A structural optimization method for active distribution network considering flexible interconnection of intelligent soft switches

Abstract

The present invention provides a method for optimizing the structural form of an active distribution network that considers the flexible interconnection of intelligent soft switches. The method comprises: constructing a clustering framework for representative scenarios of new energy in the active distribution network system, reducing the new energy scenarios in the active distribution network system to obtain clustered representative scenarios; constructing a location candidate set generation framework for the intelligent soft switches; determining factors to be considered in the intelligent soft switch size optimization framework; comprehensively considering these factors to construct a functional relationship expression for the active power loss of the voltage source converter of the intelligent soft switch; establishing operating constraints for the intelligent soft switch; establishing constraints for the active power exchange between the active distribution network system and the upstream power grid; considering the location and size of the intelligent soft switches and various constraints, and combining the functional relationship expression for the active power loss, constructing a structural form optimization model for the active distribution network system; and solving the model to obtain the optimal structural form of the active distribution network system. The present invention can achieve more efficient, reliable, and controllable power distribution.

Description

Technical Field

[0001] The present invention belongs to the field of distribution network structure morphology, and more specifically, is a method for optimizing the structure morphology of an active distribution network taking into account the flexible interconnection of intelligent soft switches. Background Art

[0002] Currently, methods for optimizing the structure of active distribution networks have become a research hotspot in the smart grid field. Related technologies have been widely applied and validated in practice. Intelligent soft switching technology has made significant progress in the intelligent transformation of distribution networks. Traditional hard switches are gradually being replaced by intelligent soft switches, which offer features such as remote control, fault detection, and automatic recovery, enabling a higher level of intelligent management and control. Flexible interconnection technology has also been widely adopted in distribution networks. Flexible interconnection enables flexible connection and communication between nodes in the distribution network, improving system reliability and stability. Furthermore, flexible interconnection enables the effective management and scheduling of new energy resources such as distributed energy and electric vehicles. Methods based on intelligent soft switching and flexible interconnection have been proposed and studied for optimizing the structure of distribution networks. These methods can maximize the performance and efficiency of distribution networks by optimizing line configuration, node layout, and equipment selection.

[0003] Existing research is of great significance to the optimization of the active distribution network structure, but there are still some defects, which are as follows:

[0004] First, research in this area requires a large amount of real-time data to support optimization decisions. This data comes from sensors and monitoring equipment, including information such as current, voltage, and power load. However, acquiring and processing this data requires cost and technical investment, and may also face issues with incomplete or inaccurate data. Existing research has not properly clustered this data, resulting in unrepresentative results and a time-consuming and labor-intensive process.

[0005] Second, distribution networks have complex nonlinear characteristics, such as voltage drop, power loss, and device interconnectivity. Therefore, optimization methods need to account for these nonlinearities and employ appropriate mathematical models and algorithms to solve them. However, solving nonlinear problems is often difficult and time-consuming. Existing research lacks a rational and scientific approach to mathematical modeling, resulting in inefficient problem solving. Summary of the Invention

[0006] In response to the above-mentioned defects or improvement needs of the prior art, the present invention aims to provide a method for optimizing the structural morphology of an active distribution network taking into account the flexible interconnection of intelligent soft switches, aiming to optimize the topological structure of the distribution network to achieve more efficient, reliable and controllable power distribution, thereby achieving the purpose of optimizing the structural morphology of the active distribution network.

[0007] In order to achieve the above object, the present invention discloses a method for optimizing the structure of an active distribution network considering flexible interconnection of intelligent soft switches, comprising the following steps:

[0008] 1) Construct a clustering framework for representative scenarios of new energy in the active distribution network system, reduce the new energy scenarios in the active distribution network system, and obtain representative scenarios after clustering;

[0009] 2) Based on the clustered representative scenarios, a location candidate set generation framework for intelligent soft switches in active distribution network systems is constructed;

[0010] 3) Determine the considerations for the intelligent soft switch sizing optimization framework in active distribution network systems;

[0011] 4) Based on the intelligent soft switch size optimization framework in the active distribution network system, the active power loss function expression of the voltage source converter of the intelligent soft switch is constructed by integrating the considerations determined in step 3);

[0012] 5) Develop an intelligent soft switch size optimization framework for active distribution network systems and establish operating constraints for intelligent soft switches;

[0013] 6) Aiming at the uncertainty scenario of the upstream power grid, construct the active power exchange constraints between the active distribution network system and the upstream power grid;

[0014] 7) Considering the location and size of the intelligent soft switch and various constraints, combined with the active power loss function expression, an active distribution network system structural optimization model is constructed;

[0015] 8) Using the quantum binary particle swarm algorithm fused with the differential evolution algorithm to solve the active distribution network system structure optimization model constructed in step 7) to obtain the optimal active distribution network system structure.

[0016] Furthermore, in step 1), a clustering framework of representative scenarios of new energy in the active distribution network system is constructed, and the new energy scenarios in the active distribution network system are reduced to obtain representative scenarios after clustering, specifically:

[0017] Step 1.1) Randomly select samples {g1,g2,...,g k} as the initial cluster centers {η1,η2,...,η k}, initialize the clustering process;

[0018] Step 1.2) Calculate the new energy operation sample g i and the new energy operation cluster center η j(j＝1,2,...,k) The Euclidean distance between them is as follows:

[0019] d E,ij =||g i -η j || 2 (1)

[0020] Then, according to their respective Euclidean distance d E,ij , each new energy running sample g i Assigned to the nearest new energy operation cluster center η j(j＝1,2,...,k) ;

[0021] Step 1.3) Calculate the mean of the new energy operation samples in each category and assign it as the new new energy operation cluster center;

[0022] Step 1.4) Repeat steps 1.2) and 1.3) until the new energy operation cluster center no longer changes, and then retain the final new energy operation cluster center;

[0023] Step 1.5) The final new energy operation cluster center is used as the representative scenario after clustering.

[0024] Furthermore, in step 2), based on the clustered representative scenarios, a location candidate set generation framework for intelligent soft switches in the active distribution network system is constructed, specifically:

[0025] Step 2.1) Using the data of the lines, loads, and distributed generation of the active distribution network system as input parameters for the power flow calculation, the Newton-Raphson method or the PQ decomposition method is used to calculate the line current of the active distribution network system, the active power and reactive power transmitted between nodes, and the node voltage of each node;

[0026] Step 2.2) Calculate the loss sensitivity index of the active distribution network system, specifically:

[0027]

[0028] in, is the loss sensitivity index of node j, n Φ is the number of phases, Φ is the index of the phase of the active distribution network system, Φ = A, B, C; n s is the total number of representative scenarios of new energy, P ij,Φ,s is the active power of branch ij, phase Φ, and scenario s, Q ij,Φ,s is the reactive power of branch ij, phase Φ, and scenario s, r ij,Φ is the resistance of branch ij, phase Φ, V j,Φ,s is the voltage at node j, phase Φ, and scenario s, β s is the probability of occurrence of the sth representative scenario of new energy;

[0029] Step 2.3) Calculate the normalized node voltage deviation index of the active distribution network system, specifically:

[0030] Calculate the voltage deviation index of each node:

[0031]

[0032] Where, VDI j is the voltage deviation index of node j, n Φ is the number of phases, Φ is the index of the phase of the active distribution network system, Φ = A, B, C; n s is the total number of representative new energy scenarios, V n is the per-unit rated voltage, V j,Φ,s is the voltage at node j, phase Φ, and scenario s, β s is the probability of occurrence of the sth representative scenario of new energy;

[0033] Use the following formula to calculate VDI j Normalization is performed to calculate the normalized node voltage deviation index of the active distribution network system:

[0034]

[0035] Where, VDI j,norm is the normalized voltage deviation index of node j, VDI max All VDI j The maximum value in ;

[0036] Step 2.4) Group the normalized node loss sensitivity index of the active distribution network system and the normalized node voltage deviation index of the active distribution network system (Ω Pos,LSI ,Ω Neg,LSI ) are combined and named and And sort them in descending order according to the weight of the two indices, set a threshold, select the nodes with weight higher than the threshold, and name them Count the number of nodes whose weight is higher than the threshold, and then take the same number of nodes from the column in, that is, The first node in the The nodes of both nodes are part of the search space, where the nodes of both nodes are connected by soft switches, excluding multiple soft switch connections of consecutive nodes on the same side. is the candidate set of locations for smart soft switches in active distribution network systems.

