (2) Merge each vertex in G0 with its adjacent nodes and calculate the modularity gain;

[0192] (3) Check whether the movement satisfies the spatial channel constraints and connectivity constraints. If the constraints are satisfied, the modularity gain remains unchanged; if the constraints are not satisfied, the modularity is corrected.

[0193] (4) If the modularity gain is greater than the optimal modularity gain, the node is placed in the community of the neighboring node with the largest modularity gain; if it is less than the current optimal modularity gain, the node move is ignored.

[0194] (5) Repeat steps (2) to (4) until the difference between the two modularity results is less than the given minimum value.

[0195] The minimum value __MIN = 1E-6 is taken, and the K value in both the spatial correction term and the connectivity correction term is taken as 10. 4 That is, the spatial correction term and connectivity correction term for modularity are expressed as:

[0196]

[0197]

[0198] The node distance threshold L is calculated as follows:

[0199]

[0200] Algorithm iteration process:

[0201] (1) For node YD, its community number is 1. Let the best community to which node YD belongs be the current community 2, and initialize the corresponding optimal modularity gain as ΔQ0 = 0. The neighboring nodes of YD are YC and QN.

[0202] We will analyze nodes YC and QN in turn:

[0203] ① If YD is added to Community 5 where QN is located, the modularity gain ΔQ YD-QN = 0.8000; since l YD-QN = 2.45km < L, therefore the node pair YD-QN satisfies the spatial channel constraint; the node pair YD-QN and its connecting edge are mapped to the topology Figure 15 , which satisfies the connectivity constraint. Therefore, the modularity gain ΔQ YD-QN * = ΔQ YD-QN = 0.8000. Since the modularity gain ΔQ YD-QN * > ΔQ0, the optimal modularity gain is ΔQ0 = ΔQ YD-QN * = 0.8000.

[0204] ② If YD is added to Community 2 where YC is located, the modularity gain ΔQ YD-YC = 0.0037; since l YD-YC = 10.1km < L, therefore the node pair YD-YC satisfies the spatial channel constraint; the node pair YD-YC and its connecting edge are mapped to the original topology, which satisfies the connectivity constraint. Therefore, the modularity gain ΔQ YD-YC * = ΔQ YD-YC = 0.0037. However, since the modularity gain ΔQ YD-YC * < ΔQ0, the optimal modularity gain still remains ΔQ0 = ΔQ YD-QN * = 0.8000.

[0205] Therefore, the optimal community for node YD is Community 5 where node QN is located, and the community to which node YD belongs is changed to C YD = 5, at this time the modularity Q of the network 1 = -0.1685, since Q 2 - Q 1 > __MIN, the loop continues.

[0206] (2) For node YCH, the number of the community where this node is located is 7. Let the optimal community that node YCH belongs to be the current community 7, and initialize the corresponding optimal modularity gain as ΔQ0 = 0. Search for neighbor nodes of YCH, which are JY and ZS.

[0207] The two nodes JY and ZS are analyzed sequentially:

[0208] ① If YCH is added to Community 6 where ZS is located, the modularity gain ΔQ YCH-ZS = 0.6019; since l YCH-ZS=10.1km<L, therefore the node pair YCH-ZS satisfies the spatial channel constraint; the node pair YCH-ZS and its connecting edge are mapped to the original topology Figure 15 , and the connectivity constraint is satisfied. Therefore, the modularity gain ΔQ YCH-ZS * =ΔQ YCH-ZS =0.6019. Since the modularity gain ΔQ YCH-ZS * >ΔQ0, the optimal modularity gain is ΔQ0=ΔQ YCH-ZS * =0.6019.

[0209] ② If YCH is added to the community 8 where JY is located, the modularity gain ΔQ YCH-JY =0.6019; since l YCH-JY =10.1km<L, therefore the node pair YCH-JY satisfies the spatial channel constraint; the node pair YCH-JY and its connecting edge are mapped to the topology Figure 15 , and the connectivity constraint is satisfied. Therefore, the modularity gain ΔQ YCH-JY * =ΔQ YCH-JY =0.6019. However, since the modularity gain ΔQ YCH-JY * =ΔQ0, the optimal modularity gain ΔQ0 remains 0.6019, and no node movement is performed.

