Park network reconstruction method and device for improving grid-connected flexibility

Abstract

The invention belongs to the technical field of park network reconstruction, and particularly relates to a park network reconstruction method and device for improving grid-connected flexibility, and the method comprises the steps: S1, constructing a park grid-connected flexibility calculation model; s2, constructing a network reconstruction model oriented to grid connection flexibility improvement; s3, a column and constraint generation algorithm Camp is adopted; and solving the two-stage adaptive robust optimization reconstruction model through CG. Through combination of network reconstruction and distributed resource flexibility aggregation, the problems of low flexibility evaluation and aggregation efficiency, high network reconstruction optimization complexity, lack of a joint optimization mechanism and the like in the prior art are effectively solved; a two-stage adaptive robust optimization model is adopted, a microgrid topological structure is optimized in the first stage, depolymerization feasibility is verified in the second stage, the mutual influence of the two stages is fully considered, the power flexibility of the public coupling point of the park is improved to the maximum extent, and the flexibility limitation caused by internal complex electrical constraints is overcome.

Description

Technical Field

[0001] The present invention belongs to the technical field of campus network reconstruction, and in particular relates to a campus network reconstruction method and device for improving grid connection flexibility. Background Art

[0002] With the widespread access to distributed energy resources (DER), the industrial park, as a system where multiple distributed energy sources coexist in the "source-grid-load-storage" model, can provide flexibility to the distribution network through the Point of Common Coupling (PCC). For example, virtual power plants can integrate decentralized flexibility resources through market mechanisms to solve problems such as power balance, peak regulation, and insufficient voltage regulation resources in the power system, which is of great significance for improving the economic efficiency and safety of power system operation.

[0003] Due to complex electrical constraints within the park, such as voltage safety, line capacity, and load balancing, the park's grid-connected flexibility has not been fully realized. Despite the adjustability of distributed resources, their flexibility cannot be fully aggregated at the common coupling point under various operating conditions, limiting the park's external regulation capabilities. Furthermore, given the wide distribution of distributed resources, effectively consolidating their flexibility at the park's common coupling point and providing a scalable power range has become a major challenge facing current park technology.

[0004] Against this backdrop, network reconfiguration technology within a park has become a key solution to this problem. By adjusting the network topology within the park and optimizing current flow paths, network reconfiguration improves the overall operational efficiency and flexibility of the power grid. Numerous studies have demonstrated that network reconfiguration can effectively unleash the potential of distributed resources and improve the flexibility of a park's grid connection.

[0005] Existing network reconfiguration research focuses primarily on the following areas: On the one hand, existing flexibility aggregation methods (such as polyhedron aggregation) provide theoretical support for evaluating campus flexibility, but these methods fail to incorporate the flexibility changes brought about by network reconfiguration. On the other hand, while some network reconfiguration technologies based on voltage security and power flow constraints have been studied, most of them ignore the actual impact of network reconfiguration on campus flexibility. The main shortcomings of existing network reconfiguration research are: Flexibility assessment and aggregation are inefficient. Existing aggregation methods, such as polyhedrons or scenario simulations, often require extensive computation and scenario enumeration to assess distributed resource flexibility, resulting in high computational complexity and poor real-time performance. Due to the high uncertainty of distributed resource states and the external environment, the model must consider numerous constraint combinations, making it difficult to solve quickly. Attempts to simplify the model have resulted in insufficient accuracy and significant deviations in flexibility estimates.

[0006] Network reconfiguration optimization is highly complex. Network reconfiguration must simultaneously satisfy multiple constraints, including voltage safety, line capacity, and power balance. The exponential growth of topological combinations leads to a massive optimization problem, making it difficult to find a globally optimal solution in a short period of time. Previous approaches have mostly relied on heuristic or metaheuristic algorithms, but these methods are prone to falling into local optimality, are difficult to tune parameters, and have unstable convergence speeds, making real-time reconfiguration difficult in practical applications.

[0007] Due to the lack of a joint optimization mechanism, existing research often optimizes flexibility assessment or network reconstruction separately, lacking a systematic method to couple the two, making it difficult to quantify the effect of reconstruction on improving aggregation flexibility.

[0008] Therefore, how to combine flexibility aggregation technology with network reconstruction methods to propose a more efficient park flexibility improvement solution has become a hot issue in the field of park research. Summary of the Invention

[0009] In view of the above deficiencies in the existing technology, the purpose of the present invention is to provide a campus network reconstruction method and device for improving grid-connected flexibility, comprehensively considering network reconstruction and distributed resource flexibility aggregation, making full use of the flexibility of all distributed resources, optimizing the campus grid-connected flexibility, and at the same time improving the security within the campus and the coordinated adjustment capabilities with mutually coupled systems.

[0010] To achieve the above objectives, the present invention provides a campus network reconstruction method for improving grid connection flexibility, comprising the following steps: S1. Build a park grid-connected flexibility calculation model. The process is as follows: S1.1. Select the park's public coupling point (PCC) as the power flexibility aggregation point. Aggregate the flexibility of all distributed resources in the park to the PCC. Determine the network topology and time scale of the park. Within the determined network topology and time scale, collect all operating states at the PCC to form the park's grid-connected power flexibility. S1.2. Flexibility modeling of distributed photovoltaic and distributed energy storage systems; S1.3. Considering network constraints, the LinDistFlow power flow model taking voltage distribution into account is used to establish a campus power flow model. S1.4. Construct a grid-connected flexibility calculation method. Write all state variables at a certain moment in time in vector form. Build a comprehensive park system based on the park power flow model, flexibility modeling, and defined vectors. Establish a two-stage adaptive robust optimization model to calculate the park PCC active power flexibility. S2. Build a network reconfiguration model for improving grid-connection flexibility. The process is as follows: S2.1. Based on the characteristics of the closed-loop design and open-loop operation of the park, two integer variables are introduced to establish the park reconstruction model; S2.2, with the goal of maximizing PCC flexibility, the campus network reconfiguration problem for improving flexibility is formulated as a two-stage adaptive robust optimization reconfiguration model; S3, using the column and constraint generation algorithm C&CG to solve the two-stage adaptive robust optimization reconstruction model, the process is: S3.1. Decompose the two-stage adaptive robust optimization reconstruction problem into a main problem and sub-problems; S3.2. By iteratively solving the main problem and sub-problems, the optimal solution is obtained to achieve campus network reconstruction.

