Virtual grid division method for promoting source network load storage multi-element cooperation

Abstract

The application discloses a virtual grid division method for improving source-grid-load-storage multi-element cooperation, and the method comprises the following steps: initializing a virtual grid division of a power grid network topology structure, forming a corresponding grid adjacency matrix for different virtual grid divisions; using a generalized node search algorithm to traverse node information of the virtual grid division, and obtaining division parameters of the virtual grid division; based on the division parameters of the current virtual grid division, solving a pre-constructed grid division optimization model, calculating a grid division target function adaptive value and a grid division result, wherein the target function in the grid division optimization model considers virtual grid communication cost, virtual grid power supply rate, virtual grid power interaction and virtual grid node control dimension; when it is determined that the target function adaptive value corresponding to the current virtual grid division is smaller than the target function adaptive value corresponding to the last virtual grid division, saving the current virtual grid division result.

Description

Technical Field

[0001] This invention relates to the field of microgrid optimization technology, specifically to a virtualized grid partitioning method for improving multi-element synergy between power generation, grid, load, and storage. Background Technology

[0002] Replacing traditional fossil fuel power generation with new energy sources to address the energy crisis and environmental pollution has become a primary objective of the power industry. Recently, distributed renewable energy generation has developed rapidly, and the increasing penetration of clean energy distributed power sources such as wind and solar power in distribution networks is an inevitable trend. The randomness and volatility of distributed power sources, along with the increasingly complex and flexible controllable resources of distribution networks, make active distribution network operation and control extremely complex. To reduce the control dimensions and difficulty of active distribution network operation and control, the distribution network, which contains a large number of distributed power sources, flexible loads, energy storage, and other flexible and controllable resources, is divided into virtual grids. This forms a control pattern of local control within the grid and coordinated interactive control between grids, thus transforming a complex and high-dimensional distribution network operation and control problem into the operation and control problem of multiple relatively simple and low-dimensional grid-partitioned small systems. Simultaneously, this can promote the local consumption of renewable energy sources such as wind and solar power, improve the utilization rate of wind and solar power, and allow the grid to operate off-grid in the event of a distribution network failure while ensuring necessary power quality. This grid-based autonomous and inter-grid coordinated distribution network control strategy has become an important topic.

[0003] Traditional power distribution network zoning typically only considers active power, using electrical distances constructed based on active power sensitivity as the basis for zoning. For example, Chinese invention patent application CN115099480A discloses a planning method for dividing power supply grids based on electrical topology relationships. This method involves predicting saturated annual load, substation placement, and 10kV feeder outgoing line planning; planning the target power distribution network structure and determining electrical topology relationships; performing grid division and verification until the verification passes. Chinese invention patent application CN113222472A discloses a power distribution network grid optimization zoning method and system based on a load clustering algorithm. This method clusters load points according to their spatial location, selects the substation closest to the load center point as the primary power supply station for each load point, selects the primary power supply station corresponding to the load center point closest to the load center point as the backup power supply station for each load point, and divides the load points into grid zones based on the correspondence between load points, primary power supply stations, and backup power supply stations.

[0004] Therefore, it is necessary to comprehensively consider distributed generation nodes, energy storage nodes, and load demand response nodes in the distribution network, taking into account the overall cost of virtual grid partitioning, and balancing virtual grid power supply rate, power interaction between virtual grids, and virtual grid node control dimensions. A virtual grid partitioning objective function model should be established, and a virtual grid partitioning strategy should be proposed. This provides a foundation for grid partitioning to promote local consumption of distributed generation, improve the renewable energy consumption rate, and achieve local autonomy and collaborative interaction control strategies between virtual grids in the distribution network. Summary of the Invention

[0005] The technical problem to be solved by this invention is how to provide a distribution network virtual grid partitioning method that comprehensively takes into account distributed power generation absorption, load demand response, energy storage support, and multi-level network topology coordination.

[0006] The present invention solves the above-mentioned technical problems through the following technical means:

[0007] This invention proposes a virtualized grid partitioning method to improve multi-source coordination of source, grid, load, and storage, the method comprising:

[0008] Initialize the virtual grid partitions of the power grid network topology, and form corresponding grid adjacency matrices for different virtual grid partitions;

[0009] A generalized node search algorithm is used to traverse the node information of the virtual grid partition to obtain the partitioning parameters of the virtual grid partition;

[0010] Based on the current virtual mesh partitioning parameters, the pre-constructed mesh partitioning optimization model is solved to calculate the fitness value of the mesh partitioning objective function and the mesh partitioning result. The objective function in the mesh partitioning optimization model considers the virtual mesh communication cost, virtual mesh power supply rate, power interaction between virtual meshes, and virtual mesh node control dimension.

[0011] When the fitness value of the objective function corresponding to the current virtual mesh partition is determined to be less than the fitness value of the objective function corresponding to the previous virtual mesh partition, the result of the current virtual mesh partition is saved.

[0012] This invention employs a generalized node search algorithm to traverse the node information of virtual grid partitions, obtaining the partitioning parameters of the virtual grid partitions. A grid partitioning optimization model is then used to solve for these parameters. The objective function of this model integrates virtual grid communication costs, virtual grid power supply rate, power interaction between virtual grids, and virtual grid node control dimensions to find the optimal partitioning method from various virtual grid partitioning scenarios. This provides a grid partitioning foundation for control strategies that promote local consumption of distributed power sources, improve the renewable energy consumption rate, and achieve local autonomy and collaborative interaction between virtual grids in the distribution network. It helps enhance the complementary and coordinated power generation of the source-grid-load-storage system within the region, promotes the local consumption capacity of renewable energy, and improves source-load matching, virtual grid power supply rate, and economic efficiency.