[0210] Therefore, the optimal community of node YCH is community 6 where node ZS is located. Change the community to which node YCH belongs to C YCH =6, at this time the modularity Q of the network 2 =-0.0201, since Q 2 -Q 1 >__MIN, the loop continues.

[0211] (3) For node JY, the community number of the community where the node is located is 8. Set the optimal community that node JY belongs to as the current community 8, initialize the corresponding optimal modularity gain as ΔQ0=0. Search that the neighbor nodes of JY are YCH and WL.

[0212] Analyze the two nodes YCH and WL sequentially:

[0213] ① If JY is added to the community 4 where WL is located, the modularity gain ΔQ JY-WL =0.7971; since l JY-WL =51km<L, therefore the node pair JY-WL satisfies the spatial channel constraint; the node pair JY-WL and its connecting edge are mapped to the topology Figure 15 , and the connectivity constraint is satisfied. Therefore, the modularity gain ΔQ JY-WL* = ΔQ JY-WL = 0.7971. Due to the modularity gain ΔQ JY-WL * > ΔQ0, the optimal modularity gain is ΔQ0 = ΔQ JY-WL * = 0.7971.

[0214] ② If JY is added to Community 7 where YCH is located, the modularity gain ΔQ JY-YCH = 0.6019; since l JY-YCH = 10.1 km < L, thus the node pair JY-YCH satisfies the spatial channel constraint; the node pair JY-YCH and its connecting edge are mapped to the topology Figure 15 and satisfies the connectivity constraint. Therefore, the modularity gain ΔQ JY-YCH * = ΔQ JY-YCH = 0.6019. However, since the modularity gain ΔQ JY-YCH * < ΔQ0, the optimal modularity gain remains ΔQ0 = ΔQ JY-YCH * = 0.7971.

[0215] Therefore, the optimal community of node JY is Community 4 where node WL is located, and the community to which node JY belongs is changed to C JY = 4, at this time the modularity Q of the network 3 = -0.0203, since Q 3 - Q 2 > __MIN, the loop continues.

[0216] (4) For node WL, the community number of the community where the node is located is 4. Let the optimal community that node WL belongs to be the current Community 4, and initialize the corresponding optimal modularity gain as ΔQ0 = 0. Search for the neighbor nodes of JY, which are YC, JY and LD.

[0217] The three nodes YC, JY and LD are analyzed sequentially:

[0218] ① If WL is added to Community 3 where LD is located, the modularity gain ΔQ WL-LD = 0.7971; the node pair WL-LD and its connecting edge are mapped to the topology Figure 15 and satisfies the connectivity constraint. However, since l WL-LD = 156 km > L, which does not satisfy the spatial channel constraint, the spatial correction term for modularity is:

[0219] P L-(WL,LD) = -10 4

[0220] Therefore, the modularity gain ΔQWL-LD * =ΔQ WL-LD +P L-(WL,LD) =0.7971-10 4 ≈-10 4 <0. Due to the modularity gain ΔQ WL-LD * If the value is less than ΔQ0, then the optimal modularity gain is ΔQ0 = 0, and the WL node does not move.

[0221] ②If WL is added to community 2 where YC is located, the module gain ΔQ WL-YC =0.0053; Maps the node pair WL-YC and its connecting edges to the topology. Figure 15 In the middle, the connectivity constraint is satisfied. However, due to l WL-YC =132km>L, which does not satisfy the spatial channel constraint, so the spatial correction term for modularity is:

[0222] P L-(WL,YC) =-10 4

[0223] Therefore, the module gain ΔQ WL-YC * =ΔQ WL-YC +P L-(WL,YC) =0.0053-10 4 ≈-10 4 <0. Due to the modularity gain ΔQ WL-YC * If the value is less than ΔQ0, then the optimal modularity gain is ΔQ0 = 0, and the WL node does not move.

