Voltage partitioning method based on community structure theory and regional reactive power supply and demand balance

Abstract

The application discloses a voltage partition method based on a community structure theory and regional reactive power supply and demand balance, and is characterized by comprising the following steps: step S1: according to the topological structure of a power system network, lossless processing is performed on the power system network, and reactive power loss on a line of the power system network is equivalent to reactive power load at nodes at two ends of the line; step S2: the community structure theory is used to perform reactive power partition on the power system network, a penalty function is set according to a difference between reactive power demand and reactive power supply in a community, and a reactive power imbalance degree is set according to a ratio between the reactive power demand and the reactive power supply in the community; a community structure quantitative evaluation index modularization function Q' is established in combination with the penalty function and the reactive power imbalance degree; step S3: arrangement and combination are performed on community partition, the community structure quantitative evaluation index modularization function Q' is calculated and updated, and the community partition with the maximum modularization function Q' value is selected as a final reactive power partition result. Therefore, the best reactive power partition result is obtained.

Description

Technical Field

[0001] This invention relates to the field of power distribution network system automation technology, and in particular to a voltage zoning method based on community structure theory and regional reactive power supply and demand balance. Background Technology

[0002] With the development of power technology, my country's current power system is characterized by large grids, long distances, high loads, and multiple generating units. This poses a significant challenge to the stability of voltage at various nodes within the system. Ensuring voltage stability within a specified range is a pressing issue that needs to be addressed in today's power environment. Reactive power is directly related to voltage. To reduce grid losses, power systems should generally avoid long-distance reactive power transmission. Therefore, regional control of reactive power and voltage is essential, and how to zonify the grid for reactive power and voltage is the foundation of Automatic Voltage Control (AVC). Currently, reactive power is used as an ancillary service, relying on the active power market for its input. However, as a crucial means of voltage regulation, an independent reactive power market will inevitably emerge in the future. Zoning reactive power facilitates regional pricing, thus laying the foundation for the establishment of a reactive power market.

[0003] Currently, my country's AVC mainly adopts a "soft" three-level voltage control mode. Among them, the second-level voltage control is the most critical link in maintaining the system voltage level and improving the system voltage stability. Therefore, it is very important to divide the entire system into multiple approximately decoupled regions for separate control. The principle of reactive voltage control zoning is to ensure strong coupling within the region and weak coupling between regions. Existing reactive voltage zoning methods mainly include: (1) Sensitivity method zoning: based on node reactive voltage sensitivity or electrical distance; (2) Heuristic algorithm zoning: based on various genetic algorithms; (3) Clustering algorithm zoning: divided into hierarchical clustering, density clustering, fuzzy clustering, etc. However, sensitivity method zoning and heuristic algorithm zoning require pre-setting the number of zoning, which limits the freedom of reactive voltage zoning. Moreover, the above zoning methods do not consider whether the reactive power supply and reactive power demand within the region are balanced, and do not consider whether the reactive power supply can meet the reactive power demand. The divided regions may have a situation where the reactive power demand is greater than the reactive power supply, which leads to the inability to perform voltage control normally.

[0004] For example, a "Black-Start Partitioning Method Based on Semi-Supervised Spectral Clustering Algorithm" disclosed in Chinese patent literature, publication number CN107330809A, includes defining line weights as the reciprocal of the electrical distance between nodes and constructing an undirected weighted graph of the power system based on this. The adjacency matrix W and diagonal matrix D are calculated separately. A generator grouping model is established to obtain generator grouping information, and pairwise constraint information between generator pairs is obtained based on the generator grouping information, with corresponding modifications to the adjacency matrix W. The denormalized Laplace matrix L is calculated. The eigenvectors corresponding to the first k smallest eigenvalues ​​of L are found, and these vectors form a column matrix H. Each row of H is treated as a vector in k-dimensional space, and clustering calculations are performed. The category of each row in the clustering result is the partition to which the n nodes in the original graph G belong. However, the above scheme does not consider whether the reactive power supply and reactive power demand in the region are balanced, nor whether the reactive power supply can meet the reactive power demand, resulting in some issues.

