Charging station capacity day-ahead market allocation method and system under safety constraints of power distribution network

Abstract

A power distribution network safety constraint under charging station capacity day-ahead market allocation method and system. The method comprises the following steps: acquiring traffic network, power distribution network and charging station parameters, constructing a charging station capacity clearing model based on power distribution network constraints, including a day-ahead power distribution network cost minimization model and a day-ahead rescheduling model; constructing an ellipsoid uncertainty set according to the original simulated charging load data set, converting the ellipsoid uncertainty set into a polyhedral uncertainty set based on the vertex linking algorithm SGVL; extracting the main problem and the sub-problem from the charging station capacity clearing model, constructing a feasible cut plane constraint, solving and optimizing the main problem and the sub-problem based on the improved Benders dual decomposition formula, and obtaining the optimal solution of capacity allocation, generator output and rescheduling output. The scheme of the application reduces the system operation cost and rescheduling demand, effectively reduces the conservativeness of the model, and improves the overall robustness and economy of the system.

Description

Technical Field

[0001] This invention belongs to the field of electric vehicle charging station management and power system optimization scheduling, and specifically relates to a day-ahead market-based allocation method and system for charging station capacity under distribution network security constraints. Background Technology

[0002] In recent years, the electric vehicle industry has developed rapidly, making the construction, management, and operation of charging stations crucial. Traditional charging station capacity is predefined, neglecting the reality that charging stations do not always operate at their rated capacity due to the limited resources of the power distribution network. Capacity cannot be dynamically optimized based on the network's operational status. Therefore, there is an urgent need for a solution that, during market-based capacity allocation, ensures the safe operation of the power distribution network while guiding charging stations to compete based on their true capacity value, thereby achieving optimal resource allocation efficiency. Summary of the Invention

[0003] To address the shortcomings of existing technologies, this invention provides a day-ahead market-based allocation method and system for charging station capacity under distribution network security constraints, thereby resolving the technical problems of insufficient dynamic capacity adjustment and overly conservative uncertainty characterization methods in existing EVCS capacity market clearing studies.

[0004] To solve the above-mentioned technical problems, the present invention adopts the following technical solution.

[0005] This invention first discloses a method for day-ahead market-based allocation of charging station capacity under distribution network security constraints, the method comprising the following steps: Step 1: Obtain various parameters of the transportation network, distribution network, and charging stations. Introduce the charging capacity market in the day-ahead phase. Construct a charging station capacity clearing model based on distribution network constraints, including a day-ahead distribution network cost minimization model in the first phase and a day-ahead rescheduling model in the second phase. Perform robust optimization on the two-stage models. Step 2: Construct an ellipsoidal uncertainty set based on the original simulated charging load dataset, and transform the ellipsoidal uncertainty set into a polyhedral uncertainty set based on the vertex linking algorithm SGVL, so as to realize data-driven load modeling in an uncertain environment for describing the prediction error of electric vehicle charging load; Step 3: Based on the load modeling results, extract the main problem and sub-problems from the robustly optimized charging station capacity clearing model, construct the feasibility cutting plane constraint, and solve and optimize the main problem and sub-problems based on the improved Benders dual decomposition method to obtain the optimal solutions for capacity allocation, generator output, and rescheduled output.

[0006] The present invention further includes the following preferred embodiments: The objective function of the day-ahead distribution network cost minimization model is: (1) In the formula: This is the power generation cost coefficient; Represents a node The day-ahead forecast of generator output; the power generation cost of the distribution network only considers the cost of purchasing electricity from the upstream power grid. For a period of time; This is the EVCS capacity reduction ratio; For charging stations The day-ahead forecast of electricity purchase capacity; , These are the revenue coefficient per unit capacity and the EVCS electricity purchase cost coefficient, respectively. Represents a node The maximum capacity of the EVCS; G and M are the generator and EVCS sets, respectively.

[0007] The day-ahead distribution network cost minimization model satisfies the capacity reduction factor and power purchase constraints, as well as the distribution network power flow constraints: The capacity reduction factor and the power purchase constraint are as follows: (2) (3) The power flow constraints of the distribution network are as follows: (4) (5) (6) (7) (8) (9) In the formula: It is the set of all branches in the distribution network; It is the set of all nodes in the power distribution network; and They are nodes The active and reactive power output of the generator; and They are nodes The active and reactive power of the load; and Branch roads Active power and reactive power; and Branch roads Active power and reactive power; For nodes The square of the voltage amplitude; For nodes The square of the voltage amplitude; branch road The resistance; branch road The reactance; branch road Active power; branch road reactive power; and They are nodes The lower and upper limits of the square of the voltage amplitude; and They are nodes The lower and upper limits of generator active power output; and They are nodes The lower and upper limits of generator reactive power output.

[0008] The objective function of the day-ahead rescheduling model is: (13) (14) (15) In the formula: , , , , , It is a coefficient matrix; This is the maximum capacity vector for EVCS.