[0037] Furthermore, in step 3), the considerations for developing a size optimization framework for intelligent soft switches in an active distribution network system are as follows:

[0038] Step 3.1) Construct the line power loss function of the active distribution network system:

[0039]

[0040] in, is the line power loss function of the active distribution network system, Φ is the index of the phase of the active distribution network system, Φ = A, B, C, n s is the total number of representative scenarios of new energy, Ω l is the set of network branches, r ij,Φ is the resistance of branch ij, phase Φ, I ij,Φ,s The current of branch ij, phase Φ, and scene s, β s is the probability of occurrence of the sth representative scenario of new energy;

[0041] Step 3.2) Construct the intelligent soft-switching active power loss function of the active distribution network system:

[0042]

[0043] in, is the active power loss function of the intelligent soft switch in the active distribution network system, Ω SOP It is the collection of all intelligent soft switch connection nodes. is the active power loss of the voltage source converter with intelligent soft switching at node i, phase Φ, and scenario s;

[0044] Step 3.3) Construct the distributed generation active power output reduction function of the active distribution network system, specifically:

[0045]

[0046] in, is the distributed generation active power output curtailment function of the active distribution network system, Φ is the phase index of the active distribution network system, Ω DG is the collection of all distributed generation nodes, is the active power reduction of distributed generation at node i, phase Φ, and scenario s, and its calculation formula is:

[0047]

[0048] in, is the active power reduction of distributed generation in node i, phase Φ, and scenario s, is the maximum available output active power of distributed generation at node i, phase Φ, and scenario s, is the dispatched output active power of distributed generation at node i, phase Φ, and scenario s.

[0049] Furthermore, in step 4), for the intelligent soft switch size optimization framework in the active distribution network system, the consideration factors determined in step 3) are integrated to construct the active loss function relationship expression of the voltage source converter of the intelligent soft switch, which is specifically:

[0050] Step 4.1) Construct the active power loss of the voltage source converter of the intelligent soft switch at node i, specifically:

[0051]

[0052] in, is the active power loss of the voltage source converter with intelligent soft switching at node i, phase Φ, and scenario s, is the loss coefficient of the intelligent soft-switching voltage source converter, is the active power of the voltage source converter of the intelligent soft switch at node i, phase Φ, and scenario s, is the reactive power of the voltage source converter of the intelligent soft switch at node i, phase Φ, and scenario s;

[0053] Step 4.2) Construct the active power loss of the voltage source converter of the intelligent soft switch at node j, specifically:

[0054]

[0055] in, is the active power loss of the voltage source converter with intelligent soft switching at node j, phase Φ, and scenario s, is the loss coefficient of the intelligent soft-switching voltage source converter, is the active power of the voltage source converter of the intelligent soft switch at node j, phase Φ, and scenario s, is the reactive power of the voltage source converter of the intelligent soft switch at node j, phase Φ, and scenario s;

[0056] Step 4.3) Construct the active power balance equation of the voltage source converter of the intelligent soft switch connecting node i and node j, specifically:

[0057]

[0058] in, is the active power of the voltage source converter of the intelligent soft switch at node i, phase Φ, and scenario s, is the active power of the voltage source converter of the intelligent soft switch at node j, phase Φ, and scenario s, is the active power loss of the voltage source converter with intelligent soft switching at node i, phase Φ, and scenario s, is the active power loss of the voltage source converter with intelligent soft switching at node j, phase Φ, and scenario s.

[0059] Furthermore, in step 5), the operating constraints of the smart soft switch are constructed for the smart soft switch size optimization framework in the active distribution network system, specifically:

[0060] Step 5.1) Construct the single-phase capacity constraint of the voltage source converter with intelligent soft switching, specifically:

[0061]

[0062] in, is the active power of the voltage source converter of the intelligent soft switch at node i, phase Φ, and scenario s, is the reactive power of the voltage source converter of the intelligent soft switch at node i, phase Φ, and scenario s, is the voltage source converter capacity of the intelligent soft switch for node i, phase Φ, and scenario s;

[0063] Step 5.2) Construct the single-phase capacity expression of the intelligent soft-switching voltage source converter, specifically:

[0064]

[0065] in, is the voltage source converter capacity of the intelligent soft switch for node i, phase Φ, and scenario s, is the three-phase capacity of the voltage source converter of the intelligent soft switch at node i and scenario s;

[0066] Step 5.3) Construct the three-phase capacity expression of the intelligent soft-switching voltage source converter, specifically:

[0067]

[0068] in, is the three-phase capacity of the voltage source converter of the intelligent soft switch of node i and scenario s, Φ is the index of the phase of the active distribution network system, Φ = A, B, C, n s is the total number of representative scenarios of new energy, is the voltage source converter capacity of the smart soft switch for node i, phase Φ, and scenario s, β s is the probability of occurrence of the sth representative scenario of new energy;

[0069] Step 5.4) Construct the operating constraints of the intelligent soft switch, specifically:

[0070]

[0071] in, is the three-phase capacity of the voltage source converter of the intelligent soft switch at node i, which is the final decision variable for the size of the intelligent soft switch. is the lower limit of the three-phase capacity of the intelligent soft-switching voltage source converter, It is the upper limit of the three-phase capacity of the intelligent soft-switching voltage source converter.

[0072] Furthermore, in step 6), for the uncertainty scenario of the upstream power grid, the active power exchange constraint conditions between the active distribution network system and the upstream power grid are constructed, specifically:

[0073] Step 6.1) Construct the active power exchange constraints between the active distribution network system and the upstream power grid, specifically:

[0074]

[0075] in, is the active exchange power between the active distribution network system of node i, phase Φ, scenario s and the upstream power grid, is the lower limit of active power exchange between the active distribution network system at node i and phase Φ and the upstream power grid, is the upper limit of active power exchange between the active distribution network system at node i and phase Φ and the upstream power grid;

[0076] Step 6.2) Construct the reactive power exchange constraints between the active distribution network system and the upstream power grid, specifically:

[0077]

[0078] in, is the reactive exchange power between the active distribution network system of node i, phase Φ, scenario s and the upstream power grid, is the lower limit of reactive power exchange between the active distribution network system at node i and phase Φ and the upstream power grid, is the upper limit of reactive power exchange between the active distribution network system at node i and phase Φ and the upstream power grid;

[0079] Step 6.3) Construct the power factor constraint of the upstream power grid, specifically:

[0080]

[0081] in, is the upstream grid power factor angle for phase Φ and scenario s, is the upstream grid power three-phase factor angle of scenario s, is the average three-phase factor angle of the upstream power grid, is the lower limit of the upstream grid power factor angle limit, It is the upper limit of the upstream grid power factor angle limit.

