A multi-port power grid aggregation method based on cut-plane projection

Abstract

The application belongs to the technical field of power grid dispatching, and specifically discloses a multi-port power grid aggregation method based on a cut plane projection. The application obtains constraints by modeling a multi-port power grid based on distributed energy and solves a robust optimization problem to obtain each port independent aggregation interval, which is integrated into an initial aggregation space. Then, a cut plane problem is solved to identify an infeasible region and obtain a cut plane constraint and a worst dispatch instruction. Subsequently, a feasible dispatch instruction closest to the worst dispatch instruction in the current aggregation space is found according to the cut plane constraint, and the core function of the feasible dispatch instruction is to provide a reference point for each time interval projection. The cut plane constraint is projected to each time interval space with the reference point, the aggregation space is cut and updated, and the infeasible region is gradually cut off through iteration of the process. This gradually and accurately depicts the interconnection characteristics between multi-ports, solves the aggregation difficulty problem caused by the lack of accurate description means in the prior art, and improves the aggregation reliability compared with the prior art.

Description

Technical Field

[0001] This application belongs to the field of power grid dispatching technology, specifically relating to intelligent dispatching system technology, and more specifically, to a multi-port power grid aggregation method based on cutting plane projection. Background Technology

[0002] Currently, if distributed energy resources connected to the grid structure within a certain range are regarded as a power grid, and then aggregated to form a power grid that interacts with the outside world, the number of constraints and variables that the upper-level power grid needs to consider during dispatching can be reduced, thereby greatly reducing the complexity of power system operation and improving the efficiency of distributed energy resource management.

[0003] However, grid aggregation also presents the following challenges: First, grid aggregation requires consideration of the operational characteristics of different types of distributed energy sources and network constraints, ensuring that all dispatch commands within the aggregation interval can be de-aggregated within the grid. Second, the division of the grid is uncertain. Unlike the main grid and distribution grid, which have clearly defined geographical boundaries, the grid is merely a virtual unit composed of distributed energy sources and their grid structures with physical boundaries within the distribution grid. For example... Figure 1 As shown, when it is in a distribution network, it often forms a multi-port network with ≥3 ports connected to the external power grid. How to take into account the interconnection characteristics between multiple ports in aggregation is also a current technological challenge. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this application aims to provide a multi-port power grid aggregation method based on secant plane projection, which addresses the problem of insufficient means to accurately characterize the interconnection characteristics between multiple ports in current multi-port power grid aggregation scenarios, thus making aggregation difficult.

[0005] The first aspect of this application relates to a multi-port power grid aggregation method based on cut plane projection, comprising: step S10, modeling the multi-port power grid based on distributed energy resources set in the power grid and obtaining model constraints, and then solving the aggregation problem based on the model constraints to obtain multiple independent aggregation operation intervals for multiple ports; the aggregation problem is a robust optimization problem aimed at finding the aggregation operation intervals of the multi-port power grid under all scheduling scenarios; step S20, integrating the multiple independent aggregation operation intervals into an initial aggregation space for the multi-port power grid and determining it as the current aggregation space; step S30, solving the cut plane acquisition problem based on the current aggregation space and model constraints to obtain the current cut plane constraints and the current worst-case scheduling instruction; the cut plane acquisition problem is aimed at testing the feasibility of the current aggregation space and The goal is to generate a robust optimization problem with the objective of cutting plane constraints to remove infeasible regions. The worst-case scheduling instruction is the infeasible scheduling instruction with the largest sum of power correction values ​​across all time periods among all scheduling instructions that were originally infeasible but have been modified to be feasible in the current aggregation space. Step S40: Based on the worst-case scheduling instruction and the current cutting plane constraints, obtain the currently feasible scheduling instruction that is closest to the worst-case scheduling instruction in the current aggregation space. Step S50: Under the constraints of the closest currently feasible scheduling instruction, project the current cutting plane constraints onto the current aggregation space in different time periods to cut and update the current aggregation space. Then, execute the steps of solving the cutting plane acquisition problem based on the current aggregation space and model constraints until the current aggregation space is feasible, thereby realizing multi-port power grid aggregation.

[0006] In one embodiment, the cutting plane acquisition problem is modeled as an objective function that includes model constraints and current aggregation space constraints. The max-min robust optimization problem, in which, It is a column vector consisting entirely of integers 1; This is the vector corresponding to the upward power correction value. This is the vector corresponding to the downward power correction value.

[0007] In one embodiment, step S30 includes: performing a dual transformation on the inner-level min problem of the cut plane acquisition problem to obtain the dual problem of the cut plane acquisition problem; solving the dual problem to obtain the current worst-case scheduling instruction and the solution value of the dual variable in the current aggregation space; and generating the current cut plane constraint in the current aggregation space based on the objective that the result of the dual problem is less than or equal to 0, combined with the current worst-case scheduling instruction and the solution value of the dual variable.

[0008] In one embodiment, step S40 includes: determining the non-cutting plane constraints of the current feasible scheduling instruction based on the current cutting plane constraints; determining the grid constraints of the current feasible scheduling instruction based on the model constraints; and obtaining the current feasible scheduling instruction that is closest to the current worst-case scheduling instruction in the current aggregation space based on the non-cutting plane constraints and the grid constraints.