[0009] The robust optimization of the two-stage model further includes: calculate (twenty three) (14) (15) (20)-(21) In the formula: For the regions where rescheduling is feasible: This is the region where rescheduling is feasible.

[0010] Step 2 further includes: Step 2.1: Generate the dataset of the original simulated charging load. Randomly divided into shape learning datasets With size calibration dataset The sample sizes of each part satisfy the following relationship: (twenty four) In the formula: For dataset The total number of samples, and Each is a dataset and The sample size It is a positive integer. For the allowed probability of violation, For statistical confidence complement parameters, This is the floor operator; Step 2.2, using Construct the initial set of uncertainties for the ellipsoid; the mean vector of the dataset is denoted as... The covariance matrix is ​​denoted as The ellipsoid set can be represented as: (25) In the formula: This represents the actual charging load vector. This is the mean vector of the simulated load data; The covariance matrix of the simulated load; The simulated load vector is obtained after rotation transformation: (27) In the formula: It is a rotation matrix composed of the eigenvectors of the covariance matrix. It is a diagonal matrix composed of eigenvalues. This is the transformed load simulation vector; Step 2.3: Use the vertex linking method to convert the above ellipsoid set into a polyhedron set; the ellipsoid has Given vertices, construct a matrix of all positive and negative semi-axis vertices of the ellipsoid set. , : (28) By combining the positive and negative forms of each column in the two matrices above, we can generate... The combination matrix of each vertex W Further construct the hyperplane parameter vector of the polyhedron set. : (29) In the formula: For the first A combination matrix of vertices, This is the corresponding hyperplane normal vector; Step 2.4: Utilize the validation dataset Calibrate the radius parameters of the polyhedron set for each Define its projection value onto the polyhedron set: (30) In the formula: For data points The mapping value; Sort all projection values ​​in ascending order to get: (31) Calculate the critical parameters used to determine the radius of the polyhedral assembly. ,satisfy: (32) In the formula: Number the loads in ascending order; To meet the confidence level With the probability of violation Critical parameters; Represents the number of combinations, from Select from data The number of combinations of data points; Define the radius parameter of the polyhedron set as: (33) The above parameters are used to construct a set of polyhedral uncertainties with statistical confidence guarantees. : (34) In the formula: Indicates the sorted order of the first... One data point; The radius parameter of the polyhedron set.

[0011] The improved Benders dual decomposition method optimizes the solution of the main problem and subproblems, further including: Step 3.3.1: Input parameter data and convergence tolerance. Initialize the number of iterations Number of cutting plane constraints The cutting plane constraint set is initially empty; Step 3.3.2, Number of Update Iterations for Solve the current main problem: (49) Obtain the solution of the decision variables in the current step. ; Step 3.3.3: Apply the solution obtained from the main problem... Input a subproblem to solve to verify the robustness of the current solution: (58) If the maximum value of the subproblem Meet the conditions If the current solution does not meet the robust feasibility constraint, a new cutting plane constraint is generated and fed back to the main problem.

[0012] Update cutting plane constraints for If the iteration fails, return to step 3.3.2 and continue iterating; otherwise, proceed to step 3.3.4. Step 3.3.4: If the convergence condition is met If the feasibility of the subproblem is verified, the solution to the current main problem is the robust and feasible optimal solution, the iteration terminates, and the current optimal solution is output.

[0013] This invention also discloses a charging station capacity day-ahead market-based allocation system under distribution network security constraints, utilizing the aforementioned method for market-based allocation of charging station capacity under distribution network security constraints, comprising: The charging station capacity clearing model construction module is used to obtain various parameters of the transportation network, distribution network and charging stations. In the day-ahead stage, the charging capacity market is introduced, and the charging station capacity clearing model is constructed based on the distribution network constraints. It includes the day-ahead distribution network cost minimization model in the first stage and the day-ahead rescheduling model in the second stage. Robust optimization is performed on the two-stage model. The load modeling module is used to construct an ellipsoidal uncertainty set based on the original simulated charging load dataset, and transform the ellipsoidal uncertainty set into a polyhedral uncertainty set based on the vertex linking algorithm SGVL, so as to realize data-driven load modeling in the uncertainty environment for describing the prediction error of electric vehicle charging load. The Benders dual decomposition solution module is used to extract the main problem and subproblems from the robustly optimized charging station capacity clearing model based on the load modeling results, construct feasibility cutting plane constraints, and solve and optimize the main problem and subproblems based on the improved Benders dual decomposition method to obtain the optimal solutions for capacity allocation, generator output, and rescheduled output.

[0014] Accordingly, this application also discloses a terminal, including a processor and a storage medium; The storage medium is used to store instructions; The processor is configured to operate according to the instructions to execute the steps of the day-ahead market-based allocation method for charging station capacity under the aforementioned distribution network security constraints.

[0015] Accordingly, this application also discloses a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the aforementioned method for the day-ahead market-based allocation of charging station capacity under the security constraints of the distribution network.

