Active power distribution network multi-objective optimization method based on three-port flexible multi-state switch

Abstract

An active power distribution network multi-objective optimization method based on a three-port flexible multi-state switch comprises the steps of collecting historical operation data of an active power distribution network, analyzing and sorting the historical data to obtain conventional data, accessing the three-port flexible multi-state switch on the basis, and establishing a corresponding optimization function, the compensation power of each port is determined through an optimization algorithm, and centralized optimization of the problems of active power distribution network distributed energy consumption, voltage out-of-limit and the like is completed. According to the scheme, multi-target centralized optimization of the active power distribution network can be realized, and the stable operation capability of the active power distribution network under the access of the distributed power supply is improved.

Description

Technical Field

[0001] This invention relates to the field of power grid optimization technology, and in particular to an active distribution network multi-objective optimization method based on a three-port flexible multi-state switch. Background Technology

[0002] The safe and reliable operation of active distribution networks is crucial to the overall efficiency of the power grid and the quality of power supply for users. Currently, due to the output fluctuations and uneven distribution of distributed generation sources, high grid connection rates can lead to problems such as voltage exceeding limits, harmonic pollution, voltage fluctuations and flicker, line overload, and increased short-circuit current, jeopardizing the safe and stable operation of active distribution networks. Existing literature indicates that voltage exceeding limits and line losses are currently the most common factors limiting the capacity of distribution networks to absorb distributed energy. However, traditional active distribution network optimization methods are insufficient to meet the demands of increasingly complex active distribution networks. Therefore, researching novel active distribution network optimization and management methods is urgently needed.

[0003] With the development of new power electronics technologies, the advantages of flexible multi-state switches (FPS) have gradually become apparent, making them a research hotspot both domestically and internationally. Existing literature has modeled and analyzed issues such as reactive power optimization, distributed generation absorption, feeder power flow regulation, and voltage management in distribution networks after the integration of FPS, fully demonstrating the various advantages brought about by FPS participation in active distribution network operation optimization. However, multi-objective optimization has not been conducted for active distribution networks with three-port FPS integration.

[0004] Therefore, this invention, based on the consideration of flexible multi-state switch access to active distribution network, studies a multi-objective optimization method for active distribution network to achieve optimized management of active distribution network under three-port flexible multi-state switch access. Summary of the Invention

[0005] The purpose of this invention is to provide a stable and reliable optimization and management method for active distribution networks, thereby improving the safe and reliable operation capability of active distribution networks. This invention proposes a multi-objective optimization method for active distribution networks based on a three-port flexible multi-state switch.

[0006] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: A multi-objective optimization method for active distribution networks based on a three-port flexible multi-state switch includes the following steps: Step 1: Collect historical operating data of the active distribution network, and organize, analyze, and process the data; Step 2: Integrate a three-port flexible multi-state switch into the active distribution network, establish a multi-objective optimization model for the active distribution network, and propose corresponding constraints; Step 3: Solve the multi-objective optimization function for the active distribution network; Step 4: Using the solution obtained in Step 3, adjust the active distribution network and the corresponding flexible multi-state switch data to complete the optimization of the active distribution network.

[0007] The method is used to improve the safe and reliable operation capability of active distribution networks.

[0008] In step 1, when organizing and analyzing historical data of the active distribution network, the historical data includes photovoltaic output curves, load peak and valley records, and extreme event data; When processing data, the preprocessing includes outlier removal, missing value completion, and standardization. Standardization uses Z-score normalization to map the data to a distribution with a mean of 0 and a standard deviation of 1. In step 2, the AC side of the three-port flexible multi-state switch is connected to the AC distribution network, and the DC side is connected in parallel through a DC capacitor; the distributed power source is connected between each feeder node of the AC distribution network; the AC distribution network includes a voltage source converter, and the AC port of the three-port flexible multi-state switch is connected to the DC side capacitor through the voltage source converter; the capacity of the voltage source converter is designed according to the magnitude of the active power and reactive power adjusted by each port of the three-port flexible multi-state switch.

[0009] In step 2, the optimization objectives of the active distribution network multi-objective optimization model are: minimum voltage fluctuation, minimum network loss, maximum DG absorption level, and maximum feeder load balance.