[0009] In one embodiment, step S50 includes: decomposing and projecting the current cutting plane constraint according to different time periods to obtain independent projected cutting plane constraints for each time period; wherein, the projected cutting plane constraint for each time period ensures that its geometry is consistent with the shape of the original cutting plane in that time period, and has undergone the component of the most recent currently feasible scheduling instruction in that time period; integrating the projected cutting plane constraints of all time periods into an overall projected cutting plane constraint covering the entire time period; combining the overall projected cutting plane constraint with the current aggregate space constraint to generate a new aggregate space constraint with a reduced range, thereby completing the update of the current aggregate space.

[0010] In one embodiment, the projected cutting plane constraint for each time period ensures that its geometry is consistent with the shape of the original cutting plane in that time period, and incorporates the components of the most recent current feasible scheduling instruction within that time period. This includes: decomposing the parameters of the current cutting plane constraint into time periods based on the solution values ​​of the dual variables and the current feasible scheduling instruction, and extracting parameter slices corresponding to each time period; wherein the calculation of the parameter slices incorporates the time period components of the most recent current feasible scheduling instruction, and then performs offset adjustments, thereby constructing an independent projected cutting plane constraint for each time period.

[0011] In one embodiment, the convergence condition for the iteration process until the current aggregation space is feasible is: the result of the cutting plane problem is zero, indicating that all scheduling instructions in the current aggregation space are executable.

[0012] In a second aspect, this application provides an electronic device, comprising: at least one memory for storing a program; and at least one processor for executing the program stored in the memory, wherein when the program stored in the memory is executed, the processor is configured to execute the method described in the first aspect or any possible implementation thereof.

[0013] Thirdly, this application provides a computer-readable storage medium storing a computer program that, when run on a processor, causes the processor to perform the method described in the first aspect or any possible implementation thereof.

[0014] Fourthly, this application provides a computer program product that, when run on a processor, causes the processor to perform the method described in the first aspect or any possible implementation thereof.

[0015] Overall, the technical solutions conceived in this application have the following beneficial effects compared with the prior art:

[0016] This application models a multi-port power grid based on distributed energy resources and obtains model constraints. Then, it solves a robust optimization problem aimed at finding the aggregated operating intervals under all scheduling scenarios to obtain the independent aggregated operating intervals of each port. This provides an initial and accurate characterization of the operating characteristics of a single port, laying the foundation for aggregation. Next, these independent intervals are integrated into an initial aggregation space, which serves as the current aggregation space, to initially reflect the prototype of multi-port interconnection.

[0017] Then, by solving the cutting plane acquisition problem with the goal of testing the feasibility of the current aggregation space and generating cutting plane constraints, the current cutting plane constraints and worst-case scheduling instructions are obtained, thereby identifying infeasible regions in the aggregation space and preparing for cutting; then, based on the worst-case scheduling instructions and cutting plane constraints, the feasible scheduling instructions that are closest to the worst-case scheduling instructions in the current aggregation space are obtained, providing accurate positioning for subsequent projection;

[0018] Finally, under the constraint of the most recently feasible scheduling command, the cut plane constraint is projected onto the current aggregation space under different time periods, the current aggregation space is cut and updated, and the cut plane acquisition problem is iteratively executed until the aggregation space is feasible. This series of techniques, by iteratively removing infeasible regions and retaining feasible points, gradually and accurately characterizes the interconnection characteristics between multiple ports, solving the aggregation difficulty problem caused by the lack of accurate characterization methods in existing technologies. Compared with existing technologies, it achieves efficient and reliable aggregation of multi-port power grids, improving the feasibility of scheduling commands and the stability of power grid operation. Attached Figure Description

[0019] Figure 1 This is a conceptual diagram of the interactive power operation between a multi-port power grid and its upstream power grid, provided in an embodiment of this application.

[0020] Figure 2 This is a flowchart illustrating a multi-port power grid aggregation method based on cleaving plane projection provided in an embodiment of this application;

[0021] Figure 3 This is a topology diagram of a power grid based on a 30-node system provided in an embodiment of this application;

[0022] Figure 4 This is a daily output diagram of distributed photovoltaic power provided in the embodiments of this application;

[0023] Figure 5 This is a diagram of the total load in the power grid provided in the embodiments of this application;

[0024] Figure 6 This is the running range diagram of each of the three ports A, B, and C obtained by aggregation according to the embodiments of this application;

[0025] Figure 7It is the geometric representation diagram corresponding to the initial value of the multi-port aggregation space for each hour of the 24 hours within a day provided by the aggregation embodiment of this application;

[0026] Figure 8 This is a characterization diagram of the projection cutting plane cutting time period 10 in the first iteration provided in the embodiments of this application;

[0027] Figure 9 This is the operation interval diagram of each hour of the day considering the interconnection characteristics of ports A, B, and C, obtained by aggregation in the embodiments of this application;

[0028] Figure 10 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0030] In this application, the term "and / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent three cases: A existing alone, A and B existing simultaneously, and B existing alone. In this application, the symbol " / " indicates that the related objects are in an "or" relationship, for example, A / B means A or B.

[0031] In this application, the terms "first" and "second," etc., are used to distinguish different objects, not to describe a specific order of objects. For example, "first response message" and "second response message," etc., are used to distinguish different response messages, not to describe a specific order of response messages.