[0016] The beneficial effects of this invention are as follows: Compared with the prior art, this invention provides a day-ahead market-based allocation method and system for charging station capacity under distribution network security constraints. It introduces a market-based capacity allocation mechanism based on bid prices, ensuring grid operation safety while guiding charging stations to participate in competition based on their true capacity value, thereby achieving optimal resource allocation efficiency. To address the uncertainty of EVCS load, a two-stage robust optimization capacity clearing model considering distribution network capacity constraints is constructed. A polyhedral uncertainty set based on the Statistical-Guarantee-based Vertex Link (SGVL) algorithm is built to describe the prediction error of electric vehicle charging load, minimizing the impact of load uncertainty. An improved Benders dual decomposition algorithm is used, with improved feasibility cut constraints to adapt to the coupling relationship between uncertain variables and decision variables. This avoids the cut failure and global non-convergence problems caused by DDU shape changes in traditional methods. While reducing system operating costs and rescheduling requirements, it effectively reduces the model's conservatism and improves the overall robustness and economy of the system. Attached Figure Description

[0017] Figure 1 This is a topology diagram of the coupled system of a 12-node transportation network and a 33-node distribution network in this invention.

[0018] Figure 2 This is a parameter diagram of the power distribution network and charging station in this invention.

[0019] Figure 3 This is a schematic table of simulation scenarios in this invention.

[0020] Figure 4 This is a flowchart of the day-ahead market-based allocation method for charging station capacity under the safety constraints of the power distribution network in this invention.

[0021] Figure 5 This is a schematic diagram of the two-stage robust optimization results under different scenarios in this invention. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0023] The embodiments described in this application are merely some, not all, embodiments of the present invention. Based on the spirit of the present invention, other embodiments obtained by those skilled in the art without inventive effort are all within the protection scope of the present invention.

[0024] To address the shortcomings of existing technologies, this invention proposes a day-ahead market-based allocation method and system for charging station capacity under distribution network security constraints, considering a two-stage EVCS capacity market clearing process that takes into account both distribution network constraints and load DDU. Under distribution network capacity constraints, a day-ahead EVCS capacity clearing strategy is adopted, flexibly adjusting EVCS capacity based on actual network resource limitations to reasonably address network congestion and reduction fairness issues. Addressing the coupling relationship between the uncertainty of EVCS charging load and capacity decision-making, a data-driven polyhedral uncertainty set model is established, constructing a two-stage robust optimal capacity market clearing model for the distribution network that considers charging load DDU. The model is solved using an improved Benders dual decomposition algorithm, considering the correlation between different charging stations, effectively reducing the conservatism of optimization decisions and providing a more reliable capacity market clearing and scheduling strategy.

[0025] See Figure 4 As shown, the present invention discloses a method for day-ahead market allocation of charging station capacity under distribution network security constraints, comprising the following steps: Step 1: Obtain various parameters of the transportation network, distribution network, and charging stations. Introduce the charging capacity market in the day-ahead phase. Construct a charging station capacity clearing model based on distribution network constraints, including a day-ahead distribution network cost minimization model in the first phase and a day-ahead rescheduling model in the second phase. Perform robust optimization on the two-stage models.

[0026] In traditional day-ahead electricity markets, distributed generation sources and conventional generating units submit generation curves, and demanders (including EV charging stations) submit bids based on predicted load. System operators clear the market and publish day-ahead dispatch plans, taking into account power flow, capacity, and voltage constraints. This mechanism typically assumes that charging stations can obtain their rated capacity without obstacles. However, with the increasing scale of EV integration, this assumption does not hold true at the distribution network level: when network transmission capacity is limited or reserves are insufficient, the actual usable capacity of EVCSs may be lower than their rated value. If the actual load of EVCSs exceeds available capacity at this time, network operating constraints will be violated, thus rendering the traditional day-ahead market clearing problem ineffective.

[0027] To ensure the safe and economical operation of the distribution network, a charging capacity market is introduced during the day-ahead phase. System operators no longer simply assume full availability of the EVCS (Electronic Power Supply System) capacity; instead, they use "allowable operating capacity" as an optimization variable, with market clearing and unified allocation. When there is a discrepancy between the predicted load and available capacity, the charging capacity of the EVCS will be reduced: that is, under network constraints, some stations cannot obtain their rated capacity and can only be scheduled for charging according to the system-allocated allowable operating capacity. This capacity reduction mechanism ensures that the system remains feasible and can maintain voltage and power flow constraints when network bottlenecks occur. Incentivized by the market mechanism, charging stations with higher capacity bids will receive priority in capacity allocation during periods of capacity shortage, while charging stations with lower bids are more likely to withstand load reductions. This bid-based capacity reduction mechanism can guide charging stations to compete based on their true capacity value while ensuring grid operational safety, thereby achieving optimal resource allocation efficiency.