[0013] Furthermore, the virtual grid partitioning of the initial power grid network topology, for different virtual grid partitions, forms a corresponding grid adjacency matrix, including:

[0014] Based on the branch node information in the power grid network topology, the network connection status is recorded. For node i, we have:

[0015]

[0016] In the formula: E is the set of all nodes in the power grid network topology, M and N are the sets of adjacent nodes and non-adjacent nodes of node i, respectively. If nodes ij have a topological connection, then A ij =1, otherwise, A ik =0;

[0017] The power grid network topology is divided into branches l using a virtual mesh, where i and j are the head and tail nodes of branch l. Then A ij =0, record Q represents the total set of branches segmented by the virtual grid partition;

[0018] For different grid partition scenarios of the power grid network topology, a corresponding grid adjacency matrix is ​​formed.

[0019] Furthermore, the step of using a generalized node search algorithm to traverse the node information of the virtual grid partition and obtain the partitioning parameters of the virtual grid partition includes:

[0020] Obtain the set of connection lines between virtual grids in the virtual grid partition. l= { l pq , p,q∈H}, where H is the set of grid partitions;

[0021] A generalized node search algorithm is used to traverse the node information of the virtual mesh partition to obtain the virtual mesh partitioning parameters {P}.i Q i ,P DGi Q DGi V i ,node q ,m pq ,P lpq Q lpq ,P BESSi ,P DRi}, where node q m pq These represent the communication decision node number of the virtual mesh q and the communication distance between virtual meshes p and q, respectively. i Q i These represent the active power and reactive power of node i, respectively; P DGi Q DGi These represent the active and reactive power of the distributed power source at node i, respectively; P BESSi P DRi V represents the energy storage battery at node i and the active power of the load demand response, respectively. i P is the voltage at node i; lpq Q lpq These represent the active and reactive power of the tie lines between virtual grids p and q, respectively.

[0022] Furthermore, in the mesh partitioning optimization model, the objective function is expressed as follows:

[0023] min O=K1O1+K2O2+K3O3+K4O4

[0024] In the formula: K1 is the weighting coefficient of the virtual grid partitioning cost index, K2 is the weighting coefficient of the virtual grid power supply rate index, K3 is the weighting coefficient of the power interaction index between virtual grids, K4 is the weighting coefficient of the control dimension index, O1 is the objective function of the economic cost of virtual grid partitioning, O2 is the objective function of the virtual grid self-regulation index, O3 is the objective function of the power interaction between distribution network virtual grids, and O4 is the objective function of the control dimension of distribution network virtual grid;

[0025] The constraints of the objective function include virtual grid power supply reliability constraints, power grid power flow constraints, and distribution network virtual grid operation safety and stability constraints.

[0026] Furthermore, the formula for the objective function O1 of the virtual mesh partitioning economic cost is expressed as follows:

[0027]

[0028] In the formula: P l Let n be the average unit price of the data transmission line. cl d represents the total number of virtual grids in the distribution network. yLet P be the length of the data transmission line in the virtual grid y of the distribution network. c C represents the average unit price of the virtual grid control unit in the distribution network. DR Cost of interaction between demand response load;

[0029] The objective function O2 of the virtual grid self-regulation index is expressed as follows:

[0030]

[0031] In the formula: p g g is the power supply rate parameter for the virtual grid of the distribution network. y Let g be the power supply rate of the virtual grid y of the distribution network. ref Design reference values ​​for the power supply rate of the virtual grid of the distribution network;

[0032] The objective function O3 for power interaction between virtual grids in the distribution network is expressed as follows:

[0033]

[0034] In the formula: P Q P P These represent the reactive power and active power interaction coefficients, respectively; H is the set of virtual grid tie lines in the distribution network; P... lpq For the virtual grid branch of the distribution network l pq The meritorious trend, Q lpq For the virtual grid branch of the distribution network l pq The reactive power flow is defined as follows: T is a scheduling window, and Day is a scheduling cycle.

[0035] The objective function O4 for the control dimension of the distribution network virtual grid is expressed as follows:

[0036] O4 = p w *max(n y y = 1, 2, ..., N y

[0037] In the formula: P w n represents the control dimension coefficient of the virtual grid for the distribution network. y The total number of control nodes in the virtual mesh y.

[0038] Furthermore, the formula for calculating the power supply rate of the virtual mesh is:

[0039]

[0040] In the formula: P i,t,DG Let P be the active power of node i in DG at time t. + i,t,E ,P - i,t,ELet P be the active power emitted and absorbed by the energy storage device at node i at time t. i,t,load Let ΔP be the active power consumed by the rigid load node i at time t. i,t,DR n represents the change in load demand response adjustment. dg n represents the number of DGs in the virtual grid. + e n - e n represents the number of nodes in the virtual grid that generate and absorb power for energy storage devices. load n is the number of rigid load nodes in the virtual mesh. dr This represents the number of nodes in the virtual grid that handle the demand response load.