[0082] Furthermore, in step 7), the position and size of the intelligent soft switch and various constraints are taken into consideration, and combined with the active power loss function relationship expression, an active power distribution network system structure optimization model is constructed, specifically:

[0083] Step 7.1) Considering the location and size of the intelligent soft switch, an optimization model of the active distribution network structure is constructed. Specifically:

[0084] Considering the line power loss, active power loss of intelligent soft switches, and active power output reduction of distributed generation in the active distribution network system, the objective function of the active distribution network system structure optimization model is constructed:

[0085]

[0086] Among them, OF is the size optimization objective function of the intelligent soft switch in the active distribution network system, is the line power loss function of the active distribution network system, is the active power loss function of the intelligent soft switch in the active distribution network system, is the distributed generation active power output curtailment function of the active distribution network system;

[0087] Step 7.2) Construct the constraints of the active distribution network system structure optimization model, specifically:

[0088]

[0089] in, is the active power loss of the voltage source converter with intelligent soft switching at node i, phase Φ, and scenario s, is the loss coefficient of the intelligent soft-switching voltage source converter, is the active power of the voltage source converter of the intelligent soft switch at node i, phase Φ, and scenario s, is the reactive power of the voltage source converter of the intelligent soft switch at node i, phase Φ, and scenario s, is the active power loss of the voltage source converter with intelligent soft switching at node j, phase Φ, and scenario s, is the loss coefficient of the intelligent soft-switching voltage source converter, is the active power of the voltage source converter of the intelligent soft switch at node j, phase Φ, and scenario s, is the reactive power of the voltage source converter of the intelligent soft switch at node j, phase Φ, and scenario s, is the voltage source converter capacity of the intelligent soft switch for node i, phase Φ, and scenario s, is the three-phase capacity of the voltage source converter of the intelligent soft switch of node i and scenario s, Φ is the index of the phase of the active distribution network system, Φ = A, B, C; n sis the total number of representative scenarios of new energy, β s is the probability of occurrence of the sth representative scenario of new energy, is the three-phase capacity of the voltage source converter of the intelligent soft switch at node i, which is the final decision variable for the size of the intelligent soft switch. is the lower limit of the three-phase capacity of the intelligent soft-switching voltage source converter, is the upper limit of the three-phase capacity of the intelligent soft-switching voltage source converter, is the active exchange power between the active distribution network system of node i, phase Φ, scenario s and the upstream power grid, is the lower limit of active power exchange between the active distribution network system at node i and phase Φ and the upstream power grid, is the upper limit of active power exchange between the active distribution network system at node i and phase Φ and the upstream power grid, is the reactive exchange power between the active distribution network system of node i, phase Φ, scenario s and the upstream power grid, is the lower limit of reactive power exchange between the active distribution network system at node i and phase Φ and the upstream power grid, is the upper limit of reactive power exchange between the active distribution network system at node i and phase Φ and the upstream power grid, is the upstream grid power factor angle for phase Φ and scenario s, is the upstream grid power three-phase factor angle of scenario s, is the average three-phase factor angle of the upstream power grid, is the lower limit of the upstream grid power factor angle limit, It is the upper limit of the upstream grid power factor angle limit.

[0090] Furthermore, in step 8), the quantum binary particle swarm algorithm is fused with the differential evolution algorithm to solve the active distribution network system structure optimization model in step (7) to obtain the optimal active distribution network system structure, specifically:

[0091] Step 8.1) Using the position of the smart soft switch and the active and reactive capacities of the voltage source converters at both ends as decision variables, the quantum binary particle swarm optimization (QB-PSO) algorithm is used to solve the position of the smart soft switch in the active distribution network system. Specifically,

[0092] Initialize the particle population, set the particle population size to N, the dimension of each particle to M, and randomly assign each dimension of the particle to 0-1 with a probability of 50%, and get the initial population X of QB-PSO (0) =(x1,x2,…,x N ) (0) ;

[0093] For the t-th iteration result, the variable pbest is used to record the value status of all particles in the population, that is, pbest (t) =X (t) ,t=0,1,2,…,T;

[0094] For the t-th iteration result, the variable mbest is used to record the average value state of all particles in the population. (t) ,t=0,1,2,…,T First, the average of each dimension value of all particles in the population is calculated. If the average value of the dimension is less than 0.5, it is considered that the value of mbest in this dimension is 0. If the average value is greater than 0.5, it is considered that the value of mbest in this dimension is 1. If the average value is exactly 0.5, the value of mbest is determined by random number method.

[0095] For the t-th iteration result, the variable gbest is used to record the value state of the specific particle in the population that makes the fitness function fitness (the opposite number of formula (19)) reach the maximum value, that is,

[0096] Define pbest i The reorganization variable between and gbest is P i The calculation method adopts the "gene recombination" operation in biological genetics. First, the qth dimension of the particle is randomly selected to respectively i and gbest are truncated into four parts, and then recombined into two column vectors. Finally, a random number method is used to determine one of the column vectors as P i The value of

[0097] The binary quantum particle swarm updates the decision variable x according to the following principle i ;

[0098] ④Calculate x i With P i The difference b is calculated as

[0099] b=β·d H (x i ,mbest)·ln(1 / μ),μ=rand(0,1) (21)

[0100] In the formula, β is a constant, d H is the Hamming distance between two vectors, μ is a random number between 0 and 1;

[0101] ⑤Calculate auxiliary variable p r , which is calculated as

[0102]

[0103] Where, l d is the dimension of the particle;

[0104] ⑥Update the decision variable x i , first reorganize the variables into P i Assign to x i , then for each dimension of the decision variable, random numbers and p are used r For comparison, if p r is greater than the random number, then the decision variable x i The values ​​of this dimension are swapped between 0 and 1, otherwise they remain unchanged, and the value of the decision variable x is returned. i (t+1) ;

[0105] After the preset number of iterations T=400, the state gbest of the particle with the best fitness function in the last generation is taken (T) As the intelligent soft switch position optimization result of calling QB-PSO once;

[0106] Step 8.2) Using the intelligent soft switch position and the active and reactive capacities of the two-terminal voltage source converter as decision variables, the differential evolution algorithm (DE) is used to solve the active and reactive capacities of the two-terminal voltage source converter, specifically:

[0107] DE mutation, crossover and selection operations are used to optimize the active and reactive capacity of the two-terminal voltage source converter. For the mutation operation of the i-th DE individual, i=1,2,...,P S , first randomly select three different individuals from the population and Where r1≠r2≠r3≠i, As basis vectors, and After making a difference and scaling Superposition, that is:

[0108]

[0109] Where, is the mutation vector; F is the scaling factor;

[0110] The crossover operation uses the binary crossover method, which acts on The population of the previous iteration Finally, the trial vector is generated Right now:

[0111]

[0112] Where, and Represents vectors and The j-th dimension component, C R is the crossover probability;

[0113] Use the selection operation to select between the offspring and the parent. If the target value of the offspring vector is less than the target value of the parent vector, the offspring vector is retained and the parent vector is eliminated; otherwise, the parent vector is retained and the offspring vector is eliminated:

[0114]

[0115] Through DE mutation, crossover and selection operations, and after repeated iterations, the active and reactive capacities of the two-terminal voltage source converter are optimized and solved, and the optimal solution of the active and reactive capacities of the two-terminal voltage source converter is obtained.

[0116] In general, the above technical solutions conceived by the present invention have the following beneficial effects compared with the prior art:

[0117] 1. Flexibility: This method can flexibly adjust the structure of the distribution network according to network needs and changing conditions to adapt to different load demands and power supply conditions.

[0118] 2. Energy efficiency: By optimizing the distribution network structure, this method can maximize energy utilization efficiency, reduce energy loss and power load imbalance.

[0119] 3. Reliability: This method can achieve intelligent monitoring and automatic control of the distribution network through intelligent soft switching and flexible interconnection technology, thereby improving the reliability and stability of the system.

[0120] 4. Scalability: This method can flexibly expand the scale and capacity of the distribution network according to demand to adapt to future load growth and new energy access.

[0121] 5. Cost-effectiveness: By optimizing the distribution network structure, this method can reduce system operation and maintenance costs and improve the economic benefits of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0122] Figure 1 This is a basic configuration diagram of a distribution network with intelligent soft switches implemented in the present invention;

[0123] Figure 2 It is a modified IEEE 33-node test system adopted in the implementation of the present invention;

[0124] Figure 3 It is a comparison chart of the loss sensitivity index of all nodes implemented by the present invention;

[0125] Figure 4It is a comparison chart of voltage deviation index of all nodes implemented by the present invention;

[0126] Figure 5 It is the result of preliminary candidate node selection using the proposed method implemented in the present invention;

[0127] Figure 6 Schematic diagram of the intelligent soft switch candidate node selection strategy using the proposed method implemented in the present invention;

[0128] Figure 7 This is a flow chart of an active distribution network structural morphology optimization method considering flexible interconnection of intelligent soft switches according to an embodiment of the present invention. DETAILED DESCRIPTION

[0129] The technical solutions of the present invention are described clearly and completely below in conjunction with the accompanying drawings and specific embodiments of the present invention.