[0010] In step 2, a multi-objective optimization model for the active distribution network is established, specifically as follows: The formula for calculating the objective function of voltage fluctuation is as follows: ; In the formula, T For the total time period in dynamic power flow optimization; n This represents the total number of nodes in the distribution network. U i ( t )for t Time period nodes i The voltage amplitude; U max.i , U min.i They are nodes i Maximum and minimum voltage amplitude.

[0011] The formula for calculating the network loss objective function is as follows: ; Where: Ω i For nodes i The set of adjacent nodes; Iij ( t ), R ij Representing branch roads ij exist t The current amplitude and resistance during the time period.

[0012] The objective function for calculating the DG absorption level is as follows: ; In the formula, S DG This refers to the set of times when the distributed generation (DG) generates power in an active distribution network. P DG.t.max It was in DG t The maximum allowable output power at any given time; P DG.t For distributed power sources t Actual grid-connected power at any time.

[0013] The formula for calculating the objective function of feeder load balance is as follows: ; ; In the formula, T a For feeder a Load rate; P T.a For feeder a The power supply capacity of the medium power supply; S N.a For feeder a Rated capacity; cos φ a For feeder a The power factor. A This represents the total number of feeders in the active distribution network.

[0014] When comprehensively optimizing these objective functions, the analytic hierarchy process (AHP) is typically used to normalize the objectives. The comprehensive objective function can be expressed as: ; In the formula, λ 1. λ 2. λ 3. λ 4 represent the proportions of the sub-objectives in the overall objective function, and satisfy the following equations: .

[0015] In step 2, the constraints of the multi-objective optimization include: Node power balance constraints: ; In the formula: For the injection of the first i The active power of each node; For the first i The active load of each node; For the injection of the first i The reactive power of each node; For the first i The reactive load of each node; for i , j Electrical conductance between the two nodes; for i, j The susceptance between the two nodes; for i , j Voltage phase angle difference between two nodes; n The number of active distribution network nodes; For nodes i voltage, Let be the voltage at node j; Distributed power generation output constraints: ; In the formula: , For nodes i The active and reactive power output of distributed power sources , For nodes i Lower limits of active and reactive power output of distributed generation. , For nodes i The upper limits of active and reactive power output of distributed power sources; Constraints for the safe and stable operation of flexible multi-state switches: ; In the formula: This refers to the number of communication ports; This is the number of DC ports; Inject AC nodes into each AC port of the flexible multi-state switch. i active power, Inject DC nodes into each DC port of the flexible multi-state switch. j The active power.

[0016] System power flow constraints: ; ; Indicates the first i Each node is a set of all nodes at the end of a branch that has a starting node; Indicates the first i Each node is the set of all nodes at the beginning and end of a branch at the end; , The first t The first time period i The flow from the node to the first k Each node P and Q ; P Active power Q Reactive power; , branch road ij Reactance and resistance; For the first t The time period flows through the branch road ij The branch current, For the first t Inflow nodes in time period i The active power; For the first t Inflow nodes in time period i reactive power; For the first t Inflow nodes in time period i The active power of distributed energy sources; For the first t Inflow nodes in time period i The active power consumed by the load; For the first t Inflow nodes in time period i Distributed energy reactive power; For the first t Inflow nodes in time period i The reactive power consumed by the load; Branch capacity constraints: ; branch road ij The upper limit of the current amplitude; , branch road ij Active power and reactive power flowing through; branch road ij The current amplitude, For nodes i No. t Voltage over a time period; branch road ij active power, branch road ij The reactive power.