[0041] Furthermore, the constraints of the objective function include:

[0042] Virtual mesh power supply reliability constraints:

[0043] g y ≥g min

[0044] Where: g y For the power supply rate of the virtual mesh y, g min Lower limit for virtual grid power supply rate;

[0045] Power flow constraints in virtual distribution grids:

[0046]

[0047]

[0048]

[0049]

[0050]

[0051]

[0052] In the formula: P i Q i P i DG Q i DG These represent the active load, reactive load, active output of DG, and reactive output of node i, respectively; r ij x ij I ij P ij Q ijThese represent the branch resistance, reactance, current, active power, and reactive power, respectively; A is the set of distribution network nodes; R is the set of connected nodes in the distribution network topology; N DG A set of controllable distributed power nodes; P ki Q ki These represent the active power injected from branch ki into node i, and v. i v j Let be the voltages at nodes i and j, respectively. Let represent the reactive power and active power of DG at node i at time t. Let N be the minimum and maximum active power of the DG at node i, respectively. DG This refers to the set of DG (Distributed Generation) nodes in the distribution network. These are the minimum and maximum reactive power values ​​of the DG at node i, respectively.

[0053] Constraints on the safe and stable operation of the virtual grid of the power distribution network:

[0054]

[0055]

[0056] In the formula, v i v i,min v i,max These represent the voltage, minimum voltage, and maximum voltage at node i in the distribution network, respectively. ij I ij,max These are the branch current ij and the maximum current value, respectively.

[0057] Furthermore, the formula for the demand response load interaction cost is expressed as:

[0058]

[0059] In the formula: C t 0 C t The electricity prices before and after implementing demand response are ΔP. t Let P be the load change after demand response at time t, α be the load demand response interaction cost coefficient, and P be the load change after demand response at time t. t 0 and P t These represent the load before and after the implementation of the demand response policy.

[0060] Furthermore, the constraints of the objective function also include:

[0061] Power balance constraints of energy storage batteries:

[0062] -P dis,i,max ≤P bat,i (t)≤P ch,i,max

[0063] In the formula: P ch,i,max P dis,i,max P represents the maximum charging power and maximum discharging power of the energy storage battery at node i, respectively. bat,i (t) represents the charging and discharging power at time t;

[0064] State of charge constraints for energy storage batteries:

[0065] E i (t)=E0-∫(P dis,i (t) / η dis,i )dt+∫(η ch,i P ch,i (t))dt

[0066] SOC i (t)=E i (t) / E max,i

[0067] SOC min,i ≤SOC(t)≤SOC max,i

[0068] In the formula: E i (t), E0, E max These represent the real-time battery capacity, initial battery capacity, and maximum battery capacity, respectively, η. ch,i η dis,i These represent charging efficiency and discharging efficiency, respectively, and SOC. i (t), SOC max SOC min These represent the real-time state of charge, the maximum state of charge, and the minimum state of charge, respectively.

[0069] Furthermore, when determining that the fitness value of the objective function corresponding to the current virtual mesh partition is greater than or equal to the fitness value of the objective function corresponding to the previous virtual mesh partition, the method further includes:

[0070] Adjust the grid adjacency matrix;

[0071] Determine whether all virtual grid partition scenes have been traversed;

[0072] If so, then end the virtual mesh generation;

[0073] If not, then the generalized node search algorithm is used to traverse the node information of the virtual grid partition again to obtain the partitioning parameters of the virtual grid partition.

[0074] The advantages of this invention are:

[0075] (1) This invention obtains the partitioning parameters of the virtual grid partition by traversing the node information of the virtual grid partition using a generalized node search algorithm, and solves the partitioning parameters of the virtual grid partition using a grid partitioning optimization model. The objective function in the grid partitioning optimization model integrates the virtual grid communication cost, virtual grid power supply rate, power interaction between virtual grids and virtual grid node control dimensions to find the optimal partitioning method from each virtual grid partitioning scenario. This provides a grid partitioning basis for promoting the local consumption of distributed power sources, improving the new energy consumption rate, realizing the local autonomy between virtual grids in the distribution network, and the control strategy of inter-grid collaborative interaction. It helps to improve the complementary and coordinated power of multiple sources, grids, loads and storage in the region, promote the local consumption capacity of renewable energy, and improve the source-load matching degree, virtual grid power supply rate and economy.

[0076] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0077] Figure 1 This is a flowchart illustrating a virtualized grid partitioning method for improving multi-source coordination of source, grid, load, and storage proposed in an embodiment of the present invention.

[0078] Figure 2 This is a schematic diagram of the overall process of a virtualized grid partitioning method for improving multi-source coordination of source, grid, load and storage proposed in an embodiment of the present invention;

[0079] Figure 3 This is an instanced system diagram of a virtualized grid partitioning method for improving multi-source coordination of source, grid, load, and storage in one embodiment of the present invention;

[0080] Figure 4 This is an instanced partitioning result diagram of a virtualized grid partitioning method for improving multi-source coordination of source, grid, load, and storage in one embodiment of the present invention;

[0081] Figure 5 This is an example of the active power interaction curve between various virtual grids in one embodiment of the present invention;

[0082] Figure 6 This is an embodiment of the present invention showing the reactive power interaction curve and voltage fluctuation curve between various virtual grids;

[0083] Figure 7 This refers to the power supply rate and control dimension of each virtual grid in one embodiment of the present invention;

[0084] Figure 8 This is the power supply rate of each virtual grid without load demand response in one embodiment of the present invention;

[0085] Figure 9This is a schematic diagram of virtual mesh optimization in one embodiment of the present invention;

[0086] Figure 10 This is a schematic diagram of the adjacency matrix of a virtual mesh network topology change in one embodiment of the present invention. Detailed Implementation

[0087] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0088] like Figure 1 As shown, an embodiment of the present invention proposes a virtualized grid partitioning method to improve multi-source coordination of source, grid, load, and storage. The method includes the following steps:

[0089] S10. Initialize the virtual grid partitions of the power grid network topology, and form corresponding grid adjacency matrices for different virtual grid partitions;

[0090] S20. Use a generalized node search algorithm to traverse the node information of the virtual grid partition and obtain the partitioning parameters of the virtual grid partition;

[0091] S30. Based on the partitioning parameters of the current virtual mesh partition, solve the pre-constructed mesh partitioning optimization model, calculate the fitness value of the mesh partitioning objective function and the mesh partitioning result, wherein the objective function in the mesh partitioning optimization model considers the virtual mesh communication cost, virtual mesh power supply rate, power interaction between virtual meshes and virtual mesh node control dimension;

[0092] S40. When it is determined that the fitness value of the objective function corresponding to the current virtual mesh partition is less than the fitness value of the objective function corresponding to the previous virtual mesh partition, save the current virtual mesh partition result.