[0130] See also Figure 7 The present invention provides a method for optimizing the structure of an active distribution network considering flexible interconnection of intelligent soft switches, comprising the following steps:

[0131] Step 1) constructing a clustering framework for representative scenarios of new energy in the active distribution network system, reducing the new energy scenarios in the active distribution network system, and obtaining clustered representative scenarios;

[0132] Step 2) Based on the clustered representative scenarios, a location candidate set generation framework for intelligent soft switches in the active distribution network system is constructed;

[0133] Step 3) Develop the considerations for the intelligent soft switch sizing optimization framework in active distribution network systems;

[0134] Step 4) For the intelligent soft switch size optimization framework in the active distribution network system, the active loss function expression of the voltage source converter of the intelligent soft switch is constructed by integrating the considerations determined in step (3);

[0135] Step 5) Constructing the operating constraints of the smart soft switch based on the size optimization framework of the smart soft switch in the active distribution network system;

[0136] Step 6) For the uncertainty scenario of the upstream power grid, construct the active power exchange constraint conditions between the active distribution network system and the upstream power grid;

[0137] Step 7) Considering the location and size of the intelligent soft switch and various constraints, combined with the active power loss function relationship expression, an active distribution network system structure optimization model is constructed;

[0138] Step 8) The quantum binary particle swarm algorithm is combined with the differential evolution algorithm to solve the active distribution network system structural form optimization model described in step (7) to obtain the optimal active distribution network system structural form.

[0139] Specifically, in step 1), a clustering framework for representative scenarios of new energy in the active distribution network system is constructed, and the new energy scenarios in the active distribution network system are reduced to obtain representative scenarios after clustering, specifically:

[0140] Step 1.1) Randomly select samples {g1,g2,...,g k} as the initial cluster centers {η1,η2,...,η k}, initialize the clustering process.

[0141] Step 1.2) Calculate the new energy operation sample g i and the new energy operation cluster center η j(j＝1,2,...,k) The Euclidean distance between them. The specific formula is as follows:

[0142] d E,ij =||g i -η j || 2 (1)

[0143] Then, according to their respective Euclidean distance d E,ij , each new energy running sample g i Assigned to the nearest new energy operation cluster center η j(j＝1,2,...,k) .

[0144] Step 1.3) Calculate the mean of the new energy operation samples in each category and assign it as the new new energy operation cluster center.

[0145] Step 1.4) Repeat steps 1.2) and 1.3) until the new energy operation cluster center no longer changes, and then retain the final new energy operation cluster center.

[0146] Step 1.5) The final new energy operation cluster center is used as the new energy representative scenario.

[0147] Specifically, in step 2), based on the representative scenarios after clustering, a location candidate set generation framework for intelligent soft switches in the active distribution network system is constructed, specifically:

[0148] like Figure 1 As shown, Figure 1 A simple basic configuration of a distribution network with soft switches is given to help understand the configuration of soft switches in distribution networks.

[0149] In step 2.1, the data of the lines, loads, and distributed generation of the active distribution network system are used as input parameters for the power flow calculation. The Newton-Raphson method or the PQ decomposition method is used to calculate the line current of the active distribution network system, the active power and reactive power transmitted between nodes, and the node voltage of each node.

[0150] Step 2.2) Calculate the loss sensitivity index of the active distribution network system. Specifically:

[0151]

[0152] in, is the loss sensitivity index of node j. n Φ is the number of phases. Φ is the index of the active distribution network phase (Φ = A, B, C). n s is the total number of representative new energy scenarios. ij,Φ,s Q is the active power of branch ij, phase Φ, and scenario s. ij,Φ,s is the reactive power of branch ij, phase Φ, and scenario s. ij,Φ is the resistance of branch ij, phase Φ. V j,Φ,s is the voltage at node j, phase Φ, and scenario s. s is the probability of the sth representative scenario of new energy occurring.

[0153] Step 2.3) Calculate the normalized node voltage deviation index of the active distribution network system. Specifically:

[0154] Calculate the voltage deviation index of each node, specifically:

[0155]

[0156] Where, VDI j is the voltage deviation index of node j. Φ is the number of phases. Φ is the index of the active distribution network phase (Φ = A, B, C). n s is the total number of representative scenarios of new energy. n is the per-unit rated voltage. V j,Φ,s is the voltage at node j, phase Φ, and scenario s. s is the probability of the sth representative scenario of new energy occurring.

[0157] Use the following formula to calculate VDI j Normalization is performed to calculate the normalized node voltage deviation index of the active distribution network system, specifically:

[0158]

[0159] Where, VDI j,normis the normalized voltage deviation index of node j, VDI max All VDI j The maximum value in .

[0160] Step 2.4) Group the normalized node loss sensitivity index of the active distribution network system and the normalized node voltage deviation index of the active distribution network system (Ω Pos,LSI ,Ω Neg,LSI ) are combined and named and And sort them in descending order according to the weight of the two indices. Set a threshold and select nodes with weights higher than the threshold, named Count the number of nodes whose weight is higher than the threshold, and then take the same number of nodes from the column in, that is, The first node in the The nodes of both are components of the search space, where the nodes of both are connected by soft switches, excluding multiple soft switch connections of consecutive nodes on the same side. is the candidate set of locations for smart soft switches in active distribution network systems.

[0161] Specifically, in step 3), the factors to be considered in formulating the intelligent soft switch size optimization framework in the active distribution network system are:

[0162] Step 3.1) Construct the line power loss function of the active distribution network system, specifically:

[0163]

[0164] in, is the line power loss function of the active distribution network system. Φ is the index of the phase of the active distribution network system (Φ=A,B,C). s is the total number of representative scenarios of new energy. l is the set of network branches. ij,Φ I is the resistance of branch ij, phase Φ. ij,Φ,s The current of branch ij, phase Φ, and scenario s. s is the probability of the sth representative scenario of new energy occurring.

[0165] Step 3.2) Construct the intelligent soft switch active power loss function of the active distribution network system, specifically:

[0166]

[0167] in, is the active power loss function of the intelligent soft switch of the active distribution network system. Φ is the index of the phase of the active distribution network system (Φ=A,B,C). sis the total number of representative scenarios of new energy. SOP It is the collection of all intelligent soft switch connection nodes. is the active power loss of the voltage source converter with intelligent soft switching at node i, phase Φ, and scenario s. s is the probability of the sth representative scenario of new energy occurring.

[0168] Step 3.3) Construct the distributed generation active power output reduction function of the active distribution network system, specifically:

[0169]

[0170] in, is the distributed generation active power output curtailment function of the active distribution network system. Φ is the index of the phase of the active distribution network system (Φ=A,B,C). s is the total number of representative scenarios of new energy. DG is the set of all distributed generation nodes. s is the probability of the sth representative scenario of new energy occurring. is the active power reduction of distributed generation at node i, phase Φ, and scenario s, and its calculation formula is:

[0171]

[0172] in, is the active power reduction of distributed generation in node i, phase Φ, and scenario s. is the maximum available output active power of distributed generation at node i, phase Φ, and scenario s. is the dispatched output active power of distributed generation at node i, phase Φ, and scenario s.

[0173] Specifically, in step 4), for the intelligent soft switch size optimization framework in the active distribution network system, the consideration factors determined in step (3) are integrated to construct the active loss function relationship expression of the voltage source converter of the intelligent soft switch, which is specifically:

[0174] Step 4.1) Construct the active power loss of the voltage source converter of the intelligent soft switch at node i, specifically:

[0175]

[0176] in, is the active power loss of the voltage source converter with intelligent soft switching at node i, phase Φ, and scenario s. is the loss factor of the intelligent soft-switching voltage source converter. is the active power of the voltage source converter of the intelligent soft switch at node i, phase Φ, and scenario s. is the reactive power of the voltage source converter of the intelligent soft switch at node i, phase Φ, and scenario s.

[0177] Step 4.2) Construct the active power loss of the voltage source converter of the intelligent soft switch at node j, specifically:

[0178]

[0179] in, is the active power loss of the voltage source converter with intelligent soft switching at node j, phase Φ, and scenario s. is the loss factor of the intelligent soft-switching voltage source converter. is the active power of the voltage source converter of the intelligent soft switch at node j, phase Φ, and scenario s. is the reactive power of the voltage source converter of the intelligent soft switch at node j, phase Φ, and scenario s.