[0028] The day-ahead operation objective is to minimize the total operating cost of the distribution network, which consists of the operating cost and the difference between the capacity revenue obtained from the EVCS and the revenue from selling electricity to the EVCS. Therefore, the objective function of the day-ahead distribution network cost minimization model in the first stage is as follows: (1) In the formula: This is the power generation cost coefficient; Represents a node The day-ahead forecast of generator output; the power generation cost of the distribution network only considers the cost of purchasing electricity from the upstream power grid. For a period of time; This is the EVCS capacity reduction ratio; For charging stations The day-ahead forecast of electricity purchase capacity; , These are the revenue coefficient per unit capacity and the EVCS electricity purchase cost coefficient, respectively. Represents a node The maximum capacity of the EVCS; G and M are the generator and EVCS sets, respectively.

[0029] The capacity reduction factor and power purchase constraints for the current distribution network cost minimization model are as follows: (2) (3) The generator sets adopt a constant power control method. The power flow constraints of the day-ahead distribution network cost minimization model are as follows: (4) (5) (6) (7) (8) (9) In the formula: It is the set of all branches in the distribution network; It is the set of all nodes in the power distribution network; and They are nodes The active and reactive power output of the generator; and They are nodes The active and reactive power of the load; and Branch roads Active power and reactive power; and Branch roads Active power and reactive power; For nodes The square of the voltage amplitude; For nodes The square of the voltage amplitude; branch road The resistance; branch road The reactance; branch road Active power; branch road The reactive power on. and They are nodes The lower and upper limits of the square of the voltage amplitude. and They are nodes The lower and upper limits of the generator's active power output. and They are nodes The lower and upper limits of generator reactive power output.

[0030] Since the power flow constraints are linear power flow constraints that ignore line losses, and the distribution network is usually a radial network, constraints (4)-(9) can be expressed as a function of node power through the distribution network node-branch correlation matrix: (10) (11) (12) In the formula: and The matrices represent the active and reactive power flow of the lines, respectively; and These are matrices representing the active and reactive power of the generator, respectively. for Vector representation of; and These are matrices representing the active and reactive loads of the nodes, respectively. Represents the node-branch correlation matrix of a power distribution network; The matrix representing the inverse of the node-branch association matrix; U A matrix representing the voltage at distribution network nodes; The transpose of the inverse of the node-branch association matrix; and These are matrices representing the line resistance and reactance, respectively. It is expressed as the square of the node voltage. This represents the node reference voltage vector.

[0031] The second-stage day-ahead rescheduling model is specifically represented by the following abstract model: (13) (14) (15) In the formula: , , , , , It is a coefficient matrix; This is the maximum capacity vector for EVCS.

[0032] The constraints for the second stage of the problem include distribution network rescheduling constraints and power flow constraints, as shown below: (16) (17) (18) (19) In the formula: For nodes The rate at which generator backup power is reduced; For nodes The rate at which generator backup is increased; To redistribute power generation output.

[0033] The second-stage model can be represented as: (20) (twenty one) (twenty two) In the formula: , , , , , It is a coefficient matrix; Decision variables The vector representation of .

[0034] The two-stage robust optimization model is shown below: (twenty three) (14) (15) (20)-(21) In the formula: This is the region where rescheduling is feasible.

[0035] Figure 5 This is a schematic diagram of the two-stage robust optimization results under different scenarios in this invention.

[0036] Step 2: Construct an ellipsoidal uncertainty set based on the original simulated charging load dataset, and transform the ellipsoidal uncertainty set into a polyhedral uncertainty set based on the vertex linking algorithm SGVL, thereby realizing data-driven load modeling under uncertain environment to describe the prediction error of electric vehicle charging load.

[0037] The load modeling process includes: Step 2.1, Data Segmentation: The original simulated charging load dataset Randomly divided into shape learning datasets With size calibration dataset This ensures that the constructed uncertainty set possesses statistical confidence properties. The sample sizes of each part satisfy the following relationship: (twenty four) In the formula: For dataset The total number of samples, and Each is a dataset and The sample size It is a positive integer. The permissible probability of violation (risk level). For statistical confidence complement parameters, This is the floor operator.

[0038] Step 2.2, Shape Learning: use Construct an initial set of ellipsoidal uncertainties. The mean vector of this dataset is denoted as... The covariance matrix is ​​denoted as The ellipsoid set can be represented as: (25) In the formula: This represents the actual charging load vector. This is the mean vector of the simulated load data; This is the covariance matrix for simulating the load.

[0039] Furthermore, regarding Eigenvalue decomposition yields: (26) Therefore, the form of the simulated load vector after rotation transformation can be obtained: (27) In the formula: It is a rotation matrix composed of the eigenvectors of the covariance matrix. Let N be a diagonal matrix composed of N eigenvalues. This is the transformed load simulation vector.

[0040] Step 2.3, Shape Calibration: The above ellipsoid set is converted into a polyhedron set using the vertex linking method. The ellipsoid has... Given vertices, construct a matrix of all positive and negative semi-axis vertices of the ellipsoid set. , As shown below: (28) In the formula: , These are the vertex matrices of the positive and negative semi-axis of the ellipsoid, respectively.