[0032] In the embodiments of this application, the terms "exemplary" or "for example" are used to indicate that something is an example, illustration, or description. Any embodiment or design that is described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design. Specifically, the use of the terms "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0033] In the description of the embodiments of this application, unless otherwise stated, "multiple" means two or more, for example, multiple processing units means two or more processing units, multiple elements means two or more elements, etc.

[0034] Currently, in multi-port power grid aggregation scenarios, the lack of precise methods to characterize the interconnection characteristics between multiple ports presents aggregation difficulties. Therefore, this application proposes a multi-port power grid aggregation method based on secant plane projection. Please refer to... Figure 2 , Figure 2 This is a flowchart illustrating a multi-port power grid aggregation method based on cleaving plane projection provided in an embodiment of this application.

[0035] In this embodiment, the multi-port power grid aggregation method based on cleaving plane projection includes steps S10 to S50.

[0036] Step S10: Model the multi-port power grid based on the distributed energy resources set in the power grid and obtain the model constraints. Then solve the aggregation problem based on the model constraints to obtain multiple independent aggregation operation intervals for multiple ports.

[0037] It should be noted that multi-port grid modeling based on distributed energy resources within the power grid refers to constructing a mathematical model that can characterize the operational characteristics of each distributed energy resource within the power grid and its coupling relationship with multi-port power. Distributed energy resources include, but are not limited to, micro gas turbines, photovoltaic power generation units, energy storage systems, and various loads. Model constraints cover the physical operational constraints of these devices, such as the ramp rate of micro gas turbines, the charging and discharging power and capacity limits of energy storage, and the output prediction range of photovoltaics, as well as network constraints of port connection points determined by the external power grid, such as voltage safety limits.

[0038] by Figure 1 For example, Figure 1 This is a conceptual diagram illustrating the interactive power operation between a multi-port power grid and its upstream power grid, as provided in an embodiment of this application. Figure 1 In the grid, the components include micro gas turbines, energy storage units, and distributed photovoltaics; This represents the active power / reactive power output of the micro gas turbine. This represents the active power / reactive power output of the energy storage unit. This represents the stored energy value. This represents the active power / reactive power output of distributed photovoltaic systems. The active power / reactive power flowing on the transmission line between nodes. The square of the node voltage; For the current time period, For nodes.

[0039] Based on the above definitions, the following constraints are established to represent the restrictive relationships defined above:

[0040] (1); (2);

[0041] (3);

[0042] It should be noted that equations (1) to (3) are the operating constraints of the micro gas turbine.

[0043] (4); (5);

[0044] (6); (7);

[0045] It should be noted that equations (4) to (7) are the operating constraints of the energy storage unit.

[0046] (8);

[0047] (9);

[0048] It should be noted that equations (8) and (9) are the output constraints of distributed photovoltaic systems.

[0049] (10); (11);

[0050] (12); (13);

[0051] (14); (15);

[0052] It should be noted that equations (10) to (15) are network constraints.

[0053] Specifically, These represent the upper and lower limits of the power output of the micro gas turbine. The upper and lower limits of power output for micro gas turbines; These are the upper limits for the power generation and energy storage capacity of the energy storage unit. These are the upper and lower limits for energy storage. The self-discharge rate of the energy storage unit; This refers to the power generation of distributed photovoltaic systems. This represents the upper limit of the curtailment rate for distributed photovoltaic power. Active / reactive load, These are the upper and lower limits of the active power of the transmission line. These are the upper and lower limits of the reactive power of the transmission line. These are the upper and lower limits of the square of the node voltage. For the resistance and inductance of the transmission line; These represent the phase angles of a micro gas turbine, an energy storage unit, and a distributed photovoltaic system, respectively. The superscript "max" indicates the upper limit, and the superscript "min" indicates the lower limit.

[0054] It should be noted that the above power grid model constraints can be reorganized into the following compact form:

[0055] (16);

[0056] (17);

[0057] (18).

[0058] Understandable, Let be a vector consisting of all decision variables. Let n be the total power exchanged between the n ports and the upstream power grid; Equation (16) corresponds to all inequality constraints, and Equation (17) corresponds to all equality constraints; parameter matrix sum coefficient vector All are determined by the various constraints; Ports Total power exchanged with the upper-level power grid, of which Equation (18) represents the constraints corresponding to the power grid aggregation operation range. for and port The relationship matrix of switching power. For explicit definition, the superscript T denotes transpose; otherwise... All represent the total number of time periods.

[0059] Specifically, and corresponding , For power grid ports The upper and lower bounds of the output power are related as follows:

[0060] (19);

[0061] in, for OK The identity matrix of columns.

[0062] It should be noted that the aggregation problem is a robust optimization problem aimed at finding the aggregated operating range of a multi-port power grid under all scheduling scenarios. The aggregated operating range refers to the set of all possible operating points of the power at the two ports of the power grid, provided that all internal operating constraints are satisfied.

[0063] Understandably, for the power grid, the dispatch instructions given by external power grids are uncertain. At the same time, these dispatch instructions are based on the aggregated operating range submitted by the multi-port power grid. Therefore, we hope to find an aggregated operating range that ensures the power grid can successfully execute any dispatch instructions issued by the external power grid based on this aggregated operating range.