[0180] Step 4.3) Construct the active power balance equation of the voltage source converter of the intelligent soft switch connecting node i and node j, specifically:

[0181]

[0182] in, is the active power of the voltage source converter of the intelligent soft switch at node i, phase Φ, and scenario s. is the active power of the voltage source converter of the intelligent soft switch at node j, phase Φ, and scenario s. is the active power loss of the voltage source converter with intelligent soft switching at node i, phase Φ, and scenario s. is the active power loss of the voltage source converter with intelligent soft switching at node j, phase Φ, and scenario s.

[0183] Specifically, in step 5), the operating constraints of the smart soft switch are constructed for the size optimization framework of the smart soft switch in the active distribution network system, specifically:

[0184] Step 5.1) Construct the single-phase capacity constraint of the voltage source converter with intelligent soft switching, specifically:

[0185]

[0186] in, is the active power of the voltage source converter of the intelligent soft switch at node i, phase Φ, and scenario s. is the reactive power of the voltage source converter of the intelligent soft switch at node i, phase Φ, and scenario s. is the voltage source converter capacity of the intelligent soft switch for node i, phase Φ, and scenario s.

[0187] Step 5.2) Construct the single-phase capacity expression of the intelligent soft-switching voltage source converter, specifically:

[0188]

[0189] in, is the voltage source converter capacity of the intelligent soft switch for node i, phase Φ, and scenario s. is the three-phase capacity of the voltage source converter of the intelligent soft switch at node i and scenario s.

[0190] Step 5.3) Construct the three-phase capacity expression of the intelligent soft-switching voltage source converter, specifically:

[0191]

[0192] in, is the three-phase capacity of the voltage source converter of the intelligent soft switch at node i and scenario s. Φ is the index of the phase of the active distribution network system (Φ = A, B, C). n s It is the total number of representative scenarios of new energy. is the voltage source converter capacity of the intelligent soft switch for node i, phase Φ, and scenario s. s is the probability of the sth representative scenario of new energy occurring.

[0193] Step 5.4) Construct the operating constraints of the intelligent soft switch, specifically:

[0194]

[0195] in, is the three-phase capacity of the voltage source converter of the smart soft switch at node i, which is the final decision variable for the size of the smart soft switch. It is the lower limit of the three-phase capacity of the intelligent soft-switching voltage source converter. It is the upper limit of the three-phase capacity of the intelligent soft-switching voltage source converter.

[0196] Specifically, in step 6), for the uncertainty scenario of the upstream power grid, the active power exchange constraint conditions between the active distribution network system and the upstream power grid are constructed, specifically:

[0197] Step 6.1) Construct the active power exchange constraints between the active distribution network system and the upstream power grid, specifically:

[0198]

[0199] in, is the active exchange power between the active distribution network system of node i, phase Φ, and scenario s and the upstream power grid. is the lower limit of the active power exchange between the active distribution network system at node i and phase Φ and the upstream power grid. is the upper limit of the active power exchange between the active distribution network system at node i and phase Φ and the upstream power grid.

[0200] Step 6.2) Construct the reactive power exchange constraints between the active distribution network system and the upstream power grid, specifically:

[0201]

[0202] in, is the reactive exchange power between the active distribution network system of node i, phase Φ, and scenario s and the upstream power grid. is the lower limit of reactive power exchange between the active distribution network system at node i and phase Φ and the upstream power grid. is the upper limit of reactive power exchange between the active distribution network system at node i and phase Φ and the upstream power grid.

[0203] Step 6.3) Construct the power factor constraint of the upstream power grid, specifically:

[0204]

[0205] in, is the upstream grid power factor angle of phase Φ and scenario s. is the upstream grid power three-phase factor angle of scenario s. is the average three-phase factor angle of the upstream power grid. It is the lower limit of the upstream grid power factor angle limit. It is the upper limit of the upstream grid power factor angle limit.

[0206] Specifically, in step 7), the position and size of the intelligent soft switch and various constraints are taken into consideration, and combined with the active power loss function relationship expression, an active power distribution network system structure optimization model is constructed, specifically:

[0207] Step 7.1) Considering the location and size of the intelligent soft switch, an optimization model of the active distribution network structure is constructed. Specifically:

[0208] Considering the line power loss, intelligent soft switch active power loss and distributed generation active power output reduction in the active distribution network system, the objective function of the active distribution network system structure optimization model is constructed, specifically:

[0209]

[0210] Among them, OF is the size optimization objective function of the intelligent soft switch in the active distribution network system. is the line power loss function of the active distribution network system. It is the active power loss function of the intelligent soft switch in the active distribution network system. is the distributed generation active power output reduction function of the active distribution network system.

[0211] Step 7.2) Construct the constraints of the active distribution network system structure optimization model, specifically

[0212]

[0213] in, is the active power loss of the voltage source converter with intelligent soft switching at node i, phase Φ, and scenario s. is the loss factor of the intelligent soft-switching voltage source converter. is the active power of the voltage source converter of the intelligent soft switch at node i, phase Φ, and scenario s. is the reactive power of the voltage source converter of the intelligent soft switch at node i, phase Φ, and scenario s. is the active power loss of the voltage source converter with intelligent soft switching at node j, phase Φ, and scenario s. is the loss factor of the intelligent soft-switching voltage source converter. is the active power of the voltage source converter of the intelligent soft switch at node j, phase Φ, and scenario s. is the reactive power of the voltage source converter of the intelligent soft switch at node j, phase Φ, and scenario s. is the voltage source converter capacity of the intelligent soft switch for node i, phase Φ, and scenario s. is the three-phase capacity of the voltage source converter of the intelligent soft switch at node i and scenario s. Φ is the index of the phase of the active distribution network system (Φ = A, B, C). n s is the total number of representative scenarios of new energy. s is the probability of the sth representative scenario of new energy occurring. is the three-phase capacity of the voltage source converter of the smart soft switch at node i, which is the final decision variable for the size of the smart soft switch. It is the lower limit of the three-phase capacity of the intelligent soft-switching voltage source converter. It is the upper limit of the three-phase capacity of the intelligent soft-switching voltage source converter. is the active exchange power between the active distribution network system of node i, phase Φ, and scenario s and the upstream power grid. is the lower limit of the active power exchange between the active distribution network system at node i and phase Φ and the upstream power grid. is the upper limit of the active power exchange between the active distribution network system at node i and phase Φ and the upstream power grid. is the reactive exchange power between the active distribution network system of node i, phase Φ, and scenario s and the upstream power grid. is the lower limit of reactive power exchange between the active distribution network system at node i and phase Φ and the upstream power grid. is the upper limit of reactive power exchange between the active distribution network system at node i and phase Φ and the upstream power grid. is the upstream grid power factor angle of phase Φ and scenario s. is the upstream grid power three-phase factor angle of scenario s. is the average three-phase factor angle of the upstream power grid. It is the lower limit of the upstream grid power factor angle limit. It is the upper limit of the upstream grid power factor angle limit.

[0214] Specifically, in step 8), the quantum binary particle swarm algorithm is fused with the differential evolution algorithm to solve the active distribution network system structure optimization model in step (7) to obtain the optimal active distribution network system structure, which is specifically:

[0215] Step 8.1) Using the position of the smart soft switch and the active and reactive capacities of the voltage source converters at both ends as decision variables, the quantum binary particle swarm optimization (QB-PSO) algorithm is used to solve the position of the smart soft switch in the active distribution network system. Specifically,

[0216] Initialize the particle population, set the particle population size to N, the dimension of each particle to M, and randomly assign each dimension of the particle to 0-1 with a probability of 50%, and get the initial population X of QB-PSO (0) =(x1,x2,…,x N ) (0) .