[0041] By combining the positive and negative forms of each column in the two matrices above, we can generate... The combination matrix of each vertex W Further construct the hyperplane parameter vector of the polyhedron set. : (29) In the formula: For the first A combination matrix of vertices, is the corresponding hyperplane normal vector.

[0042] Step 2.4, Dimensional Calibration: Using the validation dataset Calibrate the radius parameters of the polyhedron set. For each Its projection onto the polyhedron set is defined as: (30) In the formula: For data points The mapping value.

[0043] Sort all projection values ​​in ascending order to get: (31) Calculate the critical parameters used to determine the radius of the polyhedral assembly. ,satisfy: (32) In the formula: Number the loads in ascending order; To meet the confidence level With the probability of violation Critical parameters; Represents the number of combinations, from Select from data The number of combinations of data points.

[0044] Finally, the radius parameter of the polyhedron set is defined as follows: (33) The above parameters can be used to construct a set of polyhedral uncertainties with statistical confidence guarantees. : (34) In the formula: Indicates the sorted order of the first... One data point; The radius parameter of the polyhedron set.

[0045] Step 3: Based on the load modeling results, extract the main problem and sub-problems from the robustly optimized charging station capacity clearing model, construct the feasibility cutting plane constraint, and solve and optimize the main problem and sub-problems based on the improved Benders dual decomposition method to obtain the optimal solutions for capacity allocation, generator output, and rescheduled output.

[0046] This algorithm solves optimization problems based on an improved Benders dual decomposition method. It adapts the coupling relationship between uncertain variables and decision variables through improved feasibility cut constraints, avoiding cut failures and global non-convergence caused by changes in the shape of the Dependent Dual Variable (DDU) in traditional methods. The optimization problem-solving method based on the improved Benders dual decomposition method specifically includes: Step 3.1, Problem Division; specifically including: Step 3.1.1, Main Problem Division: Based on the above transformation, the main problem (23) of the original problem can be further equivalently represented as this single-layer problem, in the following form: (39) (14) (15) (38) The improved Benders dual decomposition algorithm solves the main problem (39) and subproblems (35) repeatedly until the feasibility of the subproblems meets the preset convergence accuracy. This means obtaining the robust optimal solution.

[0047] Step 3.1.2, Sub-problem division: Based on the feasible region of the subproblem feedback, optimize generator output, EVCS day-ahead purchase volume and capacity reduction coefficient. Subproblems (20)-(22) do not have objective functions, so only the feasibility problem needs to be solved.

[0048] To facilitate the solution, the inner problem is dualized, resulting in the following single-layer bilinear optimization subproblem: (35) (36) (37) In the formula: yes and The corresponding matrix form of the dual variables, where These are slack variables used to relax the feasible region of subproblems; Let be the feasible region of the dual variable.

[0049] The improved Benders dual decomposition algorithm divides the original two-phase problem into two parts, and the subproblems need to pass the solved dual variables to the main problem. The feasibility of cutting plane constraints is then fed back to the main problem.

[0050] The feasibility cutting plane constraints for the k-th iteration of the algorithm are shown below: (38) In the formula: and The value of the dual variable returned for the k-th iteration subproblem; Let be the random variable of the EVCS load in the k-th iteration; This represents the total number of feasible cutting plane constraints. This indicates that all feasible cutting plane constraints are traversed.

[0051] Step 3.2, Problem Refactoring: Step 3.2.1, Duality and Linearization of the Main Problem Constraints: Replace the feasibility robustness constraint (38) in the main problem with an equivalent dual form: (40) (41) The inner problem of equation (40) is transformed using KKT conditions: (42) (43) In the formula: is the dual variable of constraint (41); constraint (43) represents the complementary relaxation condition.

[0052] Introducing binary variables By performing the Big-M transformation, the nonlinearity of the complementary condition (43) can be eliminated, resulting in the final mixed-integer linear programming form: (44) (45) (46) (47) (48) In the formula: the complementary relaxation condition (43) is replaced by constraints (46) and (47); M is a sufficiently large constant.

[0053] The mixed-integer programming form of the main problem (40) is as follows: (49) (50) In the formula: Formula (50) is the feasibility cutting plane constraint for each iteration of the algorithm.

[0054] Step 3.2.2: Linearization of the objective function of the subproblem: Since the subproblem (35) for determining robust feasibility includes a bilinear objective function term. The solution is difficult. Therefore, the subproblem will be restructured.

[0055] Subproblem (35) can be equivalently transformed into the following two-level problem: (51) In the formula: Solving the main problem yields The value of .

[0056] The inner problem of equation (51) can be transformed using KKT conditions: (52) (53) In the formula: It is the dual variable of constraint (37); The main problem decision variables The solution.