[0064] Therefore, it is understandable that the robust optimization problem proposed in this application is a decision optimization problem. Its objective is not to maximize or minimize a certain performance index while considering uncertainties within the power grid, but rather to determine the feasibility of the aggregated operation range of the power grid. This differs fundamentally from traditional economic dispatch or optimal power flow problems in that its objective function is constructed as a constant, such as 0. The solver's task is to search the boundary of the feasible solution space that makes the problem solvable, rather than finding a single optimal solution.

[0065] Specifically, the aggregation problem can be solved using a min-max robust optimization problem with an objective function of 0. The min-max robust optimization problem is an optimization method that aims to find a feasible decision scheme under the worst uncertainty scenario. In this context, solving it with an objective function of 0 aims to focus on finding a feasible power boundary that satisfies all model constraints, rather than pursuing a specific economic objective.

[0066] It should be noted that the aggregation problem can be expressed in the following form:

[0067] (20).

[0068] It is understandable that RES refers to the aggregation problem. Based on the above model constraints and aggregation runtime constraints, i.e., taking equations (16) to (18) as constraints, equation (20) can be iteratively obtained. Before iterating to the aggregation runtime, the solution to the problem is... When the iteration reaches the aggregate running interval, the solution to the problem is 0. Therefore, we hope to find an aggregate running interval such that the solution to equation (20) is... .

[0069] Specifically, although the overall model expression is consistent, the topology, line parameters and electrical distances of the power grid are different due to the differences in the physical location (i.e., nodes) of multiple ports. As a result, the key parameters in the model constraints involved in solving the maximum and minimum operating capabilities of each port individually will actually change.

[0070] Therefore, during modeling and optimization, model constraints need to be explicitly divided into first-port, second-port, third-port, etc., to ensure that the aggregated operating range calculation for each port accurately reflects the specific electrical environment of its node. From an implementation perspective, this can be accomplished through a constraint management module configured in the energy management system: this module is server-based in hardware and integrates a power grid model database in software. It can automatically call the corresponding parameter set based on the node identifier connected to the port, and generate and load the model constraints for the first-port, second-port, third-port, etc., into the optimization solver.

[0071] Understandably, this method of dividing model constraints by port fundamentally ensures that the constraints relied upon when solving the aggregation problem step by step are consistent with the actual physical connection of each port. This makes the obtained aggregation operating ranges of the first port, second port, third port, etc. not only mathematically rigorous and feasible, but also able to accurately characterize the independent operating boundaries of each port in the real power grid, providing a correct constraint basis for the subsequent de-aggregation of scheduling instructions.

[0072] Specifically, the solution process for the first port is given here; the solution process for the aggregation run interval of the remaining ports can be obtained similarly. To make the original min-max problem easier to solve, the inner max problem in the aggregation problem is transformed into duality, resulting in the... A unified min dual problem in the next iteration is as follows:

[0073] (twenty one);

[0074] (twenty two);

[0075] (twenty three).

[0076] Specifically, As dual variables, Lagrange multipliers under KKT conditions Let it be a sufficiently large positive number in the "Big M method". It is a Boolean variable. For parameter matrices, Total number of time periods for OK The identity matrix of columns.

[0077] Specifically, solving equation (21) yields the current... Worst scheduling instruction and indicating effective constraints Boolean variables The value of (0 or 1) directly defines and indicates whether the corresponding inequality constraint is a valid constraint. The complementary relaxation condition in the KKT conditions (represented as a constraint in the problem) (And the sign restrictions on λ and relaxation) through the "Big M" constraint and The values ​​of form a mathematical linkage, ensuring that =0 (effective constraint) and λ>0 occur simultaneously, while =1 (invalid constraint) and λ=0 occur simultaneously. Therefore, within this modeling framework, binary variables... These are decision variables specifically designed to identify effective constraints; the two are equivalent.

[0078] Then, using the worst-case scheduling instruction Find the nearest feasible scheduling instruction as follows:

[0079] (twenty four);

[0080] (25).

[0081] Specifically, according to and feasible scheduling instructions iteration as follows:

[0082] (26);

[0083] (27).

[0084] Specifically, It is a Boolean variable. For and Vectors of the same dimension; let and Repeat all steps of solving the aggregation problem until... .

[0085] Ultimately, multiple ports can be obtained as independent aggregated operating ranges. However, in actual interconnection, the two will inevitably be mutually constrained, so the switching capacity of each port as a whole should be less than this limit.

[0086] Step S20: Integrate multiple independent aggregation operation intervals into an initial aggregation space for a multi-port power grid, and determine it as the current aggregation space.

[0087] It should be noted that integration refers to a synthesis operation, the input of which is the multiple independently defined aggregation operating intervals obtained in step S10. These operating intervals can be mathematically represented as convex polyhedra. The initial high-dimensional polyhedral model is a higher-dimensional geometric model constructed by synthesizing the polyhedral operating intervals of the two ports mentioned above. This model fully describes the set of solutions for the power grid's external interaction power over multiple time periods, without considering interconnection factors.

[0088] It should be noted that the aggregated operating range of all ports is integrated into a high-dimensional polyhedron represented by H-, which is the initial aggregated space:

[0089] (28);

[0090] (29).

[0091] It should be noted that, for OK The identity matrix of columns; Total number of ports; The total number of time periods; this high-dimensional polyhedron This is the initial value for the multi-port aggregation space, which corresponds to one in each time period. Polyhedron; the Each axis of the space containing the polyhedron corresponds to the power of a port. Each face of a polyhedron corresponds to an axis direction. Components. Therefore , It can be expressed by equation (29).