[0217] For the t-th iteration result, the variable pbest is used to record the value status of all particles in the population, that is, pbest (t) =X (t) ,t=0,1,2,…,T。

[0218] For the t-th iteration result, the variable mbest is used to record the average value state of all particles in the population. (t) ,t=0,1,2,…,TFirst, the average value of each dimension of all particles in the population is calculated. If the average value of the dimension is less than 0.5, it is considered that the value of mbest in this dimension is 0. If the average value is greater than 0.5, it is considered that the value of mbest in this dimension is 1. If the average value is exactly 0.5, the value of mbest is determined by the random number method.

[0219] For the t-th iteration result, the variable gbest is used to record the value state of the specific particle in the population that makes the fitness function fitness (i.e., the opposite number of formula (19)) reach the maximum value, that is,

[0220] Define pbest i The reorganization variable between and gbest is P i The calculation method adopts the "gene recombination" operation in biological genetics. First, the qth dimension of the particle is randomly selected to respectively i and gbest are truncated into four parts, and then recombined into two column vectors. Finally, a random number method is used to determine one of the column vectors as P i The value of .

[0221] The binary quantum particle swarm updates the decision variable x according to the following principle i .

[0222] ⑦Calculate x i With P i The difference b is calculated as

[0223] b=β·d H (x i ,mbest)·ln(1 / μ),μ=rand(0,1) (21)

[0224] In the formula, β is a constant, d H is the Hamming distance between two vectors, and μ is a random number between 0 and 1.

[0225] ⑧Calculate auxiliary variable p r , which is calculated as

[0226]

[0227] Where, l d is the dimension of the particle.

[0228] ⑨Update decision variable x i , first reorganize the variables into P i Assign to x i , then for each dimension of the decision variable, random numbers and p are used r For comparison, if p r is greater than the random number, then the decision variable x i The values ​​of this dimension are swapped between 0 and 1, otherwise they remain unchanged, and the value of the decision variable x is returned. i (t+1) .

[0229] After the preset number of iterations T=400, the state gbest of the particle with the best fitness function in the last generation is taken (T) The intelligent soft switch position optimization result of calling QB-PSO once.

[0230] Step 8.2) Using the intelligent soft switch position and the active and reactive capacities of the two-terminal voltage source converter as decision variables, the differential evolution algorithm (DE) is used to solve the active and reactive capacities of the two-terminal voltage source converter. Specifically,

[0231] DE mutation, crossover and selection operations are used to optimize the active and reactive capacity of the two-terminal voltage source converter. For the mutation operation of the i-th DE individual (i=1,2,...,P S ), first randomly select three different individuals from the population and (r1≠r2≠r3≠i), As basis vectors, and After making a difference and scaling Superposition, that is:

[0232]

[0233] Where, is the mutation vector; F is the scaling factor.

[0234] The crossover operation uses the binary crossover method, which acts on The population of the previous iteration Finally, the trial vector is generated Right now:

[0235]

[0236] Where, and Represents vectors and The j-th dimension component, C R is the crossover probability.

[0237] A selection operation is used to select between the offspring and the parent. If the target value of the offspring vector is less than the target value of the parent vector, the offspring vector is retained and the parent vector is eliminated; otherwise, the parent vector is retained and the offspring vector is eliminated.

[0238]

[0239] Through DE mutation, crossover and selection operations, and after repeated iterations, the active and reactive capacities of the two-terminal voltage source converter are optimized and solved, and the optimal solution of the active and reactive capacities of the two-terminal voltage source converter is obtained.

[0240] Specifically, in this embodiment, the modified IEEE 33-node test system is used for analysis, as shown in FIG. Figure 2As shown in Figure 2, the system's active power and reactive power demands are 3715 kW and 2300 kW, respectively. The system has two photovoltaic systems at nodes 22 and 28, each with a rated capacity of 600 kVA; and four wind turbines at nodes 11, 14, 16, and 32, each with a rated capacity of 500 kVA. The loss coefficient of the intelligent soft switch is set to 0.02. Scenario generation and reduction techniques are used to create the scenarios. Table 1 shows the per-unit values ​​of load, wind power, and photovoltaic power, as well as their probabilities, for different scenarios.

[0241] Table 1 Per-unit values ​​and probabilities of load, wind power, and photovoltaic power in different scenarios

[0242]

[0243]

[0244] At this time, the nodes on the adjacent side are divided into a group, forming Figure 5 The nodes on the same side (abbreviated as L1, L2, L3) are shown. These nodes on adjacent sides can be connected to adjacent sides of the feeder or between different feeders through soft switching control. Then, according to the method in step 2), the loss sensitivity index of the active distribution network is calculated, as follows: Figure 3 Then the voltage deviation index of the active distribution network is calculated, as shown in Figure 4 Then, according to step 2.4), we can get Contains nodes 24, 25, and 30; Including nodes 11, 14, and 16, the overall result is as follows Figure 5 When the number of soft switches is one or two, the soft switches are connected to the active distribution network. All the permutations and combinations of soft switches are shown in Table 2. Considering the losses and soft switch capacities shown in Table 2, it is not difficult to deduce that Figure 6 The proposed method is compared with a method that only considers intelligent soft switch position optimization (referred to as the comparison method). Table 3 shows the soft switch positions, capacities, and corresponding power losses and voltage ranges for the two methods. The proposed method achieves a significantly higher soft switch capacity of 945 kVA than the comparison method, with power losses 8.31% lower than the comparison method. Furthermore, the proposed method maintains a more stable voltage range, demonstrating the strong superiority of the present invention and its ability to better optimize the structure of active distribution networks.

[0245] Table 2 Different positions, capacities and corresponding power losses of soft switches

[0246]

[0247]

[0248] Table 3 Different positions, capacities, and corresponding power losses and voltage ranges of soft switches under two methods

[0249]

[0250] It will be easily understood by those skilled in the art that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for optimizing the structure of an active distribution network considering flexible interconnection of intelligent soft switches, characterized in that: The following steps are involved: 1) Construct a clustering framework for representative scenarios of new energy in the active distribution network system, reduce the new energy scenarios in the active distribution network system, and obtain representative scenarios after clustering; 2) Based on the clustered representative scenarios, a location candidate set generation framework for intelligent soft switches in active distribution network systems is constructed; 3) Determine the considerations for the intelligent soft switch sizing optimization framework in active distribution network systems; 4) Based on the intelligent soft switch size optimization framework in the active distribution network system, the active power loss function expression of the voltage source converter of the intelligent soft switch is constructed by integrating the considerations determined in step 3); 5) Develop an intelligent soft switch size optimization framework for active distribution network systems and establish operating constraints for intelligent soft switches; 6) Aiming at the uncertainty scenario of the upstream power grid, construct the active power exchange constraints between the active distribution network system and the upstream power grid; 7) Considering the location and size of the intelligent soft switch and various constraints, combined with the active power loss function expression, an active distribution network system structural optimization model is constructed; 8) Using the quantum binary particle swarm algorithm fused with the differential evolution algorithm to solve the active distribution network system structure optimization model constructed in step 7) to obtain the optimal active distribution network system structure.

2. The method for optimizing the structure of an active distribution network considering flexible interconnection of intelligent soft switches according to claim 1, characterized in that: In step 1), a clustering framework of representative scenarios of new energy in the active distribution network system is constructed, and the new energy scenarios in the active distribution network system are reduced to obtain representative scenarios after clustering, specifically: Step 1.1) Randomly select samples {g1,g2,...,g k } as the initial cluster centers {η1,η2,...,η k }, initialize the clustering process; Step 1.2) Calculate the new energy operation sample g i and the new energy operation cluster center η j(j＝1,2,...,k) The Euclidean distance between them is as follows: d E,ij =||g i -or j || 2 (1) Then, according to their respective Euclidean distance d E,ij , each new energy running sample g i Assigned to the nearest new energy operation cluster center η j(j＝1,2,...,k) ; Step 1.3) Calculate the mean of the new energy operation samples in each category and assign it as the new new energy operation cluster center; Step 1.4) Repeat steps 1.2) and 1.3) until the new energy operation cluster center no longer changes, and then retain the final new energy operation cluster center; Step 1.5) The final new energy operation cluster center is used as the representative scenario after clustering.