[0017] In step 3, an improved *Symplocos swarm* algorithm is used to solve the multi-objective optimization function of the active distribution network. This mainly involves solving the constructed multi-objective optimization model under the aforementioned constraints using the improved *Symplocos swarm* algorithm. The individual positions of the *Symplocos swarm* are determined by the optimization variables, namely the active and reactive power command values ​​of the three ports of the flexible multi-state switch. The position of each individual corresponds to a set of optimization variable values, as shown below. X =[ p 1, p 2, p 3, q 1, q 2 , q 3] T The six variables represent the active and reactive power transmitted at the three ports of the flexible multi-state switch. The solution process includes the following steps; Step S1: Population initialization, setting the population size. N Maximum number of iterations L The current iteration number is l dimensionality D upper and lower bounds of search space ub , vb ; Step S2: Initialize the positions of individuals in the tunicate swarm within the search range according to the following formula, calculate the fitness of each individual, and after initialization, the position of each individual corresponds to a set of optimization variable values. Calculate the individual fitness using the individual positions. Fitness is the objective function value obtained by substituting the individual positions into the established active distribution network operation optimization model containing three-port flexible multi-state switches and distributed power sources, and performing power flow calculations. Then, compare the fitness of each individual position after initialization, and set the individual with the lowest fitness as the food source position. ; Step S3: Set the location of the individual with the best fitness as the food source location, designate the tunicate individual with the best fitness as the leader, and designate all other individuals as followers. Calculate the fitness of each individual and the number of iterations. l = l +1; Step S4: Calculate the convergence factor and the leader range weakening factor respectively, update the current position of the leader using the leader position update formula, and use the ensemble mutation strategy to update the current position of the individual followers of the tunicate slug swarm using the follower position update formula. Step S5: Compare the fitness of each individual tunicate after the update with the fitness of the current food source location; Step S6: Determine whether the fitness of the individual salps has reached the optimal level, and update the food source location based on the fitness value, i.e., always select the location corresponding to the optimal fitness value as the food source location. Step S7: Repeat the above iterative process until the set maximum number of iterations is reached. After the termination condition is met, output the current food position as the estimated position of the target. The values ​​of each dimension of this position are the final values ​​of the optimization variables.

[0018] In step S4, the leader's position was updated using the following formula: ; In the formula: Indicates the first l During the nth iteration, the 1st d The current location of the leader of the Uyghur people, the sea squirt. For the first d The location of food in the Middle Kingdom , The first d The upper and lower bounds of the predation space in the dimensional space. , Two random numbers are generated between [0,1], with Δ=0.5; l , L These represent the current iteration number and the maximum iteration number, respectively. Let be the convergence factor, and its value is determined by the following formula: ; The range reduction factor is determined by the following formula; .

[0019] In step S4, for followers, an adaptive ensemble mutation strategy is introduced during the follower position update phase. The specific formula for follower position update is as follows: ; In the formula: k The individual attenuation factor for the followers is denoted by , and follows an exponential distribution with a parameter of 0.5. It is the first l During the nth iteration, the 1st d Weizhongdi n The location of individual salps The corresponding fitness; For the first d Weizhongdi n -1 location of a salps individual.

[0020] In step S7, the number of iterations is predetermined based on the problem size.

[0021] Compared with the prior art, the present invention has the following technical effects: This invention studies the multi-objective optimization and governance of active distribution networks based on three-port flexible multi-state switches. By adopting an improved tunic group algorithm, it solves the current multi-objective optimization problem of active distribution networks, improves the safe and reliable operation capability of active distribution networks, and completes the simultaneous optimization of multiple objectives of active distribution networks. This is of great significance to the construction and operation of active distribution networks. Attached Figure Description

[0022] The present invention will be further described below with reference to the accompanying drawings and embodiments: Figure 1 This is a flowchart of an active distribution network multi-objective optimization governance method provided by an embodiment of the present invention; Figure 2 This is a topology diagram of an active distribution network connected to a three-port flexible multi-state switch; Figure 3 This is a flowchart of the improved tunicate group algorithm. Detailed Implementation

[0023] This invention provides a multi-objective optimization method for active distribution networks based on a three-port flexible multi-state switch, which can achieve simultaneous optimization of multiple objectives of the active distribution network and further improve the safe and reliable operation capability of the active distribution network.

[0024] The system of this invention includes an active distribution network with a three-port flexible multi-state switch. Its main components include: distributed photovoltaic power generation, an active distribution network, a three-port flexible multi-state switch, a DC capacitor, and an AC distribution network. The three-port flexible multi-state switch is connected to the AC distribution network, with its DC side connected in parallel via a DC capacitor. The distributed power source is connected between each feeder node of the AC distribution network.