[0093] This embodiment employs a generalized node search algorithm to traverse the node information of virtual grid partitions, obtaining the partitioning parameters of the virtual grid partitions. A grid partitioning optimization model is then used to solve for these parameters. The objective function of this model integrates virtual grid communication costs, virtual grid power supply rate, power interaction between virtual grids, and virtual grid node control dimensions to find the optimal partitioning method from various virtual grid partitioning scenarios. This provides a grid partitioning foundation for control strategies that promote local consumption of distributed power sources, improve the renewable energy consumption rate, and achieve local autonomy and collaborative interaction between virtual grids in the distribution network. It helps enhance the complementary and coordinated power generation of the source-grid-load-storage system within the region, promotes the local consumption capacity of renewable energy, and improves source-load matching, virtual grid power supply rate, and economic efficiency.

[0094] In one embodiment, step S10: initializing the virtual grid partitioning of the power grid network topology, and forming a corresponding grid adjacency matrix for different virtual grid partitions, includes the following steps:

[0095] S11. Based on the branch node information in the power grid network topology, record the network connection status. For node i, we have:

[0096]

[0097] In the formula: E is the set of all nodes in the power grid network topology, M and N are the sets of adjacent nodes and non-adjacent nodes of node i, respectively. If nodes ij have a topological connection, then A ij =1, otherwise, A ik =0;

[0098] S12. The power grid network topology is divided into branches l using a virtual mesh, where i and j are the head and tail nodes of branch l. Then A ij =0, record Q represents the total set of branches segmented by the virtual grid partition;

[0099] S13. For different grid partition scenarios of the power grid network topology, form corresponding grid adjacency matrices.

[0100] It should be noted that, as Figure 9 As shown, the virtual grid architecture is related to the virtual grid's interconnection lines. In the power grid, the adjacency matrix can be used to describe the relationship between branches and nodes. In order to reflect the influence of the virtual grid interconnection lines on the virtual grid partitioning, the virtual grid interconnection lines are set as decision variables, and the adjacency matrix is ​​continuously adjusted to obtain the final virtual grid.

[0101] In one embodiment, step S20: using a generalized node search algorithm to traverse the node information of the virtual grid partition and obtain the partitioning parameters of the virtual grid partition, includes the following steps:

[0102] S21. Obtain the set of connection lines between virtual grids in the virtual grid partition. l= { l pq , p,q∈H}, where H is the set of grid partitions;

[0103] S22. Use a generalized node search algorithm to traverse the node information of the virtual mesh partition and obtain the virtual mesh partitioning parameters {P}. i Q i ,P DGi Q DGi V i ,node q ,m pq ,P lpq Q lpq ,P BESSi ,P DRi}, where node q m pq These represent the communication decision node number of the virtual mesh q and the communication distance between virtual meshes p and q, respectively. i Q i These represent the active power and reactive power of node i, respectively; P DGi Q DGi These represent the active and reactive power of the distributed power source at node i, respectively; P BESSi P DRi V represents the energy storage battery at node i and the active power of the load demand response, respectively. i P is the voltage at node i; lpq Q lpq These represent the active and reactive power of the tie lines between virtual grids p and q, respectively.

[0104] It should be noted that this embodiment relies on the adjacency matrix, a decision variable representing the network topology, and combines the generalized first search algorithm to traverse the grid partitioning optimization information, continuously updating the network adjacency matrix, and optimizing the solution according to the above-mentioned objective function and constraints, finally obtaining the virtualized grid partitioning optimization result.

[0105] In one embodiment, in step S30, the objective function in the mesh partitioning optimization model is expressed as:

[0106] min O=K1O1+K2O2+K3O3+K4O4

[0107] In the formula: K1 is the weighting coefficient of the virtual grid partitioning cost index, K2 is the weighting coefficient of the virtual grid power supply rate index, K3 is the weighting coefficient of the power interaction index between virtual grids, K4 is the weighting coefficient of the control dimension index, O1 is the objective function of the economic cost of virtual grid partitioning, O2 is the objective function of the virtual grid self-regulation index, O3 is the objective function of the power interaction between distribution network virtual grids, and O4 is the objective function of the control dimension of distribution network virtual grid;

[0108] The constraints of the objective function include virtual grid power supply reliability constraints, power grid power flow constraints, and distribution network virtual grid operation safety and stability constraints.

[0109] It should be noted that, as Figure 8 As shown, this embodiment comprehensively considers the communication cost of the virtual grid, takes into account the power supply rate of the virtual grid, the power interaction between virtual grids, and the control dimension of virtual grid nodes, establishes a grid partitioning optimization model, and proposes a virtual grid partitioning strategy, which can accurately partition the virtual grid of the distribution network.