3. The method for optimizing the structure of an active distribution network considering flexible interconnection of intelligent soft switches according to claim 1, characterized in that: In step 2), based on the representative scenarios after clustering, a location candidate set generation framework for intelligent soft switches in the active distribution network system is constructed, specifically: Step 2.1) Using the data of the lines, loads, and distributed generation of the active distribution network system as input parameters for the power flow calculation, the Newton-Raphson method or the PQ decomposition method is used to calculate the line current of the active distribution network system, the active power and reactive power transmitted between nodes, and the node voltage of each node; Step 2.2) Calculate the loss sensitivity index of the active distribution network system, specifically: in, is the loss sensitivity index of node j, n Φ is the number of phases, Φ is the index of the phase of the active distribution network system, Φ = A, B, C; n s is the total number of representative scenarios of new energy, P ij,Φ,s is the active power of branch ij, phase Φ, and scenario s, Q ij,Φ,s is the reactive power of branch ij, phase Φ, and scenario s, r ij,Φ is the resistance of branch ij, phase Φ, V j,Φ,s is the voltage at node j, phase Φ, and scenario s, β s is the probability of occurrence of the sth representative scenario of new energy; Step 2.3) Calculate the normalized node voltage deviation index of the active distribution network system, specifically: Calculate the voltage deviation index of each node: Where, VDI j is the voltage deviation index of node j, n Φ is the number of phases, Φ is the index of the phase of the active distribution network system, Φ = A, B, C; n s is the total number of representative new energy scenarios, V n is the per-unit rated voltage, V j,Φ,s is the voltage at node j, phase Φ, and scenario s, β s is the probability of occurrence of the sth representative scenario of new energy; Use the following formula to calculate VDI j Normalization is performed to calculate the normalized node voltage deviation index of the active distribution network system: Where, VDI j,norm is the normalized voltage deviation index of node j, VDI max All VDI j The maximum value in ; Step 2.4) Group the normalized node loss sensitivity index of the active distribution network system and the normalized node voltage deviation index of the active distribution network system (Ω Pos,LSI ,Ω Neg,LSI ) are combined and named and And sort them in descending order according to the weight of the two indices, set a threshold, select the nodes with weight higher than the threshold, and name them Count the number of nodes whose weight is higher than the threshold, and then take the same number of nodes from the column in, that is, The first node in the The nodes of both nodes are part of the search space, where the nodes of both nodes are connected by soft switches, excluding multiple soft switch connections of consecutive nodes on the same side. is the candidate set of locations for smart soft switches in active distribution network systems.

4. The method for optimizing the structure of an active distribution network considering flexible interconnection of intelligent soft switches according to claim 1, characterized in that: In step 3), the factors to be considered in formulating the intelligent soft switch size optimization framework in the active distribution network system are as follows: Step 3.1) Construct the line power loss function of the active distribution network system: in, is the line power loss function of the active distribution network system, Φ is the index of the phase of the active distribution network system, Φ = A, B, C, n s is the total number of representative scenarios of new energy, Ω l is the set of network branches, r ij,Φ is the resistance of branch ij, phase Φ, I ij,Φ,s The current of branch ij, phase Φ, and scene s, β s is the probability of occurrence of the sth representative scenario of new energy; Step 3.2) Construct the intelligent soft-switching active power loss function of the active distribution network system: in, is the active power loss function of the intelligent soft switch in the active distribution network system, Ω SOP It is the collection of all intelligent soft switch connection nodes. is the active power loss of the voltage source converter with intelligent soft switching at node i, phase Φ, and scenario s; Step 3.3) Construct the distributed generation active power output reduction function of the active distribution network system, specifically: in, is the distributed generation active power output curtailment function of the active distribution network system, Φ is the phase index of the active distribution network system, Ω DG is the collection of all distributed generation nodes, is the active power reduction of distributed generation at node i, phase Φ, and scenario s, and its calculation formula is: in, is the active power reduction of distributed generation in node i, phase Φ, and scenario s, is the maximum available output active power of distributed generation at node i, phase Φ, and scenario s, is the dispatched output active power of distributed generation at node i, phase Φ, and scenario s.

5. The method for optimizing the structure of an active distribution network considering flexible interconnection of intelligent soft switches according to claim 4 is characterized in that: In step 4), for the intelligent soft switch size optimization framework in the active distribution network system, the consideration factors determined in step 3) are integrated to construct the active loss function relationship expression of the voltage source converter of the intelligent soft switch, which is specifically: Step 4.1) Construct the active power loss of the voltage source converter of the intelligent soft switch at node i, specifically: in, is the active power loss of the voltage source converter with intelligent soft switching at node i, phase Φ, and scenario s, is the loss coefficient of the intelligent soft-switching voltage source converter, is the active power of the voltage source converter of the intelligent soft switch at node i, phase Φ, and scenario s, is the reactive power of the voltage source converter of the intelligent soft switch at node i, phase Φ, and scenario s; Step 4.2) Construct the active power loss of the voltage source converter of the intelligent soft switch at node j, specifically: in, is the active power loss of the voltage source converter with intelligent soft switching at node j, phase Φ, and scenario s, is the loss coefficient of the intelligent soft-switching voltage source converter, is the active power of the voltage source converter of the intelligent soft switch at node j, phase Φ, and scenario s, is the reactive power of the voltage source converter of the intelligent soft switch at node j, phase Φ, and scenario s; Step 4.3) Construct the active power balance equation of the voltage source converter of the intelligent soft switch connecting node i and node j, specifically: in, is the active power of the voltage source converter of the intelligent soft switch at node i, phase Φ, and scenario s, is the active power of the voltage source converter of the intelligent soft switch at node j, phase Φ, and scenario s, is the active power loss of the voltage source converter with intelligent soft switching at node i, phase Φ, and scenario s, is the active power loss of the voltage source converter with intelligent soft switching at node j, phase Φ, and scenario s.

6. The method for optimizing the structure of an active distribution network considering flexible interconnection of intelligent soft switches according to claim 5, characterized in that: In step 5), the operating constraints of the smart soft switch are constructed for the size optimization framework of the smart soft switch in the active distribution network system, specifically: Step 5.1) Construct the single-phase capacity constraint of the voltage source converter with intelligent soft switching, specifically: in, is the active power of the voltage source converter of the intelligent soft switch at node i, phase Φ, and scenario s, is the reactive power of the voltage source converter of the intelligent soft switch at node i, phase Φ, and scenario s, is the voltage source converter capacity of the intelligent soft switch for node i, phase Φ, and scenario s; Step 5.2) Construct the single-phase capacity expression of the intelligent soft-switching voltage source converter, specifically: in, is the voltage source converter capacity of the intelligent soft switch for node i, phase Φ, and scenario s, is the three-phase capacity of the voltage source converter of the intelligent soft switch at node i and scenario s; Step 5.3) Construct the three-phase capacity expression of the intelligent soft-switching voltage source converter, specifically: in, is the three-phase capacity of the voltage source converter of the intelligent soft switch of node i and scenario s, Φ is the index of the phase of the active distribution network system, Φ = A, B, C, n s is the total number of representative scenarios of new energy, is the voltage source converter capacity of the smart soft switch for node i, phase Φ, and scenario s, β s is the probability of occurrence of the sth representative scenario of new energy; Step 5.4) Construct the operating constraints of the intelligent soft switch, specifically: in, is the three-phase capacity of the voltage source converter of the intelligent soft switch at node i, which is the final decision variable for the size of the intelligent soft switch. is the lower limit of the three-phase capacity of the intelligent soft-switching voltage source converter, It is the upper limit of the three-phase capacity of the intelligent soft-switching voltage source converter.