[0224] ③ As analyzed above regarding the process of JY node joining the community of WL node, the modularity gain ΔQ obtained by adding WL to the community of JY is... WL-JY * Since >ΔQ0, the optimal community for node WL is the community where node JY is located, because C JY =4, then C WL =4. At this point, the network's modularity Q... 4 =0.0202, because Q 4 -Q 3 >__MIN, the loop continues.

[0225] (5) For node QN, its community number is 5. Let the best community to which node QN belongs be its current community 5, and initialize the corresponding optimal modularity gain as ΔQ0 = 0. The neighboring nodes of QN are YD and ZS.

[0226] The analysis will proceed sequentially for nodes YD and ZS:

[0227] ① If node QN is added to community 6 where ZS is located, the modularity gain ΔQ QN-ZS = 0.6019; since l QN-ZS = 13.8km < L, the node pair QN-ZS satisfies the spatial channel constraint; the node pair QN-ZS and its connecting edge are mapped to the topological Figure 15 and satisfy the connectivity constraint. Therefore, the modularity gain ΔQ QN-ZS * = ΔQ QN-ZS = 0.6019. Since the modularity gain ΔQ QN-ZS * > ΔQ0, the optimal modularity gain is ΔQ0 = ΔQ QN-ZS * = 0.6019.

[0228] ② The modularity gain of adding node QN to community 1 where YD is located is ΔQ YD-QN = 0.8000 > ΔQ0, so the optimal community for node QN is the community where node YD is located. Since C YD = 5, then C QN = 5, and the modularity Q of the network at this time 5 = 0.1097. Since Q 5 - Q 4 > __MIN, the loop continues.

[0229] (6) For node LD, the community number of the community where it is located is 3. Let the optimal community that node LD belongs to be the current community 3, and initialize the corresponding optimal modularity gain as ΔQ0 = 0. Search for the neighbor node of LD, which is only node WL. It can be known from the above analysis that the node pair LD-WL cannot be divided into the same community because it does not satisfy the spatial channel constraint, so the optimal community of node LD remains unchanged, that is, C LD = 3. The modularity Q of the network at this time 8 = 0.2985. Since Q 6 - Q 5 > __MIN, the loop continues.

[0230] (7) For node ZS, the community number of the community where it is located is 6. Let the optimal community that node ZS belongs to be the current community 6, and initialize the corresponding optimal modularity gain as ΔQ0 = 0. Search for the neighbor nodes of ZS, which are QN and YCH.

[0231] Analyze the two nodes QN and YCH sequentially:

[0232] ① If node ZS is added to community 7 where YCH is located, the modularity gain ΔQ ZS-YCH = 0.6019; since l QN-ZS= 13.8km < L, thus the node pair QN-ZS satisfies the spatial channel constraint; mapping the node pair QN-ZS and its connecting edge to the topology Figure 15 , the connectivity constraint is satisfied. Therefore, the modularity gain ΔQ QN-ZS * = ΔQ QN-ZS = 0.6019. Since the modularity gain ΔQ QN-ZS * > ΔQ0, the optimal modularity gain is ΔQ0 = ΔQ QN-ZS * = 0.6019.

[0233] ② It can be known from the above analysis of the process of adding node QN to the community where node ZS is located that the modularity gain ΔQ ZS-QN * obtained by adding ZS to the community where QN is located is 0.6019 = ΔQ0, then the optimal modularity gain ΔQ0 remains 0.6019, and no node movement is performed.

[0234] Therefore, the optimal community of node ZS is the community where node YCH is located. Since C YCH = 6, change the community of node ZS to C ZS = 6, at this time the modularity Q of the network 7 = 0.1001, since Q 7 - Q 6 > __MIN, the loop continues.

[0235] (8) For node YC, the community number of the node is 2. Suppose that the optimal community of node YC is the current community 2, and initialize the corresponding optimal modularity gain as ΔQ0 = 0. The neighbor nodes of YC are WL and YD. It can be known from the above analysis that the node pair YC-WL cannot be divided into the same community because it does not satisfy the spatial channel constraint; and the modularity gain ΔQ YC-YD obtained by merging node pair YC-YD into the same community is 0.0037 > ΔQ0, therefore the optimal community of node YC is the community where node YD is located, since C YD = 5, therefore C YC = 5, at this time the modularity Q of the network 8 = 0.2985, since Q 8 - Q 7 > __MIN, the loop continues.