[0011] As a preferred solution of the present invention, in S1.1, the distributed resources include distributed photovoltaic and distributed energy storage systems. Assume that the park grid-connected power flexibility is , Expressed as: (1); Where, 、 They represent the upper and lower boundaries of the park grid-connected power flexibility at time t, respectively, t=1, 2,…, T, and T is the total number of moments.

[0012] As a preferred embodiment of the present invention, in S1.2, the distributed photovoltaic model is: (2); (3); (4); Where, 、 、 are the active power injected into node i by the photovoltaic generator at time t and its upper and lower limits; 、 、 are the reactive power injected into node i by the photovoltaic generator at time t and its upper and lower limits; is the apparent power capacity of the photovoltaic generator at node i; The distributed energy storage system is modeled as: (5); (6); (7); (8); Where, 、 、 are the power of the energy storage device at node i at time t, t-1, and T respectively; is the amount of energy storage device on node i at the starting time; is the energy loss coefficient of the energy storage device at node i as it changes with time; is the time variation; is the charging and discharging power of the energy storage device at node i at time t; 、 are the upper and lower limits of the charging and discharging power of the energy storage device at node i, respectively; 、 are the upper and lower limits of the energy storage device at node i, respectively.

[0013] As a preferred solution of the present invention, in S1.3, the park flow model is expressed as: (9); (10); (11); (12); Where, 、 are the active and reactive power of the load at node j at time t respectively; 、 are the active and reactive power injected into node j by the photovoltaic generator at time t; is the charging and discharging power of the energy storage device at node j at time t; 、 are the active and reactive power injected from node i into the branch of node j respectively; 、 are the active and reactive power flowing out of node j at time t respectively; It represents a branch or connection from node j to other node m; 、 are the voltage amplitudes of the first node i and the last node j of branch ij respectively; is the resistance of branch ij; is the reactance of branch ij; is the upper limit of the capacity of branch ij.

[0014] As a preferred solution of the present invention, in S1.4, all state variables at time t are written as vectors In the form of: (13); Where, the superscript T represents transposition; Based on the park flow model, flexibility modeling and vector , the park comprehensive system is written as formula (14)-(17): (14); (15); (16); (17); Where, 、 are the active and reactive power of the park PCC respectively; G, F, , C is The coefficient matrix of g, h, , c is a given parameter, including 、 、 、 、 、 、 、 、 ; l is an index variable used to traverse all second-order cone constraints, and L is the total number of second-order cone constraints; Introduce the following constraints: (18); (19); (20); Where E, B, and D are coefficient matrices; b and d are given parameters; is the vector form of all state variables at time 1; about The objective function is expressed as: (twenty one); Where, represents the objective function; For aggregated For any operating point in the , regulate and manage flexible resources through scheduling instructions to meet: (twenty two); Therefore, the constraints of the park grid-connected flexibility calculation model are written as: (twenty three); A two-stage adaptive robust optimization model is established. First, the uncertain parameters are introduced. To express ,Pick is a decimal between 0 and 1, equivalent to and An interpolation is taken between: (twenty four); definition Belongs to uncertainty set I: (25); The two-stage adaptive robust optimization model shown in Equations (26) and (27) is established to calculate the active power flexibility of the park PCC: (26); (27); Where, Represents a transposed all-1 vector; for The decision variable vector under ; 、 、 、 At time 1, t-1, t, and t respectively The decision variable vector under ; Represents the norm of a vector.

[0015] As a preferred solution of the present invention, in S2.1, two integer variables are introduced. and : (28); in, Indicates the switch state of branch ij, When it is 1, it means that the branch ij is in the conducting state. When it is 0, it means that branch ij is disconnected; Represents the parent node-node association matrix between node j and node i. When node j is the parent node of node i is 1, otherwise 0; According to graph theory and spanning tree theory, the campus topology has the following characteristics: each node except the common coupling point has only one parent node, and the root node has no parent node, which can be expressed as: (29); Where, Represents the parent node-node association matrix between node j and the root node; represents the set of neighbor nodes of node i; Represents the parent node-node association matrix between node i and node j. If there is a branch ij, then node i is the parent node of node j, that is, , and at the same time, it is necessary to ensure that node j will not be the parent node of node i, that is, , expressed as: (30); When branch ij is disconnected, it is necessary to ensure the active power on branch ij and reactive power is 0, so formula (12) can be rewritten as: (31); If branch ij is disconnected, due to the constraint of formula (31), and If it is limited to 0, Equation (11) will become Equation (32), which means that the voltage amplitudes at both ends of the unconnected branch ij are forced to be equal. This is incorrect. Therefore, the large M method is introduced to rewrite Equation (11) into Equations (33)-(35): (32); (33); (34); (35); Where, is an auxiliary variable; M is a constant in the big M method; is the time t ; Based on formula (21), the park reconstruction model is established and expressed as: (36); (37).