[0110] In one embodiment, the objective function O1 for the economic cost of virtual mesh partitioning is expressed as follows:

[0111]

[0112] In the formula: P l Let n be the average unit price of the data transmission line. cl d represents the total number of virtual grids in the distribution network. y Let P be the length of the data transmission line in the virtual grid y of the distribution network. c C represents the average unit price of the virtual grid control unit in the distribution network. DR Cost of demand response to load interaction.

[0113] The objective function O2 of the virtual grid self-regulation index is expressed as follows:

[0114]

[0115] In the formula: p g g is the power supply rate parameter for the virtual grid of the distribution network. y Let g be the power supply rate of the virtual grid y of the distribution network. ref Design reference value for power supply rate of virtual grid of distribution network.

[0116] Traditional distribution network zoning typically only considers active power, using electrical distances constructed based on active power sensitivity as the basis for zoning. Reactive power has a significant impact on voltage fluctuations at nodes within the virtual distribution network grid. Distributed generation (DG) within the grid acts as a reactive power source, supporting voltage at load nodes. The strength of reactive power interaction supporting load node voltage is used as an optimization index for grid zoning. Simultaneously, a virtual distribution network control scheme is formulated based on the grid zoning results. This not only improves the local absorption capacity of DG and increases resource utilization but also stabilizes voltage fluctuations at grid load nodes. Based on this, the formula for the power interaction objective function O3 between the virtual distribution network grids is expressed as follows:

[0117]

[0118] In the formula: P Q P P These represent the reactive power and active power interaction coefficients, respectively; H is the set of virtual grid tie lines in the distribution network; P... lpq For the virtual grid branch of the distribution network l pq The meritorious trend, Q lpq For the virtual grid branch of the distribution network l pq The reactive power flow is defined as follows: T is a scheduling window, and Day is a scheduling cycle.

[0119] To achieve local control of the distribution network virtual grid, the multi-node, multi-control variable distribution network is divided into multiple virtual grid subsystems, realizing local grid autonomy. Based on this, the formula for the objective function O4 of the distribution network virtual grid control dimension is expressed as follows:

[0120] O4 = p w *max(n y y = 1, 2, ..., N y

[0121] In the formula: P w n represents the control dimension coefficient of the virtual grid for the distribution network. y The total number of control nodes in the virtual mesh y.

[0122] In one embodiment, the power supply rate of the virtual mesh is calculated using the following formula:

[0123]

[0124] In the formula: P i,t,DG Let P be the active power of node i in DG at time t. + i,t,E ,P - i,t,E Let P be the active power emitted and absorbed by the energy storage device at node i at time t. i,t,loadLet ΔP be the active power consumed by the rigid load node i at time t. i,t,DR n represents the change in load demand response adjustment. dg n represents the number of DGs in the virtual grid. + e n - e n represents the number of nodes in the virtual grid that generate and absorb power for energy storage devices. load n is the number of rigid load nodes in the virtual mesh. dr This represents the number of nodes in the virtual grid that handle the demand response load.

[0125] It should be noted that the power supply rate of the virtual grid in the distribution network reflects the load support role of the active power grid nodes within the grid and is one of the indicators for improving the local absorption rate of distributed generation (DG). When a fault occurs in the distribution network, the virtual grid operates as an island, ensuring the safe and stable operation of both the distribution network and the virtual grid. During steady-state operation of the distribution network, the support role of DG for node loads within the grid can promote the local absorption of wind and solar power, reduce wind and solar curtailment, and ensure the power supply quality of the grid. Therefore, the power supply rate of the grid needs to be fully considered when dividing the virtual grid.

[0126] In one embodiment, the constraints of the objective function include:

[0127] (1) Virtual grid power supply reliability constraints:

[0128] g y ≥g min

[0129] Where: g y For the power supply rate of the virtual mesh y, g min This sets the lower limit for the power supply rate of the virtual grid.

[0130] It should be noted that when a fault occurs in the distribution network, the distribution network relay protection device will activate, and the distribution network virtual grid will operate in isolation. In order to ensure the power supply quality of critical load nodes in the virtual grid, this embodiment constructs virtual grid power supply reliability constraints.

[0131] (2) By introducing second-order cone programming to relax voltage and current auxiliary variables, the power flow constraints of the optimized distribution network virtual mesh are obtained:

[0132]

[0133]

[0134]

[0135]

[0136]

[0137]

[0138] In the formula: P i Q i P i DG Q i DG These represent the active load, reactive load, active output of DG, and reactive output of node i, respectively; r ij x ij I ij P ij Q ij These represent the branch resistance, reactance, current, active power, and reactive power, respectively; A is the set of distribution network nodes; R is the set of connected nodes in the distribution network topology; N DG A set of controllable distributed power nodes; P ki Q ki These represent the active power injected from branch ki into node i, and v. i v j Let be the voltages at nodes i and j, respectively. Let represent the reactive power and active power of DG at node i at time t. Let N be the minimum and maximum active power of the DG at node i, respectively. DG This refers to the set of DG (Distributed Generation) nodes in the distribution network. These are the minimum and maximum reactive power values ​​of the DG at node i, respectively.

[0139] (3) Constraints on the safe and stable operation of the virtual grid of the distribution network:

[0140]

[0141]

[0142] In the formula, v i v i,min v i,max These represent the voltage, minimum voltage, and maximum voltage at node i in the distribution network, respectively. ij I ij,max These are the branch current ij and the maximum current value, respectively.

[0143] In one embodiment, the formula for the demand response load interaction cost is expressed as:

[0144]

[0145] In the formula: C t 0 C t The electricity prices before and after implementing demand response are ΔP.t Let P be the load change after demand response at time t, α be the load demand response interaction cost coefficient, and P be the load change after demand response at time t. t 0 and P t These represent the load before and after the implementation of the demand response policy.