[0057] Introducing binary variables By performing the Big-M transformation, the nonlinearity of the complementary condition (53) can be eliminated, resulting in the final mixed-integer linear programming form: (54) (55) (56) (57) In the formula: the complementary condition (53) is replaced by constraints (55) and (56).

[0058] The subproblem is equivalent to the following mixed integer programming problem: (58) (59) (60) (61) (62) (63) Step 3.3, Problem Solving: Step 3.3.1, Algorithm Initialization: Input parameter data and convergence tolerance. Initialize the number of iterations Number of cutting plane constraints The cutting plane constraint set is initially an empty set.

[0059] Step 3.3.2: Solve the main problem: update the number of iterations. for Solve the current master problem (49) to obtain the solution of the decision variables in the current step. .

[0060] Step 3.3.3: Solve the subproblems and verify their feasibility: Apply the solutions obtained from the main problem to the subproblems. Input the subproblem (58) to verify the robustness of the current solution and obtain the feasibility under the worst-case scenario. If the maximum value of the subproblem is... Meet the conditions This indicates that the current solution does not satisfy the robust feasibility constraint, and a new cutting plane constraint (50) needs to be generated and fed back to the main problem to update the cutting plane constraint. for If the iteration fails, return to step 3.3.2 and continue iterating; otherwise, proceed to step 3.3.4.

[0061] Step 3.3.4, Termination Condition and Optimal Solution Output: If the convergence condition is met... If the feasibility of the subproblem is verified, the solution to the current main problem is the robust and feasible optimal solution, the iteration terminates, and the current optimal solution is output.

[0062] In a practical embodiment, to verify the effectiveness of the proposed capacity market clearing method, such as... Figure 1 As shown, a simulation analysis is conducted using a 12-node transportation network and an IEEE 33-node distribution network. The transportation network data primarily provides the charging load distribution for the distribution network. The electricity purchase cost coefficient for charging stations is also included. The cost is 0.4 yuan / (kW·h), and the capacity revenue factor is... The rate is 0.2 yuan / (kW·h), the per-unit voltage of the first node is set to 1.0, and the time period is... Take 1 hour. The three electric vehicle charging stations are connected to nodes 2, 6, and 19 respectively. Other power distribution network and charging station parameters are detailed below. Figure 2 As shown.

[0063] To analyze the economic benefits of market clearing at various capacities and the impact of uncertain set envelope shapes, four simulation scenarios were set up, such as... Figure 3 As shown. Scenario 1 is the traditional day-ahead capacity clearing method, where free reduction means that the reduction coefficients for different charging stations can be different. Scenario 4 is the capacity clearing method under uncertain conditions presented in this paper. Scenario 2 and 3 are comparative scenarios with differentiated conditions to verify the effectiveness of the proposed method. The steps are as follows: Input the parameters of the transportation network, power distribution network, and charging stations; Construct a day-ahead distribution network cost minimization model based on step 1.1; Construct the day-ahead rescheduling model based on step 1.2; Perform load uncertainty modeling based on step 2; The problems are divided according to step 3.1; Perform constraint duality and linearization according to step 3.2; The capacity allocation problem is solved according to step 3.3, and the results of capacity allocation, generator output, and rescheduled output are obtained.

[0064] The results are shown in the table below:

[0065] The beneficial effects of this invention are as follows: Compared with the prior art, this invention provides a day-ahead market-based allocation method and system for charging station capacity under distribution network security constraints. It introduces a market-based capacity allocation mechanism based on bid prices, ensuring grid operation safety while guiding charging stations to participate in competition based on their true capacity value, thereby achieving optimal resource allocation efficiency. To address the uncertainty of EVCS load, a two-stage robust optimization capacity clearing model considering distribution network capacity constraints is constructed. A polyhedral uncertainty set based on the Statistical-Guarantee-based Vertex Link (SGVL) algorithm is built to describe the prediction error of electric vehicle charging load, minimizing the impact of load uncertainty. An improved Benders dual decomposition algorithm is used, with improved feasibility cut constraints to adapt to the coupling relationship between uncertain variables and decision variables. This avoids the cut failure and global non-convergence problems caused by DDU shape changes in traditional methods. While reducing system operating costs and rescheduling requirements, it effectively reduces the model's conservatism and improves the overall robustness and economy of the system.

[0066] This invention can be a system, method, and / or computer program product. This invention also discloses a charging station capacity day-ahead market-based allocation system under distribution network security constraints, based on the aforementioned method for day-ahead market-based allocation of charging station capacity under distribution network security constraints, comprising: The charging station capacity clearing model construction module is used to obtain various parameters of the transportation network, distribution network and charging stations. In the day-ahead stage, the charging capacity market is introduced, and the charging station capacity clearing model is constructed based on the distribution network constraints. It includes the day-ahead distribution network cost minimization model in the first stage and the day-ahead rescheduling model in the second stage. Robust optimization is performed on the two-stage model. The load modeling module is used to construct an ellipsoidal uncertainty set based on the original simulated charging load dataset, and transform the ellipsoidal uncertainty set into a polyhedral uncertainty set based on the vertex linking algorithm SGVL, so as to realize data-driven load modeling in the uncertainty environment for describing the prediction error of electric vehicle charging load. The Benders dual decomposition solution module is used to extract the main problem and subproblems from the robustly optimized charging station capacity clearing model based on the load modeling results, construct feasibility cutting plane constraints, and solve and optimize the main problem and subproblems based on the improved Benders dual decomposition method to obtain the optimal solutions for capacity allocation, generator output, and rescheduled output.