[0092] Understandably, in the initial state, the initial aggregation space is used as the current aggregation space as the initial condition for subsequent operations. This initial aggregation space does not consider any interconnection characteristics between ports; it only expresses the maximum and minimum power range of each port.

[0093] Step S30: Solve the cutting plane acquisition problem based on the current aggregation space and model constraints, and obtain the current cutting plane constraints and the current worst-case scheduling instruction.

[0094] It should be noted that the cutting plane acquisition problem refers to the problem of determining whether there are infeasible scheduling instructions in the current aggregation space during the iterative process of multi-port power grid aggregation through mathematical optimization. The goal is to find the cutting plane containing the worst-case scheduling instruction to cut out infeasible regions and gradually optimize the aggregation space. Ultimately, ensuring that the aggregation space is feasible under all scheduling scenarios is the key to iterative convergence.

[0095] It should be noted that the scheduling instruction refers to the arrangement of power values ​​for each time period; the worst-case scheduling instruction refers to the infeasible scheduling instruction with the largest sum of power correction values ​​for all time periods among all scheduling instructions that were originally infeasible but have been modified to be feasible in the current aggregation space.

[0096] Therefore, the cutting plane acquisition problem is a robust optimization problem with the goal of testing the feasibility of the current aggregation space and generating cutting plane constraints for cutting infeasible regions.

[0097] In one feasible implementation, the cutting plane acquisition problem is modeled as an objective function that includes model constraints and current aggregation space constraints. The max-min robust optimization problem.

[0098] Specifically, the model is a max-min problem as follows:

[0099] (30);

[0100] (31).

[0101] Specifically, This is the vector corresponding to the upward power correction value. This is the vector corresponding to the downward power correction value. It is a column vector consisting entirely of 1s. To enable scheduling instructions The corrected value becomes feasible. and For power grid constraints, correction values ​​are taken into account; Let m be the multi-port aggregation space in the current iteration, where m is the iteration order. In the initial state, m is 1, which is the initial aggregation space mentioned earlier. It requires that both power correction values ​​be positive.

[0102] Understandably, this is due to the multi-port aggregation space in the current iteration. Dispatch instructions issued There may be parts that are not feasible, so a power correction value is used. To make it feasible, when the solution result of equation (30) is 0, it means that no matter what scheduling instruction is made, no power correction value is needed to make it feasible, which means that it is iterated to a feasible multi-port aggregation space.

[0103] In other words, the convergence condition for the iterative process until the current aggregation space is feasible is: the result of the cutting plane problem is zero, which means that all scheduling instructions in the current aggregation space are executable.

[0104] To implement step S30, the inner-level min problem of the cutting plane acquisition problem must first undergo a dual transformation to obtain the dual problem of the cutting plane acquisition problem. The obtained unified max dual problem is as follows:

[0105] (32);

[0106] (33).

[0107] in, As dual variables, Since it is a bilinear term, it is equivalently treated using the KKT conditions as follows:

[0108] (34);

[0109] (35).

[0110] in, For KKT multipliers, It is a Boolean variable.

[0111] Then, solving the dual problem yields the current worst-case scheduling instruction in the current multi-port aggregation space. And the solution value of the dual variable. .

[0112] Finally, based on the objective that the result of the dual problem is less than or equal to 0, the current cutting plane constraint of the current aggregation space is generated by combining the current worst-case scheduling instruction and the solution value of the dual variable.

[0113] It should be noted that when the algorithm has found the perfect aggregation space, any scheduling instruction Y can satisfy all internal constraints of the power grid without any power correction. At this point, the optimal value of the inner min problem is 0, so the result of the entire max-min problem is... The algorithm succeeded and can be stopped.

[0114] Understandably, in the intermediate steps of the algorithm, the current aggregation space is not perfect, and there are some infeasible scheduling instructions. For these instructions, the power grid must pay an additional cost, namely a power correction value, to operate normally. The outer max algorithm will find the infeasible scheduling instruction that requires the maximum correction cost, i.e., the worst-case scheduling instruction. At this time This represents the sum of the maximum correction costs. It is an alert that clearly indicates that there are infeasible regions in the current aggregation space.

[0115] It is understandable that, for equation (32), considering that the original max-min problem, when solving to a feasible multi-port aggregation space, During the iteration process, a power correction value is required. Therefore there is Considering strong duality, in the dual problem... Equivalent to Therefore, the cutting plane can be used to remove this infeasible space, corresponding to The situation.

[0116] At this time, the The cutting plane within the multi-port aggregation space obtained in the next iteration The following results were obtained:

[0117] (36).

[0118] Understandable, The expression is defined or assigned to indicate that the new symbol, variable, or expression on the left is defined by the expression on the right; Equation (36) is to construct an inequality, that is, a cutting plane constraint, by solving the dual variable, to express the cutting plane. The infeasible areas that were removed.

[0119] Step S40: Based on the current worst-case scheduling instruction and the current cutting plane constraints, obtain the currently feasible scheduling instruction that is closest to the current worst-case scheduling instruction in the current aggregation space.