[0146] It should be noted that the load demand response model is constructed using the electricity-price elasticity coefficient. According to economic principles, the relationship between changes in electricity price and changes in load is determined by the elasticity coefficient ε. ij It can be represented as:

[0147]

[0148] Where: ΔP i and ΔC j This refers to the load changes and electricity price changes following the implementation of the demand response policy; P i 0 and C j 0 These represent the load and electricity price before the implementation of the demand response policy, respectively; i,j represent a dispatch window, and E represents a dispatch cycle.

[0149] The change in load nodes after implementing price-based demand response can be expressed as:

[0150]

[0151] The final price-based demand response mathematical model can be expressed as:

[0152]

[0153] In the formula: P i The load at time i after the implementation of demand response.

[0154] It should be noted that the load demand response model is constructed using the electricity-price elasticity coefficient. According to economic principles, the relationship between changes in electricity prices and changes in load is represented by the elasticity coefficient ε. ij For mathematical models of distributed power sources, including wind power, photovoltaic and energy storage systems, linear relaxation modeling is performed using the Big-M method and methods such as introducing 0-1 variables. For quadratic constraints such as capacity-power, the polyhedral relaxation method is used for linearization.

[0155] In one embodiment, the constraints on the objective function further include:

[0156] The power balance constraint of energy storage batteries only needs to consider whether the charging and discharging power exceeds the limit:

[0157] -P dis,i,max ≤Pbat,i (t)≤P ch,i,max

[0158] In the formula: P ch,i,max P dis,i,max P represents the maximum charging power and maximum discharging power of the energy storage battery at node i, respectively. bat,i (t) represents the charging and discharging power at time t;

[0159] State of charge constraints for energy storage batteries:

[0160] E i (t)=E0-∫(P dis,i (t) / η dis,i )dt+∫(η ch,i P ch,i (t))dt

[0161] SOC i (t)=E i (t) / E max,i

[0162] SOC min,i ≤SOC(t)≤SOC max,i

[0163] In the formula: E i (t), E0, E max These represent the real-time battery capacity, initial battery capacity, and maximum battery capacity, respectively, η. ch,i η dis,i These represent charging efficiency and discharging efficiency, respectively, and SOC. i (t), SOC max SOC min These represent the real-time state of charge, the maximum state of charge, and the minimum state of charge, respectively.

[0164] In one embodiment, such as Figure 2 As shown, when the fitness value of the objective function corresponding to the current virtual mesh partition is determined to be greater than or equal to the fitness value of the objective function corresponding to the previous virtual mesh partition, the method further includes the following steps:

[0165] Adjust the grid adjacency matrix;

[0166] Determine whether all virtual grid partition scenes have been traversed;

[0167] If so, then end the virtual mesh generation;

[0168] If not, then the generalized node search algorithm is used to traverse the node information of the virtual grid partition again to obtain the partitioning parameters of the virtual grid partition.

[0169] In this embodiment, a distribution network system with 33 nodes is constructed. The distribution network includes a wind power system, a photovoltaic system, an energy storage system, and a load demand response node, and is established as follows: Figure 3 The simulation model diagram is shown, and the relevant parameters are shown in Tables 1, 2, and 4. Time-of-use electricity prices corresponding to load demand response are set, as shown in Table 4. A breadth-first search algorithm is used to traverse the virtual grid node information to obtain the node branch decision variables required for virtual grid partitioning. Based on the network topology of the virtual grid, a network node information adjacency matrix is ​​established, and the parameters of the virtual grid partitioning objective function are set, as shown in Table 5.

[0170] Table 1 Basic Parameters of DG

[0171]

[0172] Table 2 Energy Storage Parameters

[0173]

[0174] Table 3 Price elasticity coefficient of price-based demand response load

[0175]

[0176] Table 4 Time-of-use Electricity Prices

[0177]

[0178] Table 5 Basic Parameters of DG

[0179]

[0180] The integration of diverse and flexible resources in various virtual grids, such as Figure 4 As shown. From Figure 4 As can be seen, the modified IEEE 33-node distribution network is divided into 5 virtual grids according to the virtual grid partitioning method proposed in this paper. The grid integration results show that the 5 virtual grids effectively integrate distributed power sources, load demand response, and energy storage, demonstrating better source-load matching and creating a better environment for autonomous control within the distribution network grid and coordinated control between grids.

[0181] like Figure 4 As shown, the number of virtual grid control nodes represents the control dimension of the virtual grid. Before grid partitioning, the distribution network has 33 control nodes. Figure 4As shown, after virtual grid partitioning, the control node dimensions of the five virtual grids are 9, 3, 6, 7, and 8, respectively. Through virtual grid partitioning, local control is achieved based on the partitioning results, instead of the centralized control of the 33 nodes in the distribution network. This significantly reduces the order of data processing for the distribution network by the five low-dimensional local controllers, greatly improving calculation speed. Simultaneously, the five virtual grid local controllers interact collaboratively, achieving real-time comprehensive collaborative control of the distribution network while reducing control complexity. This is beneficial for promoting the absorption rate of distributed power sources (wind and solar) in the distribution network, as well as for power flow distribution control and power quality control.

[0182] Active interactions between grids, such as Figure 5 As shown, the power interaction between grids is relatively small compared to the total power demand of the distribution network, thus satisfying the active power support function of the grid's power sources for the grid load. The reactive power interaction between grids has a significant impact on the voltage stability of nodes within each grid. Figure 6 After grid division, the reactive power interaction between grids and the voltage fluctuation of typical central points are observed. After division, the voltage of nodes within each grid is stably distributed around the rated voltage. This satisfies the grid reactive power interaction index requirements and the voltage support effect of grid reactive power sources on nodes within the grid.