[0025] like Figure 1 As shown, the method proposed in this invention includes the following steps: Step 1: Collect historical operating data of the active distribution network, and organize and analyze it; Step 2: Integrate a three-port flexible multi-state switch into the active distribution network, establish a multi-objective optimization model for the active distribution network, and propose corresponding constraints; Step 3: Solve the multi-objective optimization function for the active distribution network; Step 4: Use the solution results to adjust the active distribution network and the corresponding flexible multi-state switch data to complete the optimization of the active distribution network.

[0026] The method is used to improve the safe and reliable operation capability of active distribution networks.

[0027] In step 1, when organizing and analyzing historical data of the active distribution network, the historical data includes, but is not limited to, photovoltaic output curves, load peak and valley records, extreme event data, etc. When processing data, preprocessing includes outlier removal, missing value completion, and standardization. Z-score normalization can be used to map the data to a distribution with a mean of 0 and a standard deviation of 1.

[0028] When employing optimization algorithms, it is necessary to first process the objectives of multi-objective optimization. Because different objective functions have their own unique characteristics and dimensions, it is essential to normalize them using special methods to transform them into a single, comprehensive objective function. Therefore, it is necessary to first determine the weight coefficients of each objective function, and then optimize to achieve the optimal comprehensive result. This type of problem is generally handled using the Analytic Hierarchy Process (AHP), which solves the problem through efficient weight decisions for the objectives.

[0029] Furthermore, in step 2, the AC side of the three-port flexible multi-state switch is connected to the AC distribution network, and the DC side is connected in parallel through a DC capacitor; distributed power sources are connected between each feeder node of the AC distribution network; the capacity of the voltage source converter is designed according to the active and reactive power regulated by each port of the three-port flexible multi-state switch. The AC distribution network should also include a voltage source converter, and the AC port of the three-port flexible multi-state switch is connected to the DC side capacitor through the voltage source converter.

[0030] The voltage source converters VSC1-VSC3 are composed of IGBTs and three-phase controllable rectifier bridges controlled by SVPWM. The AC port of the three-port flexible multi-state switch is connected to the DC side capacitor through the voltage source converter.

[0031] Furthermore, it also includes the following steps: The capacity of the voltage source converter is designed based on the active and reactive power regulated at each port of the three-port flexible multi-state switch.

[0032] Before addressing voltage over-limit issues, the active distribution network needs to be initialized, including the design of the voltage source converter capacity. To improve power supply reliability, the maximum active and reactive power that the active distribution network can regulate under distributed power source access is used as the VSC capacity. At the same time, other parameters are designed to achieve the initialization of the active distribution network.

[0033] Furthermore, the optimization objectives of the multi-objective optimization in step 2 include: minimizing voltage fluctuations, minimizing network losses, maximizing DG absorption capacity, and maximizing feeder load balance.

[0034] According to the multi-objective optimization model, the optimization objectives include: The formula for calculating the objective function of voltage fluctuation is as follows: ; In the formula, T For the total time period in dynamic power flow optimization; n This represents the total number of nodes in the distribution network. U i ( t )for t Time period nodes i The voltage amplitude; U max.i , U min.i They are nodes i Maximum and minimum voltage amplitude.

[0035] The formula for calculating the network loss objective function is as follows: ; Where: Ω i For nodes i The set of adjacent nodes; I ij ( t ), R ij Representing branch roads ij exist t The current amplitude and resistance during the time period.

[0036] The objective function for calculating the DG absorption level is as follows: ; In the formula, S DG This refers to the set of times when the distributed generation (DG) generates power in an active distribution network. P DG.t.max It was in DG t The maximum allowable output power at any given time; P DG.t For distributed power sources t Actual grid-connected power at any time.

[0037] The formula for calculating the objective function of feeder load balance is as follows: ; ; In the formula, T a For feeder a Load rate; P T.a For feeder a The power supply capacity of the medium power supply; S N.a For feeder a Rated capacity; cosφ a For feeder a The power factor. A This represents the total number of feeders in the active distribution network.

[0038] When comprehensively optimizing these objective functions, the analytic hierarchy process (AHP) is typically used to normalize the objectives. The comprehensive objective function can be expressed as: ; In the formula, λ 1. λ 2. λ 3. λ 4 represent the proportions of the sub-objectives in the overall objective function, and satisfy the following equations: .