[0016] As a preferred solution of the present invention, in S2.2, the problem of campus network reconstruction for improving flexibility is described as: aggregating the power flexibility of distributed resources and optimizing the topology structure to obtain the maximum power flexibility of the campus PCC. For any operating point in the maximum power flexibility of the PCC, there is a scheduling solution after deaggregation. Deal with it; Therefore, the outer layer of the grid-connected flexibility calculation model and park reconstruction model Combined, a two-stage adaptive robust optimization reconstruction model is established as shown in Equation (38) and Equation (39) to optimize the active power flexibility of the park PCC: (38); (39).

[0017] As a preferred solution of the present invention, in S3.1, for the main problem, according to the solution mechanism of the C&CG algorithm, the first-stage max problem of the two-stage adaptive robust optimization reconstruction model is written as Equation (40) and Equation (41), where Equation (40) is the objective function and Equation (41) is the constraint condition: (40); (41); Where, Indicates the scenario generated by the sub-problem at time t at the kth iteration. is the given initial value; k is the current iteration number, K is the total number of iterations; For the scene Adaptive solution, so in K scenarios, K adaptive ; 、 、 Respectively represent the decision variable vectors at time t-1, t, and T during the k-th iteration; After solving the main problem, the switch variables that satisfy all the constraints of the main problem will be obtained , the upper limit of the park grid-connected power flexibility at time t and the lower boundary and a , for the subproblem, it is formulated as a min-max optimization problem including the objective function of formula (42) and the constraints of formula (43): (42); (43); Since Equation (42) is a min-max double-layer optimization problem, in order to transform the inner max problem into a min problem and merge it with the outer min problem, it is necessary to find the dual of the inner max problem. Let the Lagrange multipliers of the dual variables at time t in Equation (43) be , based on the strong duality theory, Equation (42) will be transformed into: (44); Where, is the cost coefficient at time t; n is the index of the distributed resource, and N is the set of distributed resources; is the power flexibility coefficient associated with distributed resource n at time t; is a binary decision variable that represents the enabled state of distributed resource n at time t, 1 for enabled and 0 for disabled; The value of is simplified to a binary variable, as shown in Equation (45), and the constraint condition will be restated as shown in Equation (46): (45); (46); Where, is the constraint matrix of distributed resource n; is the Lagrange multiplier corresponding to the distributed resource n at time t; 、 are the dual variable vectors at time t and t-1, respectively, connecting the constraints of the time series; is the Lagrange multiplier vector; To constrain the boundaries; 、 、 is a non-negative Lagrange multiplier, ensuring that the dual variables satisfy the non-negativity constraint; There is a nonlinear term in Equation (44) , so the big M method is used to transform it into: (47); Where, 、 、 、 These are all intermediate variables generated during the linearization process; is the reference maximum power at time t; is the reference minimum power at time t; is the cost factor of power flexibility at time t; is the flexibility adjustment of the nth distributed resource at time t; when hour, , ;when hour, , , expressed as: (48); Where, is the positive auxiliary variable at time t, used to represent the upper bound or positive deviation of power flexibility; It is a negative auxiliary variable at time t, indicating the lower bound or negative deviation of power flexibility.

[0018] As a preferred solution of the present invention, in S3.2, the C&CG algorithm is adopted and the Gurobi solver is used to solve the subproblem. If the objective function value still does not meet the convergence condition after the subproblem is solved, a new variable will be generated and Equation (49) will be added to the main problem to continue iteration: (49); Where, 、 、 They represent the decision variable vectors at time 1, t-1, and t in the K+1th iteration respectively.

[0019] A campus network reconstruction device for improving grid connection flexibility includes a memory, a processor, and a computer program stored in the memory and capable of running on the processor. The above method is implemented by executing the computer program by the processor.

[0020] The beneficial effects of the present invention are: By combining network reconfiguration with distributed resource flexibility aggregation, this invention effectively addresses existing issues such as low flexibility assessment and aggregation efficiency, high network reconfiguration optimization complexity, and a lack of a joint optimization mechanism. A two-stage adaptive robust optimization model is employed: the first stage optimizes the microgrid topology, and the second stage verifies the feasibility of deaggregation. This model fully considers the mutual influence between the two, maximizing the power flexibility of the park's common coupling point and overcoming flexibility limitations caused by complex internal electrical constraints. A grid-connected power flexibility calculation model is also constructed to characterize distributed resource flexibility, introducing uncertainty parameters to ensure operational feasibility. This improves the park's grid-connected flexibility and internal safety and stability, and enhances its ability to coordinate regulation with the coupled system.

[0021] In terms of model solving, this invention employs a column and constraint generation algorithm to solve a two-stage adaptive robust optimization problem, decomposing the problem into a main problem and subproblems for iterative solution. Experimental data from a local park demonstrates that the proposed method improves the park's PCC power flexibility by nearly 28% compared to the pre-optimization level. Compared to traditional network reconfiguration methods that aim to minimize network loss, this method effectively improves the park's grid-connected power flexibility, validating the effectiveness and superiority of the proposed solution. It provides an innovative and efficient solution for park network reconfiguration and flexibility enhancement, with significant technical advantages and application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 It is a schematic diagram of the process of the present invention; Figure 2 1. It is a schematic diagram of the calculation mechanism of grid-connected flexibility in an embodiment of the present invention; Figure 3 It is a topological diagram of the campus during the verification process of the present invention; Figure 4 This is a daily trend diagram of load and photovoltaic power generation during the verification process of the present invention; Figure 5 This is a comparison chart of PCC power flexibility optimization results in two time periods during the verification process of the present invention; Figure 6This is a diagram showing the results before and after optimization of power flexibility in scenario 1 during the verification process of the present invention; Figure 7 This is a diagram showing the results before and after power flexibility optimization in scenario 2 during the verification process of the present invention. DETAILED DESCRIPTION