[0236] (9) Search node ZS again, and there is no node movement. At this time, the modularity Q of the network 9 = 0.2985, since Q 9 - Q 8 = 0 < __MIN, the loop ends.

[0237] The results of this round of iterative community division are shown in Table 5. Nodes with the same community number are regarded as a super node and renumbered as shown in Table 6.

[0238] Table 5 Results of the first round of grid node partitioning in the example.

[0239] 1 YD 5 2 YC 5 3 LD 3 4 WL 4 5 QN 5 6 ZS 6 7 YCH 6 8 JY 4

[0240] Table 6. Equivalent results of power grid nodes in the example.

[0241] 1 YD, YC, QN 1 2 LD 2 3 WL, JY 3 4 ZS、YCH 4

[0242] Following the traditional Louvain algorithm, the equivalent node graph G1 is regenerated from the equivalent results in Table 6. The above steps are then repeated on G1 until the modularity change is less than the __MIN value, at which point the loop exits. No nodes move during the second loop, and the community partitioning algorithm ends. At this point, the number of communities is 4. The results are summarized as follows: Figure 16 As shown.

[0243] Finally, the nodes {LT, HP, WT} and {YH, ZH} are merged into the same community as the nodes LD and YC, respectively. The final community division result is as follows. Figure 17 As shown in Table 7, the specific node names and corresponding community numbers included in each community are listed below.

[0244] Table 7 Final Results of Power Grid Node Community Division in the Case Study

[0245] 1 YD, YC, QN, YH, ZH 1 2 LD, LT, HP, WT 2 3 WL, JY 3 4 ZS、YCH 4

[0246] Based on the community division results, the final equivalent network structure diagram of the local power grid is as follows: Figure 18 As shown.

[0247] According to the algorithm of the present invention, the power grid nodes in the region are divided into 4 regions. Node pairs that do not meet the spatial channel constraints are assigned to different communities, and the nodes in each community maintain connectivity in the original topology.

[0248] In summary, compared with the corresponding technology, the present invention has at least the following advantages and beneficial effects:

[0249] 1) This invention proposes a method for rapidly merging low-voltage level node partitions by identifying and analyzing the electromagnetic ring network structure. A large number of radially distributed low-voltage level nodes are merged into communities containing corresponding high-voltage level nodes, while preserving the independence of low-voltage level nodes and branches within the electromagnetic ring network. This effectively reduces the node size that subsequent community partitioning algorithms must handle, while fully preserving the characteristics of the power grid topology.

[0250] 2) In view of the requirement to preserve the structural features of the power grid, this invention designs spatial constraints and connectivity constraints for node partitioning, improves the Louvain algorithm process and criteria for community identification, and forms a node community partitioning algorithm applicable to the power grid.

[0251] This embodiment also provides a power grid zoning equivalence device considering community structure, including:

[0252] At least one processor;

[0253] At least one memory for storing at least one program;

[0254] When the at least one program is executed by the at least one processor, the at least one processor performs the following: Figure 19 The method shown.

[0255] This embodiment provides a power grid zoning equivalence device that considers community structure. It can execute a power grid zoning equivalence method that considers community structure provided in the method embodiment of the present invention. It can execute any combination of implementation steps of the method embodiment and has the corresponding functions and beneficial effects of the method.

[0256] This application also discloses a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device can read the computer instructions from the computer-readable storage medium and execute the computer instructions, causing the computer device to perform... Figure 19 The method shown.

[0257] This embodiment also provides a storage medium storing instructions or programs that can execute the power grid partitioning equivalence method considering community structure provided in the method embodiment of the present invention. When the instructions or programs are run, any combination of implementation steps of the method embodiment can be executed, and the method has the corresponding functions and beneficial effects.