[0039] Furthermore, the constraints of the multi-objective optimization in step 2 include: Node power balance constraints: ; In the formula: For the injection of the first i The active power of each node; For the first i The active load of each node; For the injection of the first i The reactive power of each node; For the first i The reactive load of each node; for i , j Electrical conductance between the two nodes; for i, j The susceptance between the two nodes; for i , j Voltage phase angle difference between two nodes; n The number of active distribution network nodes; For nodes i voltage, Let be the voltage at node j; Distributed power generation output constraints: ; In the formula: , For nodes i The active and reactive power output of distributed power sources , For nodes i Lower limits of active and reactive power output of distributed generation. , For nodes i The upper limits of active and reactive power output of distributed power sources; Constraints for the safe and stable operation of flexible multi-state switches: ; In the formula: This refers to the number of communication ports; This is the number of DC ports; Inject AC nodes into each AC port of the flexible multi-state switch. i active power, Inject DC nodes into each DC port of the flexible multi-state switch. j The active power.

[0040] System power flow constraints: ; ; Indicates the first i Each node is a set of all nodes at the end of a branch that has a starting node; Indicates the first i Each node is the set of all nodes at the beginning and end of a branch at the end; , The first t The first time period i The flow from the node to the first k Each node P and Q ; P Active power Q Reactive power; , branch road ij Reactance and resistance; For the first t The time period flows through the branch road ij The branch current, For the first t Inflow nodes in time period i The active power; For the first t Inflow nodes in time period i reactive power; For the first t Inflow nodes in time period i The active power of distributed energy sources; For the first t Inflow nodes in time period i The active power consumed by the load; For the first t Inflow nodes in time period i Distributed energy reactive power; For the first t Inflow nodes in time period i The reactive power consumed by the load; Branch capacity constraints: ; branch road ij The upper limit of the current amplitude; , branch road ij Active power and reactive power flowing through; branch road ij The current amplitude, For nodes i No. t Voltage over a time period; branch road ij active power, branch road ij The reactive power.

[0041] These constraints are the rules that must be met when performing multi-objective optimization of active distribution networks. If these constraints are not met or cannot be met, the multi-objective optimization results of the active distribution network are very likely to deviate from the actual operation of the active distribution network, making the optimization results impractical and deviating from the original intention of multi-objective operation optimization of the active distribution network. Therefore, these constraints must be met when performing multi-objective operation optimization of active distribution networks.

[0042] Furthermore, the optimization algorithm in step 3 is mainly an improved *Symplocos swarm* algorithm. The improved *Symplocos swarm* algorithm is used to solve the multi-objective optimization function of the active distribution network. This mainly involves solving the multi-objective optimization model constructed above, under the aforementioned constraints, using the improved *Symplocos swarm* algorithm. The position of each *Symplocos swarm* individual is determined by optimization variables, namely the active and reactive power command values ​​of the three ports of the flexible multi-state switch. The position of each individual corresponds to a set of optimization variable values, and the *Symplocos swarm* individual positions are represented as follows: X =[ p 1, p 2, p 3, q 1, q 2 ,q 3] T The six variables represent the active and reactive power transmitted at the three ports of the flexible multi-state switch. The solution process includes the following steps; Step S1: Population initialization, setting the population size. N Maximum number of iterations L The current iteration number is l dimensionality D upper and lower bounds of search spaceub , vb ; Step S2: Initialize the positions of individuals in the tunicate swarm within the search range according to the following formula, calculate the fitness of each individual, and after initialization, the position of each individual corresponds to a set of optimization variable values. Calculate the individual fitness using the individual positions. Fitness is the objective function value obtained by substituting the individual positions into the established active distribution network operation optimization model containing three-port flexible multi-state switches and distributed power sources, and performing power flow calculations. Then, compare the fitness of each individual position after initialization, and set the individual with the lowest fitness as the food source position. ; Step S3: Set the location of the individual with the best fitness as the food source location, designate the tunicate individual with the best fitness as the leader, and designate all other individuals as followers. Calculate the fitness of each individual and the number of iterations. l = l +1; Step S4: Calculate the convergence factor and the leader range weakening factor respectively, update the current position of the leader using the leader position update formula, and use the ensemble mutation strategy to update the current position of the individual followers of the tunicate slug swarm using the follower position update formula. Step S5: Compare the fitness of each individual tunicate after the update with the fitness of the current food source location; Step S6: Determine whether the fitness of the individual salps has reached the optimal level, and update the food source location based on the fitness value, i.e., always select the location corresponding to the optimal fitness value as the food source location. Step S7: Repeat the above iterative process until the set maximum number of iterations is reached. After the termination condition is met, output the current food position as the estimated position of the target. The values ​​of each dimension of this position are the final values ​​of the optimization variables.