[0120] It is understood that step S40 includes: determining the non-cutting plane constraints of the current feasible scheduling instruction based on the current cutting plane constraints; determining the grid constraints of the current feasible scheduling instruction based on the model constraints; and obtaining the current feasible scheduling instruction that is closest to the current worst-case scheduling instruction in the current aggregation space based on the non-cutting plane constraints and the grid constraints.

[0121] It is understandable that this non-cutting plane constraint means that the solution value does not fall within the part to be cut by the cutting plane.

[0122] Specifically, the solution process is as follows:

[0123] (37);

[0124] (38).

[0125] Equation (37) is used to solve the current feasible scheduling instructions using the L2 distance. Objective function, constraints , To explain It is a dispatch instruction that is feasible given the constraints of the power grid. In order to make lie in Within the cut area.

[0126] Understandably, the goal of finding the most recent feasible scheduling instruction is to achieve accurate positioning along subsequent time-segmentation.

[0127] Step S50: Under the constraints of the most recent feasible scheduling instruction, project the current cutting plane constraint onto the current aggregation space under different time periods to cut and update the current aggregation space. Then, execute the step of solving the cutting plane acquisition problem based on the current aggregation space and model constraints until the current aggregation space is feasible, thereby realizing multi-port power grid aggregation.

[0128] As can be understood, projection specifically refers to decomposing a global cut plane constraint describing the entire scheduling cycle, such as a 24-hour period, into a set of local constraints describing each time period, such as the t-th hour. Its technical essence is dimensionality reduction and decoupling. Through projection, complex constraints coupling the port power of all time periods are transformed into a simple set of independent constraints within each time period, involving only the port power of that specific time period. This significantly reduces the complexity of subsequent calculations.

[0129] In addition, the multi-port aggregation space is essentially the aggregation of the coupling characteristics between various ports. These coupling characteristics are determined to some extent by the network within the power grid, and the network constraints within the power grid are time-varying. Cutting plane projection can better handle time-varying situations.

[0130] Understandingly, "cutting" refers to adding a linear inequality constraint—that is, the set of cut planes for each time interval obtained after projection—as a new constraint to the constraint set of the current aggregation space. "Update," on the other hand, refers to using this new, more stringent set of constraints to define the aggregation space used in the next iteration. This new space is a subset of the original space; those regions deemed infeasible are cut off.

[0131] In one feasible implementation, the current cutting plane constraint is first decomposed and projected according to different time periods to obtain independent projected cutting plane constraints for each time period. The projected cutting plane constraint for each time period ensures that its geometry is consistent with the shape of the original cutting plane in that time period and must pass through the component of the current feasible scheduling instruction in that time period.

[0132] Specifically, based on the solution values ​​of the dual variables and the current feasible scheduling instructions, the parameters of the current cutting plane constraint can be decomposed into time periods, and parameter slices corresponding to each time period can be extracted. The calculation of the parameter slices incorporates the time period components of the current feasible scheduling instructions, and offset adjustments are made using Kronecker product operations, thereby constructing an independent projected cutting plane constraint for each time period, forcing it to pass through feasible points and maintaining consistency with the geometric projection of the original cutting plane in that time period.

[0133] Therefore, the cutting plane Projected to different time periods In the space containing the polyhedron, the projected cutting plane for each time period is obtained. as follows:

[0134] (39).

[0135] in, Represents the first vector within the parentheses The element to the first One element; For Kronecker product.

[0136] It is understandable that the matrix obtained by equation (39) This is to ensure that the projected cutting plane and the cutting plane before projection are in the same order at each time period. The shape is consistent during this period; vector This is to ensure that the cutting plane at each time interval after projection can be processed by feasible scheduling instructions. The portion within that time period.

[0137] Then, the projection cut plane constraints for all time periods are integrated into a global projection cut plane constraint covering the entire time period. (This refers to the projection cut plane constraints for all time periods.) Integrate into the following form:

[0138] (40).

[0139] It is understandable that, through equation (40), the integrated form of all projective cutting planes can be obtained. .

[0140] Finally, the global projection cutting plane constraint is combined with the current aggregate space constraint to generate a new aggregate space constraint with a cut-down range, thereby completing the update of the current aggregate space.

[0141] It is understandable that the multi-port aggregation space is formed after iteration through the projection cutting plane. as follows:

[0142] (41).

[0143] It should be noted that, following the order Then, the steps of solving the cutting plane problem based on the current aggregation space and model constraints are executed until the current aggregation space is feasible, that is... This enables multi-port power grid aggregation.

[0144] In this embodiment, by modeling the multi-port power grid based on distributed energy and obtaining model constraints, and then solving a robust optimization problem aimed at finding the aggregated operating interval under all scheduling scenarios, the independent aggregated operating interval of each port is obtained, thereby initially and accurately characterizing the operating characteristics of a single port and laying the foundation for aggregation; then these independent intervals are integrated into an initial aggregation space as the current aggregation space, so as to initially reflect the prototype of multi-port interconnection.

[0145] Then, by solving the cutting plane acquisition problem with the goal of testing the feasibility of the current aggregation space and generating cutting plane constraints, the current cutting plane constraints and worst-case scheduling instructions are obtained, thereby identifying infeasible regions in the aggregation space and preparing for cutting; then, based on the worst-case scheduling instructions and cutting plane constraints, the feasible scheduling instructions that are closest to the worst-case scheduling instructions in the current aggregation space are obtained, providing accurate positioning for subsequent projection.