[0005] The region may experience a situation where reactive power demand exceeds reactive power supply, leading to problems with normal voltage control. Summary of the Invention

[0006] This invention aims to overcome the supply-demand imbalance problem in existing power system network partitioning technologies. It provides a voltage partitioning method based on community structure theory and regional reactive power supply-demand balance, which considers the balance of reactive voltage and reactive power supply within a region, and ensures that reactive power sources within a region can effectively regulate and control the node voltage within the region, thereby achieving a better voltage partitioning effect.

[0007] To achieve the above objectives, the present invention adopts the following technical solution:

[0008] A voltage zoning method based on community structure theory and regional reactive power supply and demand balance is characterized by the following steps:

[0009] Step S1: Based on the topology of the power system network, perform lossless processing on the power system network, and convert the reactive power loss on the power system network lines into reactive power load at the nodes at both ends of the lines.

[0010] Step S2: The power system network is partitioned using community structure theory. A penalty function is set based on the difference between reactive power demand and reactive power supply within each community, and the reactive power imbalance degree is set based on the ratio of reactive power demand to reactive power supply within each community. A modular function for quantitative evaluation of community structure is established by combining the penalty function and the reactive power imbalance degree. ;

[0011] Step S3: Arrange and combine the community zones, and calculate and update the modular function of the quantitative evaluation index of community structure. Select modular functions The community partition with the maximum value is the final reactive partition result.

[0012] The circuit adopts a π-type equivalent circuit. This refers to the reactive power injected into node 1; This represents the reactive power flowing out of node 2; The voltages at nodes 1 and 2; This is the equivalent capacitance across the two ends of the π-type equivalent circuit. This is the corresponding reactive power input to its ground capacitance; This refers to reactive power loss on the line, and This allows the reactive power losses on the line to be equated to the reactive power load at the nodes at both ends of the line. ,

[0013] If a network's internal community structure is well-defined, then the weight of edges connecting within each community should be significantly higher than the weight of edges in a random network with the same number of nodes. A higher value indicates a better division of the community, ensuring that the reactive power sources within the area can effectively regulate and control the node voltage within the area, thereby achieving a better reactive voltage zoning effect.

[0014] Preferably, step S1 further includes:

[0015] Step S12: Perform preliminary power flow calculations on the power system network and construct the Newton-Raphson algorithm to obtain the Jacobian matrix of the power system network:

[0016]

[0017] in, The active power injected into the node. Reactive power injected into the node The voltage amplitude at the node. The phase angle of the node; These are the Jacobian matrix coefficients;

[0018] Considering only the reactive power and voltage control relationship, let ΔP = 0; then we get: ;

[0019] make Then, the sensitivity relationship of node voltage to reactive power changes can be obtained:

[0020] ΔV = SΔQ;

[0021] Where the elements in matrix S are This represents the voltage change at node i when the reactive power injection at node j changes by one unit.

[0022] Define an electrical distance mapping matrix M between two arbitrary nodes; the elements of matrix M satisfy the following relationship:

[0023]

[0024] Under normal circumstances This is to characterize the symmetry of the coupling degree between two nodes.

[0025] Preferably, step S1 further includes constructing a node connection matrix A based on the topology of the power system network; the matrix A satisfies the following relationship:

[0026] If node i and node j are connected, then If node i and node j are not connected, then The main diagonal elements of matrix A are all 0. A is a symmetric matrix.

[0027] Preferably, the weight matrix of the node connection edges in the power system network is set to W, and the matrix W satisfies: ;

[0028] Right now:

[0029] Where matrix W is the Hadamard product of matrices A and M; the elements of matrix W... It represents the weight of the edge connecting node i and node j. The matrix W is the Hadamard product of matrices A and M, which is obtained by multiplying the elements in the same row and column of matrices A and M respectively. W is the weight matrix representing the weights of the nodes connecting the edges in the power system network.