[0023] The embodiments of the present invention are further described below with reference to the accompanying drawings: Example 1: Figure 1 As shown, a campus network reconstruction method for improving grid connection flexibility includes the following steps: S1. Build a park grid-connected flexibility calculation model. The process is as follows: S1.1. Select the park's public coupling point (PCC) as the power flexibility aggregation point. Aggregate the flexibility of all distributed resources in the park to the PCC. Determine the park's network topology and time scale (1 hour). Within the determined network topology and time scale, collect all operating states at the PCC to form the park's grid-connected power flexibility. S1.2. Flexibility modeling of distributed photovoltaic and distributed energy storage systems; S1.3. Considering network constraints, the LinDistFlow power flow model taking voltage distribution into account is used to establish a campus power flow model. S1.4. Construct a grid-connected flexibility calculation method. Write all state variables at a certain moment in time in vector form. Build a comprehensive park system based on the park power flow model, flexibility modeling, and defined vectors. Establish a two-stage adaptive robust optimization model to calculate the park PCC active power flexibility. S2. Build a network reconfiguration model for improving grid-connection flexibility. The process is as follows: S2.1. Based on the characteristics of the closed-loop design and open-loop operation of the park, two integer variables are introduced to establish the park reconstruction model; S2.2, with the goal of maximizing PCC flexibility, the campus network reconfiguration problem for improving flexibility is formulated as a two-stage adaptive robust optimization reconfiguration model; S3, using the column and constraint generation algorithm C&CG to solve the two-stage adaptive robust optimization reconstruction model, the process is: S3.1. Decompose the two-stage adaptive robust optimization reconstruction problem into a main problem and sub-problems; S3.2. By iteratively solving the main problem and sub-problems, the optimal solution is obtained to achieve campus network reconstruction.

[0024] Flexibility aggregation describes the flexibility of all distributed resources in the park to the common coupling point, under the premise of ensuring voltage safety within the park and overloading of branches. Flexibility resources include distributed photovoltaics, distributed energy storage, distributed wind power, adjustable loads, electric vehicles, generators, etc. The essence is: projecting a high-dimensional space determined by the state variables of the flexibility resources onto the injection space of the state variables of the park's common coupling point in Euclidean space. Grid-connected power flexibility is essentially also the feasible domain of power at the common coupling point. For dispatching, it is also the dispatchable range of the power system, and for the power system itself, it is also the operating safety domain of the power system. The park's grid-connected power flexibility gives the maximum grid-connected flexibility range of the power system that meets safety constraints, which facilitates the evaluation and optimization of gateway power flexibility.

[0025] In S1.1, distributed resources include distributed photovoltaic and distributed energy storage systems. The park grid-connected power flexibility is set to , expressed as a high-dimensional polyhedron using formula (1), Expressed as: (1); Where, 、 They represent the upper and lower boundaries of the park grid-connected power flexibility at time t, respectively, t=1, 2,…, T, and T is the total number of moments (value is 24).

[0026] In S1.2, distributed photovoltaics are modeled as: (2); (3); (4); Where, 、 、 are the active power injected into node i by the photovoltaic generator at time t and its upper and lower limits; 、 、 are the reactive power injected into node i by the photovoltaic generator at time t and its upper and lower limits; is the apparent power capacity of the photovoltaic generator at node i; Equations (2) and (3) are the upper and lower limit constraints of the active and reactive power of the photovoltaic motor, respectively. Equation (4) is the capacity constraint of the photovoltaic generator at node i at time t.

[0027] The distributed energy storage system is modeled as: (5); (6); (7); (8); Where, 、 、 are the power of the energy storage device at node i at time t, t-1, and T respectively; is the amount of energy storage device on node i at the starting time; is the energy loss coefficient of the energy storage device at node i as it changes with time; is the time variation; is the charging and discharging power of the energy storage device at node i at time t; 、 are the upper and lower limits of the charging and discharging power of the energy storage device at node i, respectively; 、 are the upper and lower limits of the energy storage device at node i, respectively.

[0028] Equations (5) and (6) are the energy conservation constraints at time t = T and 1 ≤ t ≤ T - 1, respectively; Equations (7) and (8) are the upper and lower limit constraints on the charging and discharging power and the upper and lower limit constraints on the energy storage device at node i at time t, respectively.

[0029] Considering that voltage is the main operational limitation of a high proportion of distributed power generation feeding into the park, the LinDistFlow power flow model taking into account voltage distribution is adopted. This model has the advantages of linear processing, high-precision voltage estimation, flexibility and scalability, and has important application value in power system analysis and optimization.

[0030] In S1.3, the park power flow model is expressed as: (9); (10); (11); (12); Where, 、 are the active and reactive power of the load at node j at time t respectively; 、 are the active and reactive power injected into node j by the photovoltaic generator at time t; is the charging and discharging power of the energy storage device at node j at time t; 、 are the active and reactive power injected from node i into the branch of node j respectively; 、 are the active and reactive power flowing out of node j at time t respectively; It represents a branch or connection from node j to other node m; 、 are the voltage amplitudes of the first node i and the last node j of branch ij respectively; is the resistance of branch ij; is the reactance of branch ij; is the upper limit of the capacity of branch ij.

[0031] Equations (9) and (10) are the node active and reactive power balance equations, respectively; Equation (11) is the branch voltage drop equation, and Equation (12) is the branch capacity constraint.

[0032] In S1.4, all state variables at time t are written as vectors (equivalent to a scheduling plan) in the form of: (13); Where, the superscript T represents transposition; Based on the park flow model, flexibility modeling and vector , the park comprehensive system is written as formula (14)-(17): (14); (15); (16); (17); Where, 、 are the active and reactive power of the park PCC respectively; G, F, , C is The coefficient matrix of g, h, , c is a given parameter, including 、 、 、 、 、 、 、 、 ; l is an index variable used to traverse all second-order cone constraints, and L is the total number of second-order cone constraints; Equations (14) and (15) are the active and reactive power balance equations of the park PCC, respectively; Equation (16) contains all second-order cone constraints such as Equations (4) and (12); and Equation (17) contains other network constraints.