[0067] Based on the spirit of this invention, those skilled in the art will readily conceive of a computer program product that can be derived from the aforementioned method for day-ahead market-based allocation of charging station capacity under distribution network security constraints. The computer program product may include a computer-readable storage medium on which computer-readable program instructions are loaded to enable a processor to implement various aspects of this disclosure. That is, this application also includes a terminal comprising a processor and a storage medium; the storage medium is used to store instructions; the processor is used to operate according to the instructions to execute the steps of the aforementioned method for day-ahead market-based allocation of charging station capacity under distribution network security constraints.

[0068] Computer-readable storage media can be tangible devices capable of holding and storing instructions for use by an instruction execution device. Computer-readable storage media can be, for example, but not limited to, electrical storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory sticks, floppy disks, mechanical encoding devices, such as punch cards or recessed protrusions storing instructions thereon, and any suitable combination of the foregoing. The computer-readable storage media used herein are not to be construed as transient signals themselves, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., light pulses through fiber optic cables), or electrical signals transmitted through wires.

[0069] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.

[0070] Computer program instructions used to perform the operations of this disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, status setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The computer-readable program instructions may execute entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing the status information of the computer-readable program instructions to implement various aspects of this disclosure.

[0071] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. A method for day-ahead market-based allocation of charging station capacity under distribution network security constraints, characterized in that, Includes the following steps: Step 1: Obtain various parameters of the transportation network, distribution network, and charging stations. Introduce the charging capacity market in the day-ahead phase. Construct a charging station capacity clearing model based on distribution network constraints, including a day-ahead distribution network cost minimization model in the first phase and a day-ahead rescheduling model in the second phase. Perform robust optimization on the two-stage models. Step 2: Construct an ellipsoidal uncertainty set based on the original simulated charging load dataset, and transform the ellipsoidal uncertainty set into a polyhedral uncertainty set based on the vertex linking algorithm SGVL, so as to realize data-driven load modeling in an uncertain environment for describing the prediction error of electric vehicle charging load; Step 3: Based on the load modeling results, extract the main problem and sub-problems from the robustly optimized charging station capacity clearing model, construct the feasibility cutting plane constraint, and solve and optimize the main problem and sub-problems based on the improved Benders dual decomposition method to obtain the optimal solutions for capacity allocation, generator output, and rescheduled output.

2. The method for day-ahead market-based allocation of charging station capacity under distribution network security constraints according to claim 1, characterized in that, The objective function of the day-ahead distribution network cost minimization model is: （1） In the formula: This is the power generation cost coefficient; Represents a node The day-ahead forecast of generator output; the power generation cost of the distribution network only considers the cost of purchasing electricity from the upstream power grid. For a period of time; This is the EVCS capacity reduction ratio; For charging stations The day-ahead forecast of electricity purchase capacity; , These are the revenue coefficient per unit capacity and the EVCS electricity purchase cost coefficient, respectively. Represents a node The maximum capacity of the EVCS; G and M are the generator and EVCS sets, respectively.

3. The method for day-ahead market-based allocation of charging station capacity under distribution network security constraints according to claim 2, characterized in that, The day-ahead distribution network cost minimization model satisfies the capacity reduction factor and power purchase constraints, as well as the distribution network power flow constraints: The capacity reduction factor and the power purchase constraint are as follows: （2） （3） The power flow constraints of the distribution network are as follows: （4） （5） （6） （7） （8） （9） In the formula: It is the set of all branches in the distribution network; It is the set of all nodes in the power distribution network; and They are nodes The active and reactive power output of the generator; and They are nodes The active and reactive power of the load; and Branch roads Active power and reactive power; and Branch roads Active power and reactive power; For nodes The square of the voltage amplitude; For nodes The square of the voltage amplitude; branch road The resistance; branch road The reactance; branch road Active power; branch road reactive power; and They are nodes The lower and upper limits of the square of the voltage amplitude; and They are nodes The lower and upper limits of generator active power output; and They are nodes The lower and upper limits of generator reactive power output.

4. The method for day-ahead market-based allocation of charging station capacity under distribution network security constraints according to claim 3, characterized in that, The objective function of the day-ahead rescheduling model is: （13） （14） （15） In the formula: , , , , , It is a coefficient matrix; This is the maximum capacity vector for EVCS.