[0030] Preferably, the penalty function set in step S2 based on the difference between reactive power demand and reactive power supply within the community satisfies the following relationship:

[0031] Penalty function

[0032] Where u is the club number; For step function, when hour, ;when hour, ; For the total reactive power supply within the community, The total reactive power demand within the community. Let u be the reactive power supply-demand difference within the community. A penalty function is introduced. The goal is to ensure that, in the scheme for dividing the network into communities, the reactive power supply within each community is not less than its reactive power demand, thus satisfying the constraint condition. .

[0033] Preferably, the reactive power imbalance setting in step S2, based on the ratio of reactive power demand to reactive power supply within the community, satisfies the following relationship:

[0034] reactive power imbalance

[0035] in, Total reactive power supply within community u The total reactive power demand within community u. When reactive power imbalance occurs... The larger the value, the more it indicates an imbalance between reactive power supply and demand within the community (u); when the reactive power imbalance... When the value approaches 0, it indicates that the reactive power supply and demand within community u are relatively balanced. This prevents an excessively large difference between reactive power supply and demand within a community, i.e., reactive power supply far exceeding its reactive power demand. Preferably, the modular function Q' of the community structure quantitative evaluation index described in step S2 satisfies the following relationship:

[0036]

[0037] in, This represents the weight of the edge connecting node i and node j; m is the sum of the weights of all connected edges. These are the sum of the weights of the edges connected to nodes i and j, respectively. Assign community numbers to nodes i and j, when hour, ,otherwise ; , , These are constant coefficients, obtained from experimental results. This is used to determine whether reactive power supply and demand belong to community u. When reactive power supply and demand belong to community u, ,otherwise If a network's internal community structure is well-defined, then the weight of edges connecting within each community should be significantly higher than the weight of edges in a random network with the same number of nodes. A higher value indicates a better effect in dividing the community.

[0038] Preferably, step S3 involves arranging and combining community partitions and calculating and updating the modular function of the community structure quantitative evaluation index. Select modular functions The community partition with the maximum value is the final reactive partition result, which includes the following detailed steps:

[0039] Step S31: Divide each node in the power system network into a community and calculate the initial value of the defined modular function. ;

[0040] Step S32: Traverse the elements in matrix W, if If so, then community i and community j are combined and merged to form a new community, which is regarded as a new node, and the weight W of the corresponding connecting edge is updated;

[0041] Step S33: Calculate the reactive power supply and reactive power demand of the new community, and calculate the modular function of the community structure quantitative evaluation index for the new community. ;

[0042] Step S34: Calculate the incremental modular function of the updated community quantitative evaluation index , The updated club structure

[0043] '

[0044] Subtract the previously updated modular function value of the community structure quantitative evaluation index from the modular function value of the quantitative evaluation index, and store the ΔQ value in the pending queue.

[0045] Step S35: Repeat steps S32 to S34 until all elements in matrix W have been traversed;

[0046] Step S36: Traverse the queue to be processed. If there is a queue to be processed... Then select The largest possible community combination is to merge the new community, update the weight W of the corresponding connecting edges, update the matrix W and the total reactive power supply and demand within the new community, and treat the new community as a new node. If a certain community does not exist in the queue to be processed... If the current community combination scheme is selected, then the optimal reactive power partitioning result can be obtained.

[0047] Preferably, step S32 involves merging community i and community j to form a new community, which is considered a new node, and updating the weight W of the corresponding connecting edges. This includes the following steps: the connecting edges between the new node and other nodes outside the new community are the connecting edges between the original community i and community j and nodes outside the new community. If the original two communities are connected to the same node outside the new community, they are merged into the same edge, and the weight of the connecting edge is the sum of the weights of the corresponding edges of community i and j.