[0033] Since the constraints in the battery energy storage system model have time coupling characteristics, in order to simplify the expression of the coefficient matrix and facilitate problem solving, the following constraints need to be introduced: (18); (19); (20); Where E, B, and D are coefficient matrices; b and d are given parameters, also including 、 、 、 、 、 、 、 、 ; is the vector form of all state variables at time 1; Equations (18) and (19) are model expressions of Equation (6); Equation (20) is a model expression of Equation (5).

[0034] about The objective function is expressed as: (twenty one); Where, represents the objective function; Power flexibility at the point of common coupling With the maximum interior approximation to the true solution, for the aggregate For any operating point in the , regulate and manage flexible resources through scheduling instructions to meet: (twenty two); Therefore, the constraints of the park grid-connected flexibility calculation model are written as: (twenty three); The mathematical perspective of the model is to project a high-dimensional space determined by the state variables of DER (Distributed Energy Resources) into the low-dimensional space where the state variables of the park PCC are located in the Euclidean space. Any operating point in its low-dimensional space can be mapped to the high-dimensional space determined by the DER state variables, which satisfies the disaggregation feasibility. Figure 2 Schematic diagram of the grid-connected flexibility calculation mechanism.

[0035] Therefore, the above problem can be formulated as a two-stage adaptive robust optimization model. To verify whether the maximum power flexibility obtained by aggregation satisfies the feasibility of deaggregation, randomly selecting some operating points for verification is not conservative. It is obviously infeasible to verify every operating point within the maximum power flexibility space of the campus PCC.

[0036] A two-stage adaptive robust optimization model is established. First, the uncertain parameters are introduced. To express ,Pick is a decimal between 0 and 1, equivalent to and An interpolation is taken between them to ensure that every operating point within the maximum power flexibility space of the park common coupling point can be obtained: (twenty four); definition Belongs to uncertainty set I: (25); The two-stage adaptive robust optimization model shown in Equations (26) and (27) is established to calculate the active power flexibility of the park PCC: (26); (27); Where, Represents a transposed all-1 vector; for The decision variable vector under ; 、 、 、 At time 1, t-1, t, and t respectively The decision variable vector under ; Represents the norm of a vector.

[0037] The optimization objective (26) in the first stage is to optimize the upper and lower limits of the maximum power flexibility, in order to find the maximum power flexibility of the park's public coupling point; in the second stage, Two-level optimization ensures that the uncertainty variables When the worst scenario of uncertainty set I is obtained, there is a flexible resource scheduling solution To achieve this, we can ensure the feasibility of deaggregation. The constraint formula (27) contains all network topology constraints and aggregation and deaggregation feasibility constraints.

[0038] In S2.1, two integer variables are introduced and : (28); in, Indicates the switch state of branch ij, When it is 1, it means that the branch ij is in the conducting state. When it is 0, it means that branch ij is disconnected; Represents the parent node-node association matrix between node j and node i. When node j is the parent node of node i is 1, otherwise 0; According to graph theory and spanning tree theory, the campus topology has the following characteristics: each node except the common coupling point has only one parent node, and the root node has no parent node, which can be expressed as: (29); Where, Represents the parent node-node association matrix between node j and the root node; represents the set of neighbor nodes of node i; Represents the parent node-node association matrix between node i and node j. If there is a branch ij, then node i is the parent node of node j, that is, , and at the same time, it is necessary to ensure that node j will not be the parent node of node i, that is, , expressed as: (30); When branch ij is disconnected, it is necessary to ensure the active power on branch ij and reactive power is 0, so formula (12) can be rewritten as: (31); If branch ij is disconnected, due to the constraint of formula (31), and If it is limited to 0, Equation (11) will become Equation (32), which means that the voltage amplitudes at both ends of the unconnected branch ij are forced to be equal. This is incorrect. Therefore, the large M method is introduced to rewrite Equation (11) into Equations (33)-(35): (32); (33); (34); (35); Where, is an auxiliary variable; M is a constant in the big M method; is the time t ; Based on formula (21), the park reconstruction model is established and expressed as: (36); (37).

[0039] Through network reconstruction, we can tap into the flexibility of new energy sources distributed at various nodes and improve the grid-connected flexibility of the park. Therefore, based on the network reconstruction optimization model and the power flexibility aggregation model, the park network reconstruction problem for improving grid-connected flexibility can be written as a two-stage adaptive robust optimization reconstruction model.

[0040] In S2.2, the campus network reconstruction problem for improving flexibility is described as: aggregating the power flexibility of distributed resources and optimizing the topology to obtain the maximum power flexibility of the campus PCC. For any operating point in the PCC maximum power flexibility, there is a scheduling solution after deaggregation. Deal with it; Therefore, the outer layer of the grid-connected flexibility calculation model and park reconstruction model Combined, a two-stage adaptive robust optimization reconstruction model is established as shown in Equation (38) and Equation (39) to optimize the active power flexibility of the park PCC: (38); (39).

[0041] The optimization objective (38) in the first phase includes the line switch state and the upper and lower limits of the maximum power feasible region. The goal is to find the optimal topology for the park PCC with maximum power flexibility. Other specific meanings are the same as those in (26). The constraint (39) includes all network topology constraints and aggregation and disaggregation feasibility constraints. Unlike the park grid-connected flexibility calculation model, (39) includes the tree topology constraints (28) to (30) that must be satisfied by network reconstruction. The network constraints must also be replaced with power flow constraints that satisfy network reconstruction optimization.