[0146] Finally, under the constraint of the most recently feasible scheduling command, the cut plane constraint is projected onto the current aggregation space under different time periods, the current aggregation space is cut and updated, and the cut plane acquisition problem is iteratively executed until the aggregation space is feasible. This series of techniques, by iteratively removing infeasible regions and retaining feasible points, gradually and accurately characterizes the interconnection characteristics between multiple ports, solving the aggregation difficulty problem caused by the lack of accurate characterization methods in existing technologies. Compared with existing technologies, it achieves efficient and reliable aggregation of multi-port power grids, improving the feasibility of scheduling commands and the stability of power grid operation.

[0147] The following description, using a specific application scenario, further illustrates the beneficial effects achievable in this embodiment. For a 30-node three-port grid system, comprising two micro gas turbines, two energy storage units, and eight distributed photovoltaic systems, ports A, B, and C are located at nodes 1, 21, and 30, respectively. Its topology is shown below. Figure 3 As shown, the daily output of distributed photovoltaic power is as follows: Figure 4 As shown, the daily load is as follows Figure 5 As shown, the data for the micro gas turbine and energy storage unit are presented in Tables 1 and 2, respectively:

[0148] Table 1. Parameters of the micro gas turbine:

[0149]

[0150] Table 2 Energy Storage Unit Parameters:

[0151]

[0152] A program was written in MATLAB R2024a to call the Yalmip solver with Gurobi. The optimization problem was solved using a Legion laptop with an Intel Core i7-10510U processor, 32GB of RAM, and running Windows 11 Professional. The aggregated execution ranges for ports A, B, and C are as follows: Figure 6 As shown.

[0153] It should be noted that integrating the aggregation and operation ranges of ports A, B, and C yields a high-dimensional polyhedron. This is the initial value of the multi-port aggregation space. For example... Figure 7 As shown, the cube in each subgraph is... In the initial multi-port aggregation space within each time period, the unit of each coordinate axis is MW.

[0154] It should be noted that the schematic diagram of the projected cutting plane is as follows: Figure 8 As shown, the unit for each coordinate axis is MW. This projected cutting plane is defined in the first iteration, i.e. The cutting plane at time The initial multi-port aggregation space is cut, and the initial multi-port aggregation space is the grass-green cuboid above before it was cut. The specific form of the projection cutting plane at that time is:

[0155] (42).

[0156] in for The power at ports A, B, and C is shown. The red plane is the projection cutting plane, the blue polyhedron represents the infeasible portion cut out, and the light green polyhedron below represents the portion cut by the projection cutting plane in the first iteration. exist The space corresponding to time. The projection plane at that time can be deduced by analogy.

[0157] It should be noted that the aggregated operating range for each hour of the 24-hour period within a day considers the interconnection characteristics of ports A, B, and C. Figure 9 As shown, the three-dimensional polyhedron in each subgraph represents the aggregated operation range, and the unit of each coordinate axis is MW. 10,000 dispatch instructions are generated within the aggregated operation range of the power grid, and all dispatch instructions are input into the power grid to test whether the power grid can operate normally.

[0158] Table 3: Dispatch results for power grid operating sections:

[0159]

[0160] The results in Table 3 show that all 10,000 dispatch instructions can be successfully executed by the power grid, proving the feasibility of the proposed method. The aggregated operating intervals obtained using the multi-port power grid aggregation method are used in power grid dispatching, and the results are compared with dispatching without the power grid aggregation method.

[0161] Table 4. Dispatch results for power grid operating sections:

[0162]

[0163] As shown in Table 4, the proposed method can significantly reduce the number of variables and constraints in the power grid that need to be considered during scheduling, thereby accelerating the scheduling operation of the power grid and reducing the solution time from more than 1 minute to less than 10 seconds, proving the effectiveness of the proposed method.

[0164] Based on the methods in the above embodiments, such as Figure 10 As shown in the illustration, this application provides an electronic device that may include a processor, a communications interface, a memory, and a communication bus. The processor, communications interface, and memory communicate with each other via the communication bus. The processor can invoke logical instructions stored in the memory to execute the methods described in the above embodiments.

[0165] Furthermore, the logical instructions in the aforementioned memory can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application.

[0166] Based on the methods in the above embodiments, this application provides a computer-readable storage medium storing a computer program that, when run on a processor, causes the processor to execute the methods in the above embodiments.

[0167] Based on the methods in the above embodiments, this application provides a computer program product that, when run on a processor, causes the processor to execute the methods in the above embodiments.

[0168] It is understood that the processor in the embodiments of this application can be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. A general-purpose processor can be a microprocessor or any conventional processor.

[0169] The method steps in this application embodiment can be implemented in hardware or by a processor executing software instructions. The software instructions can consist of corresponding software modules, which can be stored in random access memory (RAM), flash memory, read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), registers, hard disks, portable hard disks, CD-ROMs, or any other form of storage medium known in the art. An exemplary storage medium is coupled to the processor, enabling the processor to read information from and write information to the storage medium. Of course, the storage medium can also be a component of the processor. The processor and the storage medium can reside in an ASIC.

[0170] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted through the computer-readable storage medium. The computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state disk (SSD)).