[0043] In the traditional tunic swarm optimization algorithm, the leader is typically selected based on the individual with the best fitness, and the leader's position update within the search space is unconstrained. However, followers update their positions sequentially according to the leader's position, without considering the influence of followers with better fitness. Therefore, this paper proposes necessary improvements to the traditional tunic swarm optimization algorithm to better meet the requirements of solving the objective function.

[0044] Furthermore, in step S4, the leader's position is updated, and the leader position update formula is: ; In the formula: Indicates the first l During the nth iteration, the 1st d The current location of the leader of the Uyghur people, the sea squirt. For the first d The location of food in the Middle Kingdom , The first d The upper and lower bounds of the predation space in the dimensional space. , Two random numbers are generated between [0,1], with Δ=0.5; l , L These represent the current iteration number and the maximum iteration number, respectively. Let be the convergence factor, and its value is determined by the following formula: ; The range reduction factor is determined by the following formula; ; In the traditional tunic swarm optimization algorithm, the leader's position update in each iteration is unrestricted within the foraging space, leading to reduced accuracy and efficiency. To address this issue, a range reduction factor is introduced during the leader's position update phase, referencing the convergence factor. Since the range reduction factor is a non-linearly decreasing function, the leader's search range is negatively correlated with the number of iterations. The presence of the range reduction factor reduces the movement range of the tunic swarm, thereby effectively improving the algorithm's efficiency.

[0045] Furthermore, in step S4, for followers, an adaptive ensemble mutation strategy is introduced during the follower position update stage. The specific formula for follower position update is as follows: ; In the formula: k The individual attenuation factor for the followers is denoted by , and follows an exponential distribution with a parameter of 0.5. It is the first l During the nth iteration, the 1st d Weizhongdi n The location of individual salps The corresponding fitness; For the first d Weizhongdi n -1 location of a salps individual.

[0046] In the traditional tunic swarm optimization algorithm, the follower position update phase does not consider enhancing the influence weight of individuals in better positions during the tunic swarm's movement. Therefore, an adaptive ensemble mutation strategy is introduced in the follower position update phase. After introducing this strategy, the fitness of two connected followers is used to determine the quality of individuals. Then, an individual weakening factor is used to reduce the influence weight of weaker individuals and enhance the influence weight of stronger individuals. This assists the leader in the search and improves optimization efficiency.

[0047] Furthermore, the number of iterations in step S7 needs to be predetermined based on the problem size, and should also include the following steps: Step S0: Determine the number of iterations based on the size of the problem to be solved.

[0048] Since the improved tunic group algorithm is still an iterative algorithm, the iteration must have a termination point; otherwise, it will loop indefinitely. Therefore, when solving such problems, the number of iterations is usually predetermined based on the size of the problem to be solved, thus determining the algorithm's termination time.

[0049] In the final data adjustment process, adjustments must be made strictly according to the optimization results. However, in reality, achieving complete accuracy is unlikely. Therefore, a certain degree of error is allowed during the data adjustment process, but it must be ensured that the data does not deviate too much from the actual operation of the active distribution network or deviate too much from the operational optimization results; otherwise, adjustments are not permitted based on this data.

[0050] This invention relates to a study on multi-objective optimization and governance of active distribution networks based on a three-port flexible multi-state switch. The simulation was performed using the MATLAB / Simulink platform. All main circuits in the experiment were built in the MATLAB / Simulink simulation platform environment, and the working mode of the flexible multi-state switch port was controlled by SVPWM control.