[0258] In some alternative embodiments, the functions / operations mentioned in the block diagrams may not occur in the order shown in the operation diagrams. For example, depending on the functions / operations involved, two consecutively shown blocks may actually be executed substantially simultaneously, or the blocks may sometimes be executed in reverse order. Furthermore, the embodiments presented and described in the flowcharts of this invention are provided by way of example to provide a more comprehensive understanding of the technology. The disclosed methods are not limited to the operations and logic flows presented herein. Alternative embodiments are contemplated in which the order of various operations is altered and sub-operations described as part of a larger operation are executed independently.

[0259] Furthermore, although the invention has been described in the context of functional modules, it should be understood that, unless otherwise stated, one or more of the described functions and / or features may be integrated into a single physical device and / or software module, or one or more functions and / or features may be implemented in a separate physical device or software module. It is also understood that a detailed discussion of the actual implementation of each module is unnecessary for understanding the invention. Rather, given the properties, functions, and internal relationships of the various functional modules in the apparatus disclosed herein, the actual implementation of the module will be understood within the scope of conventional skill of an engineer. Therefore, those skilled in the art can implement the invention as set forth in the claims using ordinary techniques without excessive experimentation. It is also understood that the specific concepts disclosed are merely illustrative and not intended to limit the scope of the invention, which is determined by the full scope of the appended claims and their equivalents.

[0260] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, essentially, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0261] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0262] More specific examples of computer-readable media (a non-exhaustive list) include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0263] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0264] In the foregoing description of this specification, references to terms such as "one embodiment," "another embodiment," or "some embodiments" indicate that a specific feature, structure, material, or characteristic described in connection with an embodiment or example is included in at least one embodiment or example of the present invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0265] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

[0266] The above is a detailed description of the preferred embodiments of the present invention. However, the present invention is not limited to the above embodiments. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of the present invention. All such equivalent modifications or substitutions are included within the scope defined by the claims of this application.

Claims

1. A method for equivalence of power grid zoning considering community structure, characterized in that, Includes the following steps: Acquire power grid data and preprocess the power grid data; For the preprocessed power grid data, an automatic electromagnetic ring network identification algorithm based on the power grid topology is used to retain the first node related to the electromagnetic ring network and merge the radial second nodes into communities. To address the requirement of preserving power grid structural features, spatial channel constraints and connectivity constraints for node partitioning are designed, and the Louvain algorithm process and criteria are improved. After partitioning the first node using the improved Louvain algorithm, the second node is added to the community to which the first node belongs, forming the final community partitioning result. The automatic electromagnetic ring network identification algorithm obtains the electromagnetic ring network subgraph in the following way: Let the set of nodes for high voltage levels be { The corresponding voltage level is The set of nodes for low voltage levels is { The corresponding voltage level is represented as} ; Let { contain }and{ The complete power grid topology diagram of the node is as follows: ; The automatic identification steps for electromagnetic ring networks are as follows: A1. Delete image All suspended nodes; A2. Delete image All { } node; delete { }back,{ The network composed of nodes will be divided into several non-connected subgraphs, denoted as . ; A3. Delete image All { } nodes, resulting in a connected subgraph containing only high-voltage nodes, is called ; A4. For each subgraph In the diagram Search for the connection edges between the subgraph and nodes outside the subgraph; the nodes on the opposite side of the edge all belong to { The nodes in the subgraph are called the high-voltage power supply nodes of that subgraph. A5, Subgraph If the number of high-voltage power supply nodes is 1, then the subgraph node and its high-voltage power supply node are assigned to the same community, and the subsequent search of the subgraph is skipped; A6. Subgraph If all high-voltage power supply nodes are connected to only one low-voltage node, then skip the subsequent search of this subgraph; A7, Subgraph If the high-voltage power supply node does not meet the conditions in steps A5 and A6, then in the connected subgraph Search subgraph Find the shortest path between all high-voltage power supply nodes, and add the found nodes and edges to the subgraph. The resulting new subgraph is called , This is a sub-graph of an electromagnetic ring network; The improved Louvain algorithm works as follows: D1. Obtain the power grid topology model; D2. Initialization: Treat each node in the power grid topology model as an independent community; D3. Clustering optimization: For each node, try to move it to the community of its neighboring nodes; D4. Determine whether the spatial channel constraints are met. If they are met, proceed to step D5. If the conditions are not met, proceed to step D6; D5. Determine whether the connectivity constraint is satisfied. If satisfied, proceed to step D7. If not satisfied, proceed to step D6; D6. Set the constraint-based correction term to a negative value with a large absolute value; D7. Calculate the improved modularity gain; D8. Determine if there is a move that increases the module level. If there is an increase, return to step D7. If there is no improvement, proceed to step D9; D9. Community Aggregation: Using the current community as a super node to build a new network; D10. On the new network, determine whether there are still moves that improve module degree. If there are, return to step D7; if not, proceed to step D11. D11. Output the final community division results.