5. The method for day-ahead market-based allocation of charging station capacity under distribution network security constraints according to claim 4, characterized in that, The robust optimization of the two-stage model further includes: calculate （23） （14） （15） （20）-（21） In the formula: This is the region where rescheduling is feasible.

6. The method for day-ahead market-based allocation of charging station capacity under distribution network security constraints according to claim 5, characterized in that, Step 2 further includes: Step 2.1: Generate the dataset of the original simulated charging load. Randomly divided into shape learning datasets With size calibration dataset The sample sizes of each part satisfy the following relationship: （24） In the formula: For dataset The total number of samples, and Each dataset and The sample size It is a positive integer. For the allowed probability of violation, For statistical confidence complement parameters, This is the floor operator; Step 2.2, using Construct the initial set of uncertainties for the ellipsoid; the mean vector of the dataset is denoted as... The covariance matrix is ​​denoted as The ellipsoid set can be represented as: （25） In the formula: This represents the actual charging load vector. This is the mean vector of the simulated load data; The covariance matrix of the simulated load; The simulated load vector is obtained after rotation transformation: （27） In the formula: It is a rotation matrix composed of the eigenvectors of the covariance matrix. A diagonal matrix composed of eigenvalues. This is the transformed load simulation vector; Step 2.3: Use the vertex linking method to convert the above ellipsoid set into a polyhedron set; the ellipsoid has Given vertices, construct a matrix of all positive and negative semi-axis vertices of the ellipsoid set. , : （28） By combining the positive and negative forms of each column in the two matrices above, we can generate... The combination matrix of each vertex W Further construct the hyperplane parameter vector of the polyhedron set. : （29） In the formula: For the first A combination matrix of vertices, This is the corresponding hyperplane normal vector; Step 2.4: Utilize the validation dataset Calibrate the radius parameters of the polyhedron set for each Define its projection value onto the polyhedron set: （30） In the formula: For data points The mapping value; Sort all projection values ​​in ascending order to get: （31） Calculate the critical parameters used to determine the radius of the polyhedral assembly. ,satisfy: （32） In the formula: Number the loads in ascending order; To meet the confidence level With probability of violation Critical parameters; Represents the number of combinations, from Select from data The number of combinations of data points; Define the radius parameter of the polyhedron set as: （33） The above parameters are used to construct a set of polyhedral uncertainties with statistical confidence guarantees. : （34） In the formula: Indicates the sorted order of the first... One data point; The radius parameter of the polyhedron set.

7. The method for day-ahead market-based allocation of charging station capacity under distribution network security constraints according to claim 6, characterized in that, The improved Benders dual decomposition method optimizes the solution of the main problem and subproblems, further including: Step 3.3.1: Input parameter data and convergence tolerance. Initialize the number of iterations Number of cutting plane constraints The cutting plane constraint set is initially empty; Step 3.3.2, Number of Update Iterations for Solve the current main problem: （49） Obtain the solution of the decision variables in the current step. ; Step 3.3.3: Apply the solution obtained from the main problem... Input a subproblem to solve to verify the robustness of the current solution: （58） If the maximum value of the subproblem Meet the conditions If the current solution does not satisfy the robust feasibility constraint, a new cutting plane constraint is generated and fed back to the main problem. Update cutting plane constraints for If the iteration fails, return to step 3.3.2 and continue iterating; otherwise, proceed to step 3.3.

4. Step 3.3.4: If the convergence condition is met If the feasibility of the subproblem is verified, the solution to the current main problem is the robust and feasible optimal solution, the iteration terminates, and the current optimal solution is output.

8. A day-ahead market-based allocation system for charging station capacity under distribution network security constraints, characterized in that, include: The charging station capacity clearing model construction module is used to obtain various parameters of the transportation network, distribution network and charging stations. In the day-ahead stage, the charging capacity market is introduced, and the charging station capacity clearing model is constructed based on the distribution network constraints. It includes the day-ahead distribution network cost minimization model in the first stage and the day-ahead rescheduling model in the second stage. Robust optimization is performed on the two-stage model. The load modeling module is used to construct an ellipsoidal uncertainty set based on the original simulated charging load dataset, and transform the ellipsoidal uncertainty set into a polyhedral uncertainty set based on the vertex linking algorithm SGVL, so as to realize data-driven load modeling in the uncertainty environment for describing the prediction error of electric vehicle charging load. The Benders dual decomposition solution module is used to extract the main problem and subproblems from the robustly optimized charging station capacity clearing model based on the load modeling results, construct feasibility cutting plane constraints, and solve and optimize the main problem and subproblems based on the improved Benders dual decomposition method to obtain the optimal solutions for capacity allocation, generator output, and rescheduled output.

9. A terminal, comprising a processor and a storage medium; characterized in that: The storage medium is used to store instructions; The processor is configured to operate according to the instructions to execute the steps of the day-ahead market-based allocation method for charging station capacity under distribution network security constraints as described in any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements the steps of the day-ahead market-based allocation method for charging station capacity under the security constraints of the distribution network as described in any one of claims 1-7.

Images