[0042] In S3.1, for the main problem, according to the solution mechanism of the C&CG algorithm, the first-stage max problem of the two-stage adaptive robust optimization reconstruction model is written as Equation (40) and Equation (41), where Equation (40) is the objective function and Equation (41) is the constraint condition: (40); (41); Where, Indicates the scenario generated by the sub-problem at time t at the kth iteration. is the given initial value; k is the current iteration number, K is the total number of iterations; For the scene Adaptive solution, so in K scenarios, K adaptive ; 、 、 Respectively represent the decision variable vectors at time t-1, t, and T during the k-th iteration; After solving the main problem, the switch variables that satisfy all the constraints of the main problem will be obtained. , the upper limit of the park grid-connected power flexibility at time t and the lower boundary and a , for the subproblem, it is formulated as a min-max optimization problem including the objective function of formula (42) and the constraints of formula (43): (42); (43); Since Equation (42) is a min-max double-layer optimization problem, in order to transform the inner max problem into a min problem and merge it with the outer min problem, it is necessary to find the dual of the inner max problem. Let the Lagrange multipliers of the dual variables at time t in Equation (43) be , based on the strong duality theory, Equation (42) will be transformed into: (44); Where, is the cost coefficient at time t; n is the index of the distributed resource, and N is the set of distributed resources; is the power flexibility coefficient associated with distributed resource n at time t; is a binary decision variable that represents the enabled state of distributed resource n at time t, 1 for enabled and 0 for disabled; The value of is simplified to a binary variable, as shown in Equation (45), and the constraint condition will be restated as shown in Equation (46): (45); (46); Where, is the constraint matrix of distributed resource n; is the Lagrange multiplier corresponding to the distributed resource n at time t (corresponding to the dual variable of the constraint); 、 are the dual variable vectors at time t and t-1, respectively, connecting the constraints of the time series; is the Lagrange multiplier vector; To constrain the boundaries; 、 、 is a non-negative Lagrange multiplier, ensuring that the dual variables satisfy the non-negativity constraint; There is a nonlinear term in Equation (44) , so the big M method is used to transform it into: (47); Where, 、 、 、 These are all intermediate variables generated during the linearization process; is the reference maximum power at time t (the upper limit of the active power at the PCC); is the benchmark minimum power at time t (the lower limit of the active power at the PCC); is the cost factor of power flexibility at time t; is the flexibility adjustment of the nth distributed resource at time t; when hour, , ;when hour, , , expressed as: (48); Where, is the positive auxiliary variable at time t, used to represent the upper bound or positive deviation of power flexibility; It is a negative auxiliary variable at time t, indicating the lower bound or negative deviation of power flexibility.

[0043] According to the above description, the subproblem will be reformulated as Equations (45)-(48), and the solution of the subproblem will be transformed into the solution of the mixed integer second-order cone problem, which can be directly solved using the Gurobi solver.

[0044] In S3.2, the C&CG algorithm is used and the Gurobi solver is used to solve the problem. If the objective function value (Equation 47) still does not meet the convergence condition after the subproblem is solved, a new variable will be generated and Equation (49) will be added to the main problem to continue the iteration: (49); Where, 、 、 They represent the decision variable vectors at time 1, t-1, and t at the K+1th iteration (K is not the strict "maximum number of iterations", but the index of the current iteration. The C&CG algorithm allows the number of iterations to grow dynamically. K+1 only represents the variable of the next iteration. The actual total number of iterations may exceed the initially set K value).

[0045] The verification process is: The feasibility of the method in this embodiment is verified by using a certain park. The topology is as follows: Figure 3As shown. The charging and discharging efficiency of the energy storage device is 0.9, the maximum charging power is 0.2MW / h, the maximum discharging power is 0.2MW / h, the initial SOC is 0.3MWh, the lower limit of the energy storage capacity is 0.15MWh, and the upper limit of the energy storage capacity is 0.8MWh; the installed capacity of the photovoltaic generator is 0.4MW. The changing trends of active and reactive loads and photovoltaic daily power generation in the park are shown as follows: Figure 4 shown.

[0046] Optimization analysis of the park's grid-connected power flexibility under two time periods: First, with the goal of optimizing the park grid-connected power flexibility within two time periods, the two-stage adaptive robust optimization model proposed in this embodiment is intuitively explained. Figure 5 The power flexibility of the campus grid connection in two time periods is shown as follows: the power flexibility of the campus grid connection before network reconstruction, the power flexibility of the campus grid connection during traditional network reconstruction with the goal of minimizing network loss, and the power flexibility of the campus grid connection after optimization using the method proposed in this embodiment. The power flexibility before network reconstruction refers to the PCC power flexibility that only considers the aggregation of flexibility resources and satisfies all security constraints.

[0047] from Figure 5 It can be seen that the aggregated PCC power flexibility is increased by 2.98 after optimization by the method proposed in this embodiment, which is nearly 28% higher than before optimization. By comparing the park grid-connected power flexibility under the traditional network reconstruction with the goal of minimizing network loss and the park grid-connected flexibility after optimization by the method proposed in this embodiment, it can be seen that the traditional network reconstruction with the goal of minimizing network loss cannot improve the park grid-connected flexibility.

[0048] Considering that the reactive power flexibility of distributed resources will affect the grid-connected power flexibility of the park, in order to further verify the effectiveness of the method proposed in this embodiment, the grid-connected power flexibility before and after optimization is simulated and analyzed for scenario one and scenario two.

[0049] Scenario 1: Optimizing the flexibility of campus grid-connected power.

[0050] Scenario 2: When the photovoltaic generator does not provide reactive power flexibility, the park's grid-connected power flexibility is optimized.