[0048] As a preferred option, it also includes making ,like If node t is selected as the dominant node for the region, then the dominant node represents the region, and its voltage reflects the voltage level of the region, facilitating the monitoring and regulation of the power system network voltage.

[0049] Therefore, the present invention has the following beneficial effects: (1) The present invention regards the power system network as a weighted, undirected connected network, and the reactive power loss on the line is equivalent to the reactive power load of the two-end nodes and the node connection edge weight matrix related to the reactive power voltage sensitivity between the nodes; (2) Based on the community structure theory, a penalty function and reactive power imbalance are introduced, and a modular function of community structure quantitative evaluation index is set. (3) Combine the elements of the weight matrix W to recombine the communities and update the modular function of the community structure quantitative evaluation index. To obtain the modularity before and after merging communities in various combination schemes. The function increments to select the optimal reactive power partitioning result. Attached Figure Description

[0050] Figure 1 is a flowchart of a voltage partitioning method based on community structure theory and regional reactive power supply and demand balance according to an embodiment of the present invention.

[0051] Figure 2 is a flowchart of an embodiment of the present invention, which uses a voltage partitioning method based on community structure theory and regional reactive power supply and demand balance to select the optimal partitioning method by arranging and combining community partitions.

[0052] Figure 3 is a schematic diagram of a lossless equivalent processing method for power system networks according to an embodiment of the present invention. Detailed Implementation

[0053] The present invention will now be further described with reference to the accompanying drawings and specific embodiments.

[0054] Example:

[0055] Figure 1 illustrates a reactive voltage zoning method based on community structure theory and regional reactive power supply and demand balance, which employs the following steps:

[0056] Step S1: Based on the topology of the power system network, perform lossless processing on the power system network, and convert the reactive power loss on the power system network lines into reactive power load at the nodes at both ends of the lines.

[0057] As shown in Figure 3, the circuit adopts a π-type equivalent circuit. This refers to the reactive power injected into node 1; This represents the reactive power flowing out of node 2; The voltages at nodes 1 and 2; This is the equivalent capacitance across the two ends of the π-type equivalent circuit. This is the corresponding reactive power input to its ground capacitance; This refers to reactive power loss on the line, and The reactive power losses on the line can then be equated to the reactive power load at the nodes at both ends of the line. .

[0058] Step S12: Perform preliminary power flow calculations on the power system network and construct the Newton-Raphson algorithm to obtain the Jacobian matrix of the power system network:

[0059]

[0060] in, The active power injected into the node. Reactive power injected into the node The voltage amplitude at the node. Phase angle of the node These are the Jacobian matrix coefficients;

[0061] Considering only the relationship between reactive power and voltage control, let We can obtain: ;

[0062] make Then, the sensitivity relationship of node voltage to reactive power changes can be obtained:

[0063] ;

[0064] Where the elements in matrix S are This represents the voltage change at node i when the reactive power injection at node j changes by one unit.

[0065] Define an electrical distance mapping matrix M between two arbitrary nodes; the elements of matrix M satisfy the following relationship:

[0066]

[0067] Under normal circumstances This is to characterize the symmetry of the coupling degree between two nodes.

[0068] make ,like If node t is selected as the dominant node for the region, then the dominant node represents the region, and its voltage reflects the voltage level of the region, facilitating the monitoring and regulation of the power system network voltage.

[0069] Construct a node connection matrix A based on the topology of the power system network; the matrix A satisfies the following relationship:

[0070] If node i and node j are connected, then If node i and node j are not connected, then The main diagonal elements of matrix A are all 0. A is a symmetric matrix.

[0071] Let W be the weight matrix of the node connection edges in the power system network, where the matrix W satisfies: W = ;Right now: ;

[0072] Where matrix W is the Hadamard product of matrices A and M; the elements of matrix W... It represents the weight of the edge connecting node i and node j. The matrix W is the Hadamard product of matrices A and M, which is obtained by multiplying the elements in the same row and column of matrices A and M respectively. W is the weight matrix representing the weights of the nodes connecting the edges in the power system network.