[0183] Power supply rate and control dimensions of each virtual grid, such as Figure 7 As shown, the power supply rates of each virtual mesh are 100%, 94.86%, 80.64%, 81.29%, and 91.75%, respectively. Figure 8 The power supply rates of the virtual grid without demand response nodes are 100%, 91.45%, 79.48%, 76.16%, and 74.46%, respectively. The power supply rates of the virtual grid are significantly lower than those of the distribution grid with demand response nodes. This indicates that through the regulation of demand response loads and the support of energy storage in the virtual grid, the power of the virtual grid can be kept balanced at all times, reducing the waste of distributed resources, improving the absorption of distributed power sources, and also giving the virtual grid a certain degree of anti-interference capability.

[0184] It should be noted that this embodiment, by considering the communication cost of the virtual grid, and taking into account the power supply rate of the virtual grid, the power interaction between virtual grids, and the control dimension of virtual grid nodes, establishes a virtual grid partitioning objective function model and proposes a virtual grid partitioning strategy. The communication cost and load demand response interaction cost of virtual grid partitioning are established; a virtual grid autonomy rate index is proposed, reflecting the load support role of active power grid nodes within the grid, and is one of the indicators for improving the local absorption rate of distributed generation (DG). With the goal of reducing the voltage fluctuation rate of nodes within the grid, the interaction of active and reactive power between virtual grids is considered; to achieve local control of the distribution network virtual grid, the multi-node, multi-control variable distribution network is divided into multiple virtual grid mini-systems to achieve local grid autonomy, and based on this, a control dimension objective function for the distribution network virtual grid is constructed. Finally, DG operation constraints, load DR constraints, network topology constraints, and power flow constraints are constructed to seek the optimal solution for the virtualized grid partitioning optimization model.

[0185] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the 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.

[0186] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0187] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A virtualized grid partitioning method for improving multi-source coordination of source, grid, load, and storage, characterized in that, The method includes: Initialize the virtual grid partitions of the power grid network topology, and form corresponding grid adjacency matrices for different virtual grid partitions; A generalized node search algorithm is used to traverse the node information of the virtual grid partition to obtain the partitioning parameters of the virtual grid partition; Based on the current virtual mesh partitioning parameters, a pre-constructed mesh partitioning optimization model is solved to calculate the fitness value of the mesh partitioning objective function and the mesh partitioning results. The objective function in the mesh partitioning optimization model considers virtual mesh communication cost, virtual mesh power supply rate, power interaction between virtual meshes, and virtual mesh node control dimensions. The formula for the virtual mesh partitioning economic cost objective function is expressed as follows: In the formula: P l This represents the average unit price of the data transmission line. n cl This represents the total number of virtual grids in the distribution network. d y Virtual grid for distribution network y The length of the data transmission line, P c The average unit price of the virtual grid control unit for the distribution network. C DR Cost of interaction between demand response load; Virtual grid power supply rate objective function O The formula for 2 is expressed as: In the formula: p g For the power supply rate parameters of the virtual grid of the distribution network, g y Virtual grid for distribution network y Power supply rate, g ref Design reference values ​​for the power supply rate of the virtual grid of the distribution network; Power interaction objective function between virtual grids of distribution network O The formula for 3 is expressed as: In the formula: P Q , P P These are the interaction coefficients for reactive power and active power, respectively. H For the set of virtual grid tie lines of the distribution network, P lpq For virtual grid branches of the distribution network l pq The meritorious trend, Q lpq For virtual grid branches of the distribution network l pq The unproductive current, T For a scheduling window, Day One scheduling cycle; Objective function of virtual grid control dimension of distribution network O The formula for 4 is expressed as: In the formula: P w For the control dimension coefficient of the virtual grid of the distribution network, n y For virtual grid y The total number of control nodes; When the fitness value of the objective function corresponding to the current virtual mesh partition is determined to be less than the fitness value of the objective function corresponding to the previous virtual mesh partition, the result of the current virtual mesh partition is saved.

2. The virtualized grid partitioning method for improving multi-source coordination of source, grid, load, and storage as described in claim 1, characterized in that, The virtual grid partitioning of the initial power grid network topology, for different virtual grid partitions, forms a corresponding grid adjacency matrix, including: Based on the branch node information in the power grid network topology, the network connection status is recorded, and for each node... i ,have: A ij = 1 ,A ik = 0 , { i∈E,j∈M,k∈N} In the formula: E The set of all nodes in the power grid network topology. M, N They are respectively with nodes i The set of adjacent nodes and the set of non-adjacent nodes, if the node ij If there is a topological relationship A ij = 1. Otherwise, A ik = 0; The power grid network topology is divided into branches using a virtual mesh. l segmentation, i, j branch road l The head and tail nodes, then A ij =0 ,Record l∈Q , Q The total set of branches is divided into virtual grid partitions; For different grid partition scenarios of the power grid network topology, a corresponding grid adjacency matrix is ​​formed.