Claims

1. A multi-objective optimization method for active distribution networks based on three-port flexible multi-state switches, characterized in that, Includes the following steps: Step 1: Collect historical operating data of the active distribution network, and organize, analyze, and process the data; Step 2: Integrate a three-port flexible multi-state switch into the active distribution network, establish a multi-objective optimization model for the active distribution network, and propose corresponding constraints; Step 3: Solve the multi-objective optimization function for the active distribution network; Step 4: Using the solution obtained in Step 3, adjust the active distribution network and the corresponding flexible multi-state switch data to complete the optimization of the active distribution network.

2. The method according to claim 1, characterized in that, The method is used to improve the safe and reliable operation capability of active distribution networks.

3. The method according to claim 1, characterized in that, In step 1, when organizing and analyzing historical data of the active distribution network, the historical data includes photovoltaic output curves, load peak and valley records, and extreme event data; When processing data, the preprocessing includes outlier removal, missing value completion, and standardization. Standardization uses Z-score normalization to map the data to a distribution with a mean of 0 and a standard deviation of 1. In step 2, the AC side of the three-port flexible multi-state switch is connected to the AC distribution network, and the DC side is connected in parallel through a DC capacitor; the distributed power source is connected between each feeder node of the AC distribution network; the AC distribution network includes a voltage source converter, and the AC port of the three-port flexible multi-state switch is connected to the DC side capacitor through the voltage source converter; the capacity of the voltage source converter is designed according to the magnitude of the active power and reactive power adjusted by each port of the three-port flexible multi-state switch.

4. The method according to any one of claims 1 to 3, characterized in that, In step 2, the optimization objectives of the active distribution network multi-objective optimization model are: minimum voltage fluctuation, minimum network loss, maximum DG absorption level, and maximum feeder load balance.

5. The method according to claim 4, characterized in that, In step 2, a multi-objective optimization model for the active distribution network is established, specifically as follows: The formula for calculating the objective function of voltage fluctuation is as follows: ； In the formula, T The total time period in dynamic power flow optimization; n This represents the total number of nodes in the distribution network. U i ( t )for t Time period nodes i The voltage amplitude; U max.i , U min.i They are nodes i Maximum and minimum voltage amplitude; The formula for calculating the network loss objective function is as follows: ； Where: Ω i For nodes i The set of adjacent nodes; I ij ( t ), R ij Representing branches ij exist t The current amplitude and resistance during the time period; The objective function for calculating the DG absorption level is as follows: ； In the formula, S DG This is the set of times when the distributed generation (DG) generates power in an active distribution network. P DG.t.max It was in DG t The maximum allowable output power at any given time; P DG.t For distributed power sources t Actual grid-connected power at any time; The formula for calculating the objective function of feeder load balance is as follows: ； ； In the formula, T a For feeder a Load rate; P T.a For feeder a The power supply capacity of the medium power supply; S N.a For feeder a Rated capacity; cos φ a For feeder a The power factor; A This represents the total number of feeders in the active distribution network. When comprehensively optimizing these objective functions, the analytic hierarchy process (AHP) is used to normalize the objectives; the comprehensive objective function is expressed as: ； In the formula, λ 1. λ 2. λ 3. λ 4 represent the proportions of the sub-objectives in the overall objective function, and satisfy the following equations: 。 6. The method according to claim 5, characterized in that, In step 2, the constraints of the multi-objective optimization include: Node power balance constraints: ； In the formula: For the injection of the first i The active power of each node; For the first i The active load of each node; For the injection of the first i The reactive power of each node; For the first i The reactive load of each node; for i , j Electrical conductance between the two nodes; for i, j The susceptance between the two nodes; for i , j Voltage phase angle difference between two nodes; n The number of active distribution network nodes; For nodes i voltage, Let be the voltage at node j; Distributed power generation output constraints: ； In the formula: , For nodes i The active and reactive power output of distributed power sources , For nodes i Lower limits of active and reactive power output of distributed generation. , For nodes i The upper limits of active and reactive power output of distributed power sources; Constraints for the safe and stable operation of flexible multi-state switches: ； In the formula: This refers to the number of communication ports; This is the number of DC ports; Inject AC nodes into each AC port of the flexible multi-state switch. i active power, Inject DC nodes into each DC port of the flexible multi-state switch j The active power; System power flow constraints: ； 　； Indicates the first i Each node is a set of all nodes at the end of a branch that has a starting node; Indicates the first i Each node is the set of all nodes at the beginning and end of a branch at the end; , The first t The first time period i The flow from the node to the first k Nodes P and Q ; P Active power Q Reactive power; , branch road ij Reactance and resistance; For the first t The time period flows through the branch road ij The branch current, For the first t Inflow nodes in time period i The active power; For the first t Inflow nodes during each time period i reactive power; For the first t Inflow nodes in time period i The active power of distributed energy sources; For the first t Inflow nodes in time period i The active power consumed by the load; For the first t Inflow nodes during each time period i Distributed energy reactive power; For the first t Inflow nodes in time period i The reactive power consumed by the load; Branch capacity constraints: ； branch road ij The upper limit of the current amplitude; , branch road ij Active power and reactive power flowing through; branch road ij The current amplitude, For nodes i No. t Voltage over a time period; branch road ij active power, branch road ij The reactive power.