[0073] Step S2: The power system network is partitioned using community structure theory. A penalty function is set based on the difference between reactive power demand and reactive power supply within each community, and the reactive power imbalance degree is set based on the ratio of reactive power demand to reactive power supply within each community. A modular function for quantitative evaluation of community structure is established by combining the penalty function and the reactive power imbalance degree. ;

[0074] The penalty function satisfies the following relationship:

[0075] Penalty function

[0076] Where u is the club number; For step function, when hour, ;when hour, ; For the total reactive power supply within the community, The total reactive power demand within the community. Let u be the reactive power supply-demand difference within the community. A penalty function is introduced. The goal is to ensure that, in the scheme for dividing the network into communities, the reactive power supply within each community is not less than its reactive power demand, thus satisfying the constraint condition. .

[0077] The reactive power imbalance degree satisfies the following relationship: Reactive power imbalance degree

[0078] in, For the total reactive power supply within the community u, The total reactive power demand within community u. When reactive power imbalance occurs... The larger the value, the more it indicates an imbalance between reactive power supply and demand within the community (u); when reactive power is unbalanced... When the value approaches 0, it indicates that the reactive power supply and demand within community u are relatively balanced. This prevents an excessively large difference between reactive power supply and demand within a community, i.e., reactive power supply far exceeding reactive power demand.

[0079] Existing technologies typically use modular Q-functions defined by Girvan and Newman to quantify the structure of communities, as follows:

[0080]

[0081] in

[0082] The weights of the edges connecting nodes i and j are represented by elements of matrix W; m is the sum of the weights of all connected edges. These are the sum of the weights of the edges connected to nodes i and j, respectively. Assign community numbers to nodes i and j, when hour, ,otherwise From equation (6), it can be seen that... Furthermore, the Q-value means that if the community structure within a network is well-partitioned, the weight of the edges within each community should be significantly greater than the weight of the edges in a random network with the same number of nodes. A larger Q-value, closer to 1, indicates a better community partitioning effect.

[0083] When considering reactive power zoning in this invention, it is also necessary to ensure that the reactive power demand within a zoning zone does not exceed the reactive power supply that the zone can provide; otherwise, the reactive power demand will exceed the range that the reactive power source can control. Therefore, this invention introduces a penalty function. This invention addresses the reactive power supply and demand situation within a specific community (or partition) u. While the goal of dividing communities is to achieve reactive power balance within each community, to prevent excessively large imbalances in reactive power supply and demand within a community (i.e., reactive power supply far exceeding demand), this invention defines a reactive power imbalance degree within community u. By incorporating penalty functions and reactive power imbalance into the evaluation indicators, a modular function for the quantitative evaluation indicators of community structure is redefined. The following relationship must be satisfied:

[0084]

[0085] in, This represents the weight of the edge connecting node i and node j; m is the sum of the weights of all connected edges. These are the sum of the weights of the edges connected to nodes i and j, respectively. Assign community numbers to nodes i and j, when hour, ,otherwise ; These are constant coefficients, obtained from experimental results. Used to define whether reactive power supply and demand belong to community u. When reactive power supply and demand belong to community u... ,otherwise If a network's internal community structure is well-defined, then the weight of edges connecting within each community should be significantly higher than the weight of edges in a random network with the same number of nodes. A higher value indicates a better effect in dividing the community.

[0086] Step S3: Arrange and combine the community zones, and calculate and update the modular function of the quantitative evaluation index of community structure. Select modular functions The community partition with the maximum value is the final reactive partition result.

[0087] Specifically, as shown in Figure 2, the steps include the following sub-steps:

[0088] Step S31: Divide each node in the power system network into a community and calculate the initial value of the defined modular function. ;

[0089] Step S32: Traverse the elements in matrix W, if If community i and community j are combined and merged, the new community formed is regarded as a new node. The edges connecting the new node to other nodes outside the new community are the edges connecting the original community i and community j to the nodes outside the new community. If the original two communities are connected to the same node outside the new community, they are merged into the same edge. The weight of the connecting edge is the sum of the weights of the corresponding edges of community i and j, and the weight W of the corresponding connecting edge is updated.