3. The virtualized grid partitioning method for improving multi-source coordination of source, grid, load, and storage as described in claim 1, characterized in that, The step of using a generalized node search algorithm to traverse the node information of the virtual grid partition and obtain the partitioning parameters of the virtual grid partition includes: Obtain the set of connection lines between virtual grids in the virtual grid partition. l= { l pq , p,q∈H }, H A set of grid partitions; A generalized node search algorithm is used to traverse the node information of the virtual mesh partition to obtain the virtual mesh partitioning parameters { P i , Q i , P DGi , Q DGi , V i , node q , m pq , P lpq , Q lpq , P BESSi , P DRi },in, node q ，m pq Virtual mesh q Communication decision node number and virtual grid p, q Communication distance between them P i , Q i They are nodes i Active power and reactive power; P DGi , Q DGi They are nodes i Active and reactive power of distributed power sources; P BESSi , P DRi They are nodes i Energy storage batteries and active power for load demand response; V i For nodes i The voltage; P lpq , Q lpq Virtual mesh p , q The active and reactive power of the inter-connection line.

4. The virtualized grid partitioning method for improving multi-source coordination of source, grid, load, and storage as described in claim 1, characterized in that, In the aforementioned mesh partitioning optimization model, the objective function is expressed as follows: In the formula: K 1 represents the weighting coefficient for the cost index of virtual mesh partitioning. K 2 represents the weighting coefficient for the virtual grid power supply rate index. K 3 represents the weighting coefficient for the power interaction index between virtual grids. K 4 represents the weighting coefficient for the control dimension indicators. O 1 represents the objective function for the economic cost of virtual mesh partitioning. O 2 represents the objective function for the virtual grid power supply rate index. O 3 represents the objective function for power interaction between virtual grids in the distribution network. O 4 represents the objective function for the virtual grid control dimension of the distribution network; The constraints of the objective function include virtual grid power supply reliability constraints, power grid power flow constraints, and distribution network virtual grid operation safety and stability constraints.

5. The virtualized grid partitioning method for improving multi-source coordination of source, grid, load, and storage as described in claim 1, characterized in that, The formula for calculating the power supply rate of the virtual grid is: In the formula: P i,t,DG For nodes i exist t The active power of DG at any given time. P + i,t,E , P - i,t,E For nodes i Energy storage devices t Active power constantly being generated and absorbed. P i,t,load Rigid load node i exist t The active power consumed by the load at any given moment. P i,t,DR This represents the change in load demand response adjustment. n dg The number of DGs in the virtual mesh. n + e , n - e This represents the number of nodes in the virtual grid that generate and absorb power for energy storage devices. n load The number of rigid load nodes in the virtual mesh. n dr This represents the number of nodes in the virtual mesh that handle demand response loads.

6. The virtualized grid partitioning method for improving multi-source coordination of source, grid, load, and storage as described in claim 4, characterized in that, The constraints of the objective function include: Virtual mesh power supply reliability constraints: In the formula: g y For virtual grid y Power supply rate, g min Lower limit for virtual grid power supply rate; Power flow constraints in virtual distribution grids: In the formula: P i , Q i , P i DG , Q i DG They are nodes i Active load, reactive load DG Efforts that yield results and effort that yields no results; r ij , x ij , I ij , P ij , Q ij Branch roads ij Resistance, reactance, current, active power, and reactive power; A For the set of distribution network nodes; R It is the set of connected nodes in the distribution network topology. N DG A set of controllable distributed power supply nodes; P ki , Q ki They are respectively from branch roads ki To the node i Injected active power, v i , v j They are nodes i , j voltage, , For nodes i exist t The reactive power and active power of DG at any given time. , They are nodes i The minimum and maximum active power of the DG. This refers to the set of DG (Distributed Generation) nodes in the distribution network. , , respectively, nodes i Minimum and maximum reactive power of the DG, and safety and stability constraints for the operation of the distribution network virtual grid: In the formula, v i , v i,min , v i,max Distribution network nodes i Voltage, minimum voltage, maximum voltage I ij , I ij,max Branch roads ij Current and maximum current, N For nodes i The set of non-adjacent nodes.

7. The virtualized grid partitioning method for improving multi-source coordination of source, grid, load, and storage as described in claim 1, characterized in that, The formula for the demand response load interaction cost is expressed as follows: In the formula: C t 0 , C t These are the electricity prices before and after the implementation of demand response. P t for t The change in load after demand response at any given moment. α This represents the interaction cost coefficient for load demand response. P t 0 and P t These represent the load before and after the implementation of the demand response policy.

8. The virtualized grid partitioning method for improving multi-source coordination of source, grid, load, and storage as described in claim 4, characterized in that, The constraints on the objective function also include: Power balance constraints of energy storage batteries: In the formula: P ch,i,max , P dis,i,max They are nodes i The maximum charging power and maximum discharging power of the energy storage battery P bat,i ( t )for t The charging and discharging power at any given moment; State of charge constraints for energy storage batteries: In the formula: E i ( t ), E 0、 E max These are the real-time battery level, initial battery level, and maximum battery level, respectively. η ch,i , η dis,i These are charging efficiency and discharging efficiency, respectively. SOC i ( t ), SOC max , SOC min These represent the real-time state of charge, the maximum state of charge, and the minimum state of charge, respectively. For nodes i Energy storage batteries in t Discharge power at any given time For nodes i Energy storage batteries in t The charging power at any given time.

9. The virtualized grid partitioning method for improving multi-source coordination of source, grid, load, and storage as described in any one of claims 1 to 8, characterized in that, When determining that the fitness value of the objective function corresponding to the current virtual mesh partition is greater than or equal to the fitness value of the objective function corresponding to the previous virtual mesh partition, the method further includes: Adjust the grid adjacency matrix; Determine whether all virtual grid partition scenes have been traversed; If so, then end the virtual mesh generation; If not, then the generalized node search algorithm is used to traverse the node information of the virtual grid partition again to obtain the partitioning parameters of the virtual grid partition.

Images