2. The power grid zoning equivalence method considering community structure according to claim 1, characterized in that, The preprocessing of the power grid data includes: Identify the set of nodes within the same substation according to the preset node naming rules; If the impedance of the connecting branch between nodes is less than a preset value, identify the nodes operating in parallel within the substation and merge these nodes. After preprocessing, the topological relationships between nodes are represented as a graph. , It is a combination of nodes. It is a set of transmission lines connecting nodes, called the edge set; the edge information is stored using {{node number and name, voltage level, and site number of one end}, {node number and name, voltage level, and site number of the other end}}.

3. The power grid zoning equivalence method considering community structure according to claim 1, characterized in that, Low-voltage level nodes are grouped into communities using the following method: B1. Identify electromagnetic loops in the network using an automatic electromagnetic loop identification algorithm; B2, Retention and { } Nodes form an electromagnetic ring network structure { }node; B3. Using the cut point method, transform the original network topology. Electromagnetic ring network Other nodes outside the node { The nodes are identified and grouped, and then merged with the corresponding cut points into a community.

4. The power grid zoning equivalence method considering community structure according to claim 3, characterized in that, The execution steps of the cut point method are as follows: C1. Identify the original network topology The cut points in the middle, and All articulation points are stored in a set. middle; C2, For sets { in The nodes are traversed sequentially, with each traversal revisiting the original network topology. Execution on: For sets Nodes in , will node From the original network topology Remove, and obtain a new network topology. ;statistics Number of connected subgraphs N ;like N =1, then skip the node. The loop; if N >1, and the connected subgraph formed by the separated nodes is entirely composed of { If the nodes are composed of a set, they are stored in the set. In the middle, Nodes and nodes Belong to the same community; otherwise skip the node. The loop continues until the collection is complete. Traverse all nodes in the array.

5. The power grid zoning equivalence method considering community structure according to claim 1, characterized in that, The spatial channel constraint detection detects any two nodes. , Spatial distance between Does it exceed the threshold L? Where, the node distance to the threshold is... The calculation formula is as follows: In the formula, Indicates the node distance threshold. and These represent the total east-west span and total north-south span of the power grid coverage area, respectively. The set proportional coefficient; Based on the detection results of spatial channel constraints, the spatial correction term for modularity is calculated, as shown in the following expression: In the formula, This is a spatial correction term for modularity; For nodes and nodes Spatial distance between them It is a positive value with a very large absolute value. This is the default value; The connectivity constraint detection node From the original club Divide to node Club Afterwards, will the two clubs lose their connection? In each iteration, nodes After removing from one community and joining another, ensure the node... The connectivity of the two communities before and after the move in the original topology; based on the detection results of connectivity constraints, the connectivity correction term of the modularity is calculated, as shown in the following expression: In the formula, This is a connectivity correction term for modularity.

6. The power grid zoning equivalence method considering community structure according to claim 5, characterized in that, The improved modularity gain is calculated using the following formula: In the formula, This represents the improved modularity gain. This is a spatial correction term for modularity; This is a connectivity correction term for modularity; This is the module gain in Louvain's algorithm.

7. A power grid zoning equivalence device considering community structure, characterized in that, include: At least one processor; At least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements the method according to any one of claims 1-6.

8. A computer-readable storage medium storing a processor-executable program, characterized in that, The processor-executable program, when executed by the processor, is used to perform the method as described in any one of claims 1-6.

Images