[0090] Step S33: Calculate the reactive power supply and reactive power demand of the new community, and calculate the modular function of the community structure quantitative evaluation index for the new community. If a network's internal community structure is well-defined, then the weight of edges connecting within each community should be significantly higher than the weight of edges in a random network with the same number of nodes. A higher value indicates a better effect in dividing the community.

[0091] Step S34: Calculate the updated modular function increment ΔQ' of the community quantitative evaluation index. ΔQ' is the value of the modular function of the community structure quantitative evaluation index updated this time minus the value of the modular function of the community structure quantitative evaluation index updated last time. Store the value of ΔQ' in the queue to be processed.

[0092] Step S35: Repeat step S35 Continue until all elements in matrix W have been traversed;

[0093] Step S36: Traverse the queue to be processed. If there is a queue to be processed... Then select The largest possible community combination is to merge the new community, update the weight W of the corresponding connecting edges, update the matrix W and the total reactive power supply and demand within the new community, and treat the new community as a new node. If a certain community does not exist in the queue to be processed... If the current community combination scheme is selected, then the optimal reactive power partitioning result can be obtained.

[0094] The specific embodiments described herein are merely illustrative of the spirit of the invention. Those skilled in the art to which this invention pertains may make various modifications or additions to the described specific embodiments or use similar methods to substitute them, without departing from the spirit of the invention or exceeding the scope defined by the appended claims.

[0095] Although this paper uses terms such as reactive power loss, reactive power demand, reactive power imbalance, modular function, and partitioning extensively, the possibility of using other terms is not excluded. These terms are used merely to facilitate the description and explanation of the essence of this invention; interpreting them as any additional limitation would contradict the spirit of this invention.

Claims

1. A voltage partitioning method based on community structure theory and regional reactive power supply and demand balance, characterized in that, Comprising the following steps: Step S1: According to the topology of the power system network, lossless processing is performed on the power system network, and the reactive power loss on the line of the power system network is equivalent to the reactive power load at both ends of the line; Step S2: The community structure theory is used to perform reactive power partitioning on the power system network, a penalty function is set according to the difference between the reactive power demand and the reactive power supply in the community, and a reactive power imbalance degree is set according to the ratio of the reactive power demand to the reactive power supply in the community; and a community structure quantitative evaluation index modularization function Q' is established in combination with the penalty function and the reactive power imbalance degree; Step S3: The community partitions are arranged and combined, the community structure quantitative evaluation index modularization function Q' is updated, and the community partition with the maximum modularization function Q' value is selected as the final reactive power partitioning result.

2. The method of claim 1, wherein the method is characterized by: The step S1 further comprises: Step S12: Preliminary power flow calculation is performed on the power system network to construct a Newton-Raphson algorithm to obtain a Jacobian matrix of the power system network: where ΔP is the active power injected at the node, ΔQ is the reactive power injected at the node, ΔV is the voltage magnitude at the node, and Δθ is the phase angle at the node; J PV , Pθ , QV , Qθ is the Jacobian matrix coefficient; Considering only the reactive power and voltage control relationship, let ΔP = 0; we can get: Let The sensitivity of the node voltage to the reactive power change is obtained as follows: ΔV=SΔQ; where the elements of matrix S are S ij represents the change in voltage at node i when the reactive power injection at node j is changed by one unit. An electrical distance relationship mapping matrix M between any two nodes is defined, and the elements in the matrix M satisfy the following relationship:

3. The method of claim 2, wherein the method further comprises: The step S1 further comprises constructing a node connection matrix A according to the topology of the power system network, and the matrix A satisfies the following relationship: If node i and node j are connected, then a ij = 1; if node i and node j are not connected, then a ij = 0; the main diagonal elements of matrix A are 0.