[0051] The feasible domains of the common coupling point power before and after optimization of scenario 1 and scenario 2 are: Figure 6 and Figure 7 The blue and pink areas in the figure are the feasible regions of active power in each period before and after optimization, respectively, and the purple area is the crossover area before and after optimization.

[0052] To visually analyze the results before and after optimizing active power flexibility at the point of common coupling, the power flexibility for each time period in both scenarios was summed and compared. The analysis showed that the proposed method in this example improved the park's grid-connected power flexibility by 1% in Scenario 1 and by 7% in Scenario 2.

[0053] Example 2: A campus network reconstruction device for improving grid-connected flexibility includes a memory, a processor, and a computer program stored in the memory and capable of running on the processor. The method in Example 1 is implemented by executing the computer program by the processor.

Claims

1. A campus network reconstruction method for improving grid connection flexibility, characterized in that The following steps are involved: S1. Build a park grid-connected flexibility calculation model. The process is as follows: S1.

1. Select the park's public coupling point (PCC) as the power flexibility aggregation point. Aggregate the flexibility of all distributed resources in the park to the PCC. Determine the network topology and time scale of the park. Within the determined network topology and time scale, collect all operating states at the PCC to form the park's grid-connected power flexibility. S1.

2. Flexibility modeling of distributed photovoltaic and distributed energy storage systems; S1.

3. Considering network constraints, the LinDistFlow power flow model taking voltage distribution into account is used to establish a campus power flow model. S1.

4. Construct a grid-connected flexibility calculation method. Write all state variables at a certain moment in time in vector form. Build a comprehensive park system based on the park power flow model, flexibility modeling, and defined vectors. Establish a two-stage adaptive robust optimization model to calculate the park PCC active power flexibility. S2. Build a network reconfiguration model for improving grid-connection flexibility. The process is as follows: S2.

1. Based on the characteristics of the closed-loop design and open-loop operation of the park, two integer variables are introduced to establish the park reconstruction model; S2.2, with the goal of maximizing PCC flexibility, the campus network reconfiguration problem for improving flexibility is formulated as a two-stage adaptive robust optimization reconfiguration model; S3, using the column and constraint generation algorithm C&CG to solve the two-stage adaptive robust optimization reconstruction model, the process is: S3.

1. Decompose the two-stage adaptive robust optimization reconstruction problem into a main problem and sub-problems; S3.

2. By iteratively solving the main problem and sub-problems, the optimal solution is obtained to achieve campus network reconstruction.

2. A campus network reconfiguration method for improving grid connection flexibility according to claim 1, characterized in that: In S1.1, distributed resources include distributed photovoltaic and distributed energy storage systems. Assume that the park grid-connected power flexibility is , Expressed as: （1）； Where, 、 They represent the upper and lower boundaries of the park grid-connected power flexibility at time t, respectively, t=1, 2,…, T, and T is the total number of moments.

3. A campus network reconfiguration method for improving grid connection flexibility according to claim 2, characterized in that: In S1.2, the distributed photovoltaic model is: （2）； （3）； （4）； Where, 、 、 are the active power injected into node i by the photovoltaic generator at time t and its upper and lower limits; 、 、 are the reactive power injected into node i by the photovoltaic generator at time t and its upper and lower limits; is the apparent power capacity of the photovoltaic generator at node i; The distributed energy storage system is modeled as: （5）； （6）； （7）； （8）； Where, 、 、 are the power of the energy storage device at node i at time t, t-1, and T respectively; is the amount of energy storage device on node i at the starting time; is the energy loss coefficient of the energy storage device at node i as it changes with time; is the time variation; is the charging and discharging power of the energy storage device at node i at time t; 、 are the upper and lower limits of the charging and discharging power of the energy storage device at node i, respectively; 、 are the upper and lower limits of the energy storage device at node i, respectively.

4. A campus network reconfiguration method for improving grid connection flexibility according to claim 3, characterized in that: In S1.3, the park power flow model is expressed as: （9）； （10）； （11）； （12）； Where, 、 are the active and reactive power of the load at node j at time t respectively; 、 are the active and reactive power injected into node j by the photovoltaic generator at time t; is the charging and discharging power of the energy storage device at node j at time t; 、 are the active and reactive power injected from node i into the branch of node j respectively; 、 are the active and reactive power flowing out of node j at time t respectively; It represents a branch or connection from node j to other node m; 、 are the voltage amplitudes of the first node i and the last node j of branch ij respectively; is the resistance of branch ij; is the reactance of branch ij; is the upper limit of the capacity of branch ij.

5. A campus network reconfiguration method for improving grid connection flexibility according to claim 4, characterized in that: In S1.4 above, all state variables at time t are written as vectors In the form of: （13）； Where, the superscript T represents transposition; Based on the park flow model, flexibility modeling and vector , the park comprehensive system is written as formula (14)-(17): （14）； （15）； （16）； （17）； Where, 、 are the active and reactive power of the park PCC respectively; G, F, , C is The coefficient matrix of g, h, , c is a given parameter, including 、 、 、 、 、 、 、 、 ; l is an index variable used to traverse all second-order cone constraints, and L is the total number of second-order cone constraints; Introduce the following constraints: （18）； （19）； （20）； Where E, B, and D are coefficient matrices; b and d are given parameters; is the vector form of all state variables at time 1; about The objective function is expressed as: （21）； Where, represents the objective function; For aggregated For any operating point in the , regulate and manage flexible resources through scheduling instructions to meet: （22）； Therefore, the constraints of the park grid-connected flexibility calculation model are written as: （23）； A two-stage adaptive robust optimization model is established. First, the uncertain parameters are introduced. To express ,Pick is a decimal between 0 and 1, equivalent to and An interpolation is taken between: （24）； definition Belongs to uncertainty set I: （25）； The two-stage adaptive robust optimization model shown in Equations (26) and (27) is established to calculate the active power flexibility of the park PCC: （26）； （27）； Where, represents a transposed all-1 vector; for The decision variable vector under ; 、 、 、 At time 1, t-1, t, and t respectively The decision variable vector under ; Represents the norm of a vector.