7. The method according to claim 6, characterized in that, In step 3, an improved *Symplocos swarm* algorithm is used to solve the multi-objective optimization function of the active distribution network. This mainly involves solving the multi-objective optimization model of the active distribution network under the constraints of the multi-objective optimization. The individual positions of the *Symplocos swarm* are determined by optimization variables, namely the active and reactive power command values ​​of the three ports of the flexible multi-state switch. The position of each individual corresponds to a set of optimization variable values. The individual positions of the *Symplocos swarm* are represented as follows: X =[ p 1, p 2, p 3, q 1, q 2 ,q 3] T The six variables represent the active and reactive power transmitted at the three ports of the flexible multi-state switch, respectively; the solution process includes the following steps; Step S1: Population initialization, setting the population size. N Maximum number of iterations L The current iteration number is l dimensionality D upper and lower bounds of search space ub , vb ; Step S2: Initialize the position of each individual in the tunicate swarm within the search range according to the following formula, calculate the fitness of each individual, and after initialization, the position of each individual corresponds to the value of a set of optimization variables; the fitness is the objective function value obtained by substituting the individual position into the established active distribution network operation optimization model containing three-port flexible multi-state switches and distributed power sources and performing power flow calculation. Then, the fitness of each individual's location after initialization is compared, and the individual with the lowest fitness is set as the food source location. ； Step S3: Set the location of the individual with the best fitness as the food source location, the tunicate individual with the best fitness as the leader, and all other individuals as followers; calculate the fitness of each individual and the number of iterations. l = l +1; Step S4: Calculate the convergence factor and the leader range weakening factor respectively, update the current position of the leader using the leader position update formula, and use the ensemble mutation strategy to update the current position of the individual followers of the tunicate slug swarm using the follower position update formula. Step S5: Compare the fitness of each individual tunicate after the update with the fitness of the current food source location; Step S6: Determine whether the fitness of the individual salps has reached the optimal level, and update the food source location according to the fitness value, that is, always select the location corresponding to the optimal fitness as the food source location. Step S7: Repeat the above iterative process until the maximum number of iterations is reached. After the termination condition is met, output the current food position as the estimated position of the target; the values ​​of each dimension of this position are the final values ​​of the optimization variables.

8. The method according to claim 7, characterized in that, In step S4, the leader's position was updated using the following formula: ； In the formula: Indicates the first l During the nth iteration, the 1st d The current location of the leader of the Uyghur people, the sea squirt. For the first d The location of food in the Middle Kingdom , The first d The upper and lower bounds of the predation space in the dimensional space. , Two random numbers are generated between [0,1], with Δ=0.5; l , L These represent the current iteration number and the maximum iteration number, respectively. Let be the convergence factor, and its value is determined by the following formula: ； The range reduction factor is determined by the following formula; 。 9. The method according to claim 7, characterized in that, In step S4, for followers, an adaptive ensemble mutation strategy is introduced during the follower position update phase. The specific formula for follower position update is as follows: ； In the formula: k The individual weakening factor of the followers is given, and the followers follow an exponential distribution with a parameter of 0.

5. It is the first l During the nth iteration, the 1st d Weizhongdi n The location of individual salps The corresponding fitness; For the first d Weizhongdi n -1 location of a salps individual.

10. The method according to claim 7, characterized in that, In step S7, the number of iterations is predetermined based on the problem size.

Images