4. The method of claim 3, wherein the method further comprises: The weight matrix of the node connection edges in the power system network is set as W, and the matrix W satisfies: That is, where matrix W is the Hadamard product of matrix A and matrix M; the elements w ij represent the weight of the connection edge between node i and node j.

5. The method of claim 4, wherein the method further comprises: The penalty function according to the difference between the reactive power demand and the reactive power supply in the community of step S2 satisfies the following relationship: The penalty function ε(ΔQ u ) = ε(∑Q u.exp -∑Q u.sup ). where u is the community number; ε(ΔQ u ) is a step function, ε(ΔQ u ) = 1 when ΔQ u ≥ 0; ε(ΔQ u ) = 0 when ΔQ u < 0; ∑Q u.sup is the total reactive power supply in community u, ∑Q u.exp is the total reactive power demand in community u, and ΔQ u = ∑Q u.exp - ∑Q u.sup is the difference between the supply and demand of reactive power in community u.

6. The method of claim 5, wherein the method further comprises: The reactive imbalance degree set in Step S2 according to the ratio of the reactive demand to the reactive supply in the community satisfies the following relationship: reactive imbalance degree where ∑Q u.sup is the total reactive supply within community u, ∑Q u.exp is the total reactive demand within community u.

7. The method of claim 6, wherein the method further comprises: The community structure quantitative evaluation index modularization function Q' of step S2 satisfies the following relationship: where w ij denotes the weight of the edge connecting node i and node j; m is the total weight of all connected edges; k i , k j are the total weight of the edges connected to node i and node j, respectively; c i , c j are the community numbers of node i and node j, when c i =c j , δ(c i , c j ) = 1, otherwise δ(c i , c j ) = 0; C1, C2, C3 are constant coefficients; δ(u) is used to limit whether the reactive power supply and demand belong to community u, when the reactive power supply and demand belong to community u, δ(u) = 1, otherwise δ(u) = 0.

8. The method of claim 7, wherein the method further comprises: The step S3 of arranging and combining the community partitions, updating the community structure quantitative evaluation index modularization function Q', and selecting the community partition with the maximum modularization function Q' value as the final reactive power partitioning result comprises the following detailed steps: Step S31: Each node in the power system network is divided into a community respectively, and the initial value Q'0 of the defined modularization function is calculated; Step S32: traverse the elements in matrix W, if w ij ≠0, then combine the communities i and j, and make the new community formed as a new node, and update the weight W of the corresponding connection edge; Step S33: The reactive power supply and demand of the new community are calculated, and the community structure quantitative evaluation index modularization function Q' of the new community is calculated; Step S34: The incremental value ΔQ' of the updated community quantitative evaluation index modularization function is calculated, the value of the community structure quantitative evaluation index modularization function after this update is subtracted from the value of the community structure quantitative evaluation index modularization function after the last update, and the value of ΔQ' is stored in the processing queue; Step S35: Steps S32-S34 are repeated until all elements in the matrix W are traversed; Step S36: The processing queue is traversed, if there is ΔQ' > 0 in the processing queue, the community combination scheme with the maximum ΔQ' is selected, a new community is obtained by merging, the weight W of the corresponding connection edge is updated, the matrix W and the total reactive power supply and demand in the new community are updated, and the new community is regarded as a new node, if there is no ΔQ' > 0 in the processing queue, the current community combination scheme is selected.

9. The method of claim 8, wherein the method further comprises: The step S32 combines and merges the community i and the community j, so that a new community is formed and regarded as a new node, and the weight W of the corresponding connecting edge is updated; including the following steps: the connecting edge between the new node and other nodes outside the new community is the connecting edge between the original community i and the community j and the nodes outside the new community, if the original two communities are connected with the same node outside the new community, then they are merged into the same edge, and the weight of the connecting edge is the cumulative weight of the corresponding edges of the communities i and j.

Images