6. A campus network reconfiguration method for improving grid connection flexibility according to claim 5, characterized in that: In S2.1, two integer variables are introduced and : （28）； in, Indicates the switch state of branch ij, When it is 1, it means that the branch ij is in the conducting state. When it is 0, it means that branch ij is disconnected; Represents the parent node-node association matrix between node j and node i. When node j is the parent node of node i is 1, otherwise 0; According to graph theory and spanning tree theory, the campus topology has the following characteristics: each node except the common coupling point has only one parent node, and the root node has no parent node, which can be expressed as: （29）； Where, Represents the parent node-node association matrix between node j and the root node; represents the set of neighbor nodes of node i; Represents the parent node-node association matrix between node i and node j. If there is a branch ij, then node i is the parent node of node j, that is, , and at the same time, it is necessary to ensure that node j will not be the parent node of node i, that is, , expressed as: （30）； When branch ij is disconnected, it is necessary to ensure the active power on branch ij and reactive power is 0, so formula (12) can be rewritten as: （31）； If branch ij is disconnected, due to the constraint of formula (31), and If it is limited to 0, Equation (11) will become Equation (32), which means that the voltage amplitudes at both ends of the unconnected branch ij are forced to be equal. This is incorrect. Therefore, the large M method is introduced to rewrite Equation (11) into Equations (33)-(35): （32）； （33）； （34）； （35）； Where, is an auxiliary variable; M is a constant in the big M method; is the time t ; Based on formula (21), the park reconstruction model is established and expressed as: （36）； （37）。 7. A campus network reconfiguration method for improving grid connection flexibility according to claim 6, characterized in that: In S2.2, the problem of campus network reconstruction for flexibility improvement is described as: aggregating distributed resource power flexibility and optimizing the topology to obtain the maximum power flexibility of the campus PCC. For any operating point in the PCC maximum power flexibility, there is a scheduling solution after deaggregation. Deal with it; Therefore, the outer layer of the grid-connected flexibility calculation model and park reconstruction model Combined, a two-stage adaptive robust optimization reconstruction model is established as shown in Equation (38) and Equation (39) to optimize the active power flexibility of the park PCC: （38）； （39）。 8. A campus network reconfiguration method for improving grid connection flexibility according to claim 7, characterized in that: In S3.1, for the main problem, according to the solution mechanism of the C&CG algorithm, the first-stage max problem of the two-stage adaptive robust optimization reconstruction model is written as Equation (40) and Equation (41), where Equation (40) is the objective function and Equation (41) is the constraint condition: （40）； （41）； Where, Indicates the scenario generated by the sub-problem at time t at the kth iteration. is the given initial value; k is the current iteration number, K is the total number of iterations; For the scene Adaptive solution, so in K scenarios, K adaptive ; 、 、 Respectively represent the decision variable vectors at time t-1, t, and T during the k-th iteration; After solving the main problem, the switch variables that satisfy all the constraints of the main problem will be obtained , the upper limit of the park grid-connected power flexibility at time t and the lower boundary and a , for the subproblem, it is formulated as a min-max optimization problem including the objective function of formula (42) and the constraints of formula (43): （42）； （43）； Since Equation (42) is a min-max double-layer optimization problem, in order to transform the inner max problem into a min problem and merge it with the outer min problem, it is necessary to find the dual of the inner max problem. Let the Lagrange multipliers of the dual variables at time t in Equation (43) be , based on the strong duality theory, Equation (42) will be transformed into: （44）； Where, is the cost coefficient at time t; n is the index of the distributed resource, and N is the set of distributed resources; is the power flexibility coefficient associated with distributed resource n at time t; is a binary decision variable that represents the enabled state of distributed resource n at time t, 1 for enabled and 0 for disabled; The value of is simplified to a binary variable, as shown in Equation (45), and the constraint condition will be restated as shown in Equation (46): （45）； （46）； Where, is the constraint matrix of distributed resource n; is the Lagrange multiplier corresponding to the distributed resource n at time t; 、 are the dual variable vectors at time t and t-1, respectively, connecting the constraints of the time series; is the Lagrange multiplier vector; To constrain the boundaries; 、 、 is a non-negative Lagrange multiplier, ensuring that the dual variables satisfy the non-negativity constraint; There is a nonlinear term in Equation (44) , so the big M method is used to transform it into: （47）； Where, 、 、 、 These are all intermediate variables generated during the linearization process; is the reference maximum power at time t; is the reference minimum power at time t; is the cost factor of power flexibility at time t; is the flexibility adjustment of the nth distributed resource at time t; when hour, , ;when hour, , , expressed as: （48）； Where, is the positive auxiliary variable at time t, used to represent the upper bound or positive deviation of power flexibility; It is a negative auxiliary variable at time t, indicating the lower bound or negative deviation of power flexibility.

9. A campus network reconfiguration method for improving grid connection flexibility according to claim 8, characterized in that: In S3.2, the C&CG algorithm is used and the Gurobi solver is used to solve the subproblem. If the objective function value still does not meet the convergence condition after the subproblem is solved, a new variable will be generated and Equation (49) will be added to the main problem to continue the iteration: （49）； Where, 、 、 They represent the decision variable vectors at time 1, t-1, and t in the K+1th iteration respectively.

10. A campus network reconfiguration device for improving grid connection flexibility, characterized by: The method comprises a memory, a processor, and a computer program stored in the memory and capable of running on the processor, wherein the method according to any one of claims 1 to 9 is implemented by executing the computer program by the processor.

Images