7. The method for optimizing the structure of an active distribution network considering flexible interconnection of intelligent soft switches according to claim 6, characterized in that: In step 6), for the uncertainty scenario of the upstream power grid, the active power exchange constraint conditions between the active distribution network system and the upstream power grid are constructed, specifically: Step 6.1) Construct the active power exchange constraints between the active distribution network system and the upstream power grid, specifically: in, is the active exchange power between the active distribution network system of node i, phase Φ, scenario s and the upstream power grid, is the lower limit of active power exchange between the active distribution network system at node i and phase Φ and the upstream power grid, is the upper limit of active power exchange between the active distribution network system at node i and phase Φ and the upstream power grid; Step 6.2) Construct the reactive power exchange constraints between the active distribution network system and the upstream power grid, specifically: in, is the reactive exchange power between the active distribution network system of node i, phase Φ, scenario s and the upstream power grid, is the lower limit of reactive power exchange between the active distribution network system at node i and phase Φ and the upstream power grid, is the upper limit of reactive power exchange between the active distribution network system at node i and phase Φ and the upstream power grid; Step 6.3) Construct the power factor constraint of the upstream power grid, specifically: in, is the upstream grid power factor angle for phase Φ and scenario s, is the upstream grid power three-phase factor angle of scenario s, is the average three-phase factor angle of the upstream power grid, is the lower limit of the upstream grid power factor angle limit, It is the upper limit of the upstream grid power factor angle limit.

8. The method for optimizing the structure of an active distribution network considering flexible interconnection of intelligent soft switches according to claim 7, characterized in that: In step 7), the position and size of the intelligent soft switch and various constraints are taken into consideration, and combined with the active power loss function relationship expression, an active distribution network system structural morphology optimization model is constructed, specifically: Step 7.1) Considering the location and size of the intelligent soft switch, an optimization model of the active distribution network system structure is constructed. Specifically: Considering the line power loss, active power loss of intelligent soft switches, and active power output reduction of distributed generation in the active distribution network system, the objective function of the active distribution network system structure optimization model is constructed: Among them, OF is the size optimization objective function of the intelligent soft switch in the active distribution network system, is the line power loss function of the active distribution network system, is the active power loss function of the intelligent soft switch in the active distribution network system, is the distributed generation active power output curtailment function of the active distribution network system; Step 7.2) Construct the constraints of the active distribution network system structure optimization model, specifically: in, is the active power loss of the voltage source converter with intelligent soft switching at node i, phase Φ, and scenario s, is the loss coefficient of the intelligent soft-switching voltage source converter, is the active power of the voltage source converter of the intelligent soft switch at node i, phase Φ, and scenario s, is the reactive power of the voltage source converter of the intelligent soft switch at node i, phase Φ, and scenario s, is the active power loss of the voltage source converter with intelligent soft switching at node j, phase Φ, and scenario s, is the loss coefficient of the intelligent soft-switching voltage source converter, is the active power of the voltage source converter of the intelligent soft switch at node j, phase Φ, and scenario s, is the reactive power of the voltage source converter of the intelligent soft switch at node j, phase Φ, and scenario s, is the voltage source converter capacity of the intelligent soft switch for node i, phase Φ, and scenario s, is the three-phase capacity of the voltage source converter of the intelligent soft switch of node i and scenario s, Φ is the index of the phase of the active distribution network system, Φ = A, B, C; n s is the total number of representative scenarios of new energy, β s is the probability of occurrence of the sth representative scenario of new energy, is the three-phase capacity of the voltage source converter of the intelligent soft switch at node i, which is the final decision variable for the size of the intelligent soft switch. is the lower limit of the three-phase capacity of the intelligent soft-switching voltage source converter, is the upper limit of the three-phase capacity of the intelligent soft-switching voltage source converter, is the active exchange power between the active distribution network system of node i, phase Φ, scenario s and the upstream power grid, is the lower limit of active power exchange between the active distribution network system at node i and phase Φ and the upstream power grid, is the upper limit of active power exchange between the active distribution network system at node i and phase Φ and the upstream power grid, is the reactive exchange power between the active distribution network system of node i, phase Φ, scenario s and the upstream power grid, is the lower limit of reactive power exchange between the active distribution network system at node i and phase Φ and the upstream power grid, is the upper limit of reactive power exchange between the active distribution network system at node i and phase Φ and the upstream power grid, is the upstream grid power factor angle for phase Φ and scenario s, is the upstream grid power three-phase factor angle of scenario s, is the average three-phase factor angle of the upstream power grid, is the lower limit of the upstream grid power factor angle limit, It is the upper limit of the upstream grid power factor angle limit.

9. The method for optimizing the structure of an active distribution network considering flexible interconnection of intelligent soft switches according to claim 8, characterized in that: In step 8), the quantum binary particle swarm algorithm is combined with the differential evolution algorithm to solve the active distribution network system structure optimization model in step (7) to obtain the optimal active distribution network system structure, specifically: Step 8.1) Using the position of the smart soft switch and the active and reactive capacities of the voltage source converters at both ends as decision variables, the quantum binary particle swarm optimization (QB-PSO) algorithm is used to solve the position of the smart soft switch in the active distribution network system. Specifically, Initialize the particle population, set the particle population size to N, the dimension of each particle to M, and randomly assign each dimension of the particle to 0-1 with a probability of 50%, and get the initial population X of QB-PSO (0) =(x1,x2,…,x N ) (0) ; For the t-th iteration result, the variable pbest is used to record the value status of all particles in the population, that is, pbest (t) =X (t) ,t=0,1,2,…,T; For the t-th iteration result, the variable mbest is used to record the average value state of all particles in the population. (t) ,t=0,1,2,…,T First, the average of each dimension value of all particles in the population is calculated. If the average value of the dimension is less than 0.5, it is considered that the value of mbest in this dimension is 0. If the average value is greater than 0.5, it is considered that the value of mbest in this dimension is 1. If the average value is exactly 0.5, the value of mbest is determined by random number method. For the t-th iteration result, the variable gbest is used to record the value state of the specific particle in the population that makes the fitness function fitness (the opposite number of formula (19)) reach the maximum value, that is, Define pbest i The reorganization variable between and gbest is P i The calculation method adopts the "gene recombination" operation in biological genetics. First, the qth dimension of the particle is randomly selected to respectively i and gbest are truncated into four parts, and then recombined into two column vectors. Finally, a random number method is used to determine one of the column vectors as P i The value of The binary quantum particle swarm updates the decision variable x according to the following principle i ; ①Calculate x i With P i The difference b is calculated as b=β·d H (x i ,mbest)·ln(1 / μ),μ=rand(0,1) (21) In the formula, β is a constant, d H is the Hamming distance between two vectors, μ is a random number between 0 and 1; ②Calculate the auxiliary variable p r , which is calculated as Where, l d is the dimension of the particle; ③Update the decision variable x i , first reorganize the variables into P i Assign to x i , then for each dimension of the decision variable, random numbers and p are used r For comparison, if p r is greater than the random number, then the decision variable x i The values ​​of this dimension are swapped between 0 and 1, otherwise they remain unchanged, and the value of the decision variable x is returned. i (t+1) ; After the preset number of iterations T=400, the state gbest of the particle with the best fitness function in the last generation is taken (T) As the intelligent soft switch position optimization result of calling QB-PSO once; Step 8.2) Using the intelligent soft switch position and the active and reactive capacities of the two-terminal voltage source converter as decision variables, the differential evolution algorithm (DE) is used to solve the active and reactive capacities of the two-terminal voltage source converter, specifically: DE mutation, crossover and selection operations are used to optimize the active and reactive capacity of the two-terminal voltage source converter. For the mutation operation of the i-th DE individual, i=1,2,...,P S , first randomly select three different individuals from the population and Where r1≠r2≠r3≠i, As basis vectors, and After making a difference and scaling Superposition, that is: Where, is the mutation vector; F is the scaling factor; The crossover operation uses the binary crossover method, which acts on The population of the previous iteration Finally, the trial vector is generated Right now: Where, and Represents vectors and The j-th dimension component, C R is the crossover probability; Use the selection operation to select between the offspring and the parent. If the target value of the offspring vector is less than the target value of the parent vector, the offspring vector is retained and the parent vector is eliminated; otherwise, the parent vector is retained and the offspring vector is eliminated: Through DE mutation, crossover and selection operations, and after repeated iterations, the active and reactive capacities of the two-terminal voltage source converter are optimized and solved, and the optimal solution of the active and reactive capacities of the two-terminal voltage source converter is obtained.

Images