[0171] It is understood that the various numerical designations used in the embodiments of this application are merely for the convenience of description and are not intended to limit the scope of the embodiments of this application.

[0172] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A multi-port power grid aggregation method based on secant plane projection, characterized in that, include: Step S10: Model the multi-port power grid based on the distributed energy resources set in the power grid and obtain the model constraints. Then, solve the aggregation problem based on the model constraints to obtain multiple independent aggregation operation intervals for multiple ports. The aggregation problem is a robust optimization problem aimed at finding the aggregation operation intervals of the multi-port power grid under all scheduling scenarios. Step S20: Integrate multiple independent aggregation operation intervals into an initial aggregation space for a multi-port power grid, and determine it as the current aggregation space; Step S30: Solve the cut plane acquisition problem based on the current aggregation space and the model constraints to obtain the current cut plane constraints and the current worst-case scheduling instruction; the cut plane acquisition problem is a robust optimization problem with the goal of testing the feasibility of the current aggregation space and generating cut plane constraints for cutting infeasible regions; the current worst-case scheduling instruction refers to the infeasible scheduling instruction with the largest sum of power correction values ​​in all time periods among all scheduling instructions that were originally infeasible but have become feasible after modification in the current aggregation space. Step S40: Based on the current worst-case scheduling instruction and the current cut plane constraint, obtain the current feasible scheduling instruction that is closest to the current worst-case scheduling instruction in the current aggregation space; Step S50: Under the constraint of the most recent currently feasible scheduling instruction, project the current cutting plane constraint onto the current aggregation space under different time periods to cut and update the current aggregation space, and then execute the step of solving the cutting plane acquisition problem based on the current aggregation space and the model constraints until the current aggregation space is feasible, thereby realizing multi-port power grid aggregation.

2. The multi-port power grid aggregation method based on secant plane projection as described in claim 1, characterized in that, The problem of obtaining the cutting plane is modeled as an objective function that includes the model constraints and the current aggregation space constraints. The max-min robust optimization problem, in which, It is a column vector consisting entirely of integers 1; This is the vector corresponding to the upward power correction value. This is the vector corresponding to the downward power correction value.

3. The multi-port power grid aggregation method based on secant plane projection as described in claim 2, characterized in that, Step S30 includes: By performing a dual transformation on the inner min problem of the cutting plane acquisition problem, the dual problem of the cutting plane acquisition problem can be obtained. Solving the dual problem yields the current worst-case scheduling instruction and the solution value of the dual variable in the current aggregation space; Based on the objective that the result of the dual problem is less than or equal to 0, the current cutting plane constraint in the current aggregation space is generated by combining the current worst-case scheduling instruction and the solution value of the dual variable.

4. The multi-port power grid aggregation method based on secant plane projection as described in claim 1, characterized in that, Step S40 includes: Based on the current cutting plane constraints, determine the non-cutting plane constraints of the current feasible scheduling instructions; The grid constraints for the current feasible scheduling instructions are determined based on the model constraints. Based on the non-cutting plane constraint and the power grid constraint, obtain the currently feasible scheduling instruction that is closest to the current worst-case scheduling instruction in the current aggregation space.

5. The multi-port power grid aggregation method based on secant plane projection as described in claim 1, characterized in that, Step S50 includes: The current cutting plane constraint is decomposed and projected according to different time periods to obtain independent projected cutting plane constraints for each time period; wherein, the projected cutting plane constraint for each time period ensures that its geometry is consistent with the shape of the original cutting plane in that time period, and passes through the component of the most recent currently feasible scheduling instruction in that time period. The projection cut plane constraints for all time periods are integrated into a global projection cut plane constraint that covers the entire time period. The overall projection cutting plane constraint is combined with the current aggregate space constraint to generate a new aggregate space constraint with a reduced range, thereby completing the update of the current aggregate space.

6. The multi-port power grid aggregation method based on secant plane projection as described in claim 5, characterized in that, The projected cutting plane constraint for each time period ensures that its geometry is consistent with the shape of the original cutting plane in that time period, and includes the components of the most recent currently feasible scheduling instruction within that time period, including: Based on the solution values ​​of the dual variables and the current feasible scheduling instructions, the parameters of the current cutting plane constraint are decomposed into time periods, and parameter slices corresponding to each time period are extracted. The calculation of the parameter slices incorporates the time period component of the most recent current feasible scheduling instructions, and then offset adjustments are made to construct an independent projective cutting plane constraint for each time period.

7. The multi-port power grid aggregation method based on secant plane projection as described in claim 1, characterized in that, The iterative process continues until the convergence condition for the current aggregation space is met: the result of the problem obtained by the cutting plane is zero, indicating that all scheduling instructions in the current aggregation space are executable.

8. An electronic device, characterized in that, Includes memory and one or more processors; The memory is coupled to the one or more processors, and the memory is used to store computer program code, the computer program code including computer instructions; The one or more processors invoke the computer instructions to cause the electronic device to perform the method as described in any one of claims 1 to 7.

9. A computer-readable storage medium comprising instructions, characterized in that: When the instructions are executed on an electronic device, the electronic device performs the method as described in any one of claims 1 to 7.

10. A computer program product, comprising a computer program or instructions, characterized in that: When the computer program or instructions are run on an electronic device, the electronic device causes the electronic device to perform the method as described in any one of claims 1 to 7.

Images