Cooperative fault recovery system for main power grid and power distribution network

Abstract

The invention relates to a cooperative fault recovery system for a main power grid and a power distribution network, and the system comprises a fault recovery model building module which is used for building a fault recovery model based on the maximum random operation number of controllable distributed power supplies and the minimum power loss and following a system operation safety constraint condition under multi-source load access; the model solving module is used for solving and calculating to obtain an optimal fault recovery scheme based on the fault recovery model; and the power generation calculation module is used for sending a fault recovery control instruction to perform fault recovery based on the optimal fault recovery scheme. The invention relates to the technical field of power grids. The method starts from four key targets, follows system operation safety constraint conditions under multi-source load access, establishes a fault recovery model, and uses an improved binary particle swarm optimization and genetic algorithm combined method to find an optimal fault recovery scheme, so as to fully cooperate with various controllable DGs, improve power supply reliability, and improve power supply reliability. And power distribution network fault recovery requirements are met.

Description

Technical Field

[0001] The present invention relates to the technical field of power grids, and in particular to a coordinated fault recovery system for a main power grid and a distribution network. Background Art

[0002] In recent years, with the large-scale access of various distributed power sources such as photovoltaics and wind power, the traditional distribution network is undergoing an evolution, taking a key step towards the active distribution network (ADN). This evolution is accompanied by the continuous intelligence of ADN, which brings stronger load control capabilities to the power grid. It is particularly worth noting that the fault recovery speed plays a vital role in reducing economic losses and improving power supply quality. Therefore, it is necessary to comprehensively consider the characteristics of distributed generation (DG) to improve power supply reliability and reduce economic losses.

[0003] In the traditional power grid fault recovery strategy, different controllable DGs are generally used to supply power to the active distribution network in stages and at different times, or different controllable DGs are isolated and microgrids are independently operated in different regions to supply power to the active distribution network. However, the above methods all fully consider the potential of controllable DGs to participate in scheduling, and therefore do not fully coordinate the interaction of various controllable DGs to meet the distribution network fault recovery needs. Therefore, the present invention proposes a coordinated fault recovery system for the main power grid and the distribution network. Summary of the invention

[0004] In view of the deficiencies in the prior art, an object of the present invention is to provide a coordinated fault recovery system for a main power grid and a distribution network to solve the above technical problems.

[0005] The above technical objectives of the present invention are achieved through the following technical solutions:

[0006] A coordinated fault recovery system for a main power grid and a distribution network includes a fault recovery model building module for building a fault recovery model based on maximizing the number of random operations of controllable distributed power sources and minimizing power losses and following system operation safety constraints under multi-source load access;

[0007] A model solving module, used to solve and calculate the best fault recovery solution based on the fault recovery model;

[0008] The power generation calculation module is used to send fault recovery control instructions and perform fault recovery based on the best fault recovery plan.

[0009] Through the above technical scheme, the present invention starts from the four key goals of minimizing grid loss F1, the number of switch operations F2, the operation of controllable DG F3 and the average user downtime F4, follows the system operation safety constraints under multi-source load access, establishes a fault recovery model, and adopts a combination method of improved binary particle swarm algorithm and genetic algorithm to find the best fault recovery scheme, so as to fully coordinate various controllable DGs, improve power supply reliability, and meet the distribution network fault recovery requirements.

[0010] In a preferred example, the present invention can be further configured as follows: the fault recovery model includes establishing an objective function and fault recovery constraints, and performing normalization processing based on the objective function and the constraints to further optimize the fault recovery model;

[0011] The objective function includes minimization of grid loss F1, number of switch operations F2, controllable DG operation F3 and average user downtime F4. The fault recovery model is as follows:

[0012] F R =min(α1F1+α2F2+α4F4)+maxα3F3

[0013] Among them, α1, α2, α3, and α4 are weight coefficients, and their sum is 1;

[0014] The fault recovery constraints include node voltage constraints, branch capacity constraints, ring network constraints and power flow constraints.

[0015] In a preferred example, the present invention can be further configured as follows: the normalization process is performed based on the objective function and the constraint conditions to further optimize the fault recovery model, including minimum network loss and maximum power failure recovery;

[0016] The calculation method of the minimum network loss is:

[0017]

[0018] In the above formula, E is the total network loss after the power grid fault is restored, E s The total network loss before the active distribution network is restored;

[0019] The calculation method of the maximum power failure recovery is:

[0020]

[0021] In the above formula, P is the restored power supply load power, P s is the total power of the power-off load.

[0022] In a preferred example, the present invention can be further configured as follows: the objective function includes minimization of grid loss, number of switch operations, controllable DG operation and average user downtime, wherein the minimization of grid loss is expressed as:

[0023]

[0024] In the above formula, N is the total number of branches in the distribution network;

[0025] S i is the position of the switch on branch i, which is 0 when not connected and 1 when actively connected;

[0026] P i and Q i are the active power and reactive power losses on branch i respectively;

[0027] U i is the voltage of its i-node, R i is the resistance of the i-th branch;

[0028] The number of switch operations is expressed as:

[0029]

[0030] In the above formula, S s Indicates the initial closed state of the Sth branch switch;

[0031] T s and O s They represent the pre-fault section switches and the pre-fault tie switch set respectively;

[0032] The controllable DG operation is expressed as:

[0033]

[0034] In the above formula, N is the total number of DGs put into operation;

[0035] m represents the mth unit in operation;

[0036] Represents the output power of unit m;

[0037] The average user downtime is expressed as:

[0038]

[0039] In a preferred example, the present invention can be further configured as follows: the fault recovery constraint condition includes a node voltage constraint, a branch capacity constraint, a ring network constraint and a power flow constraint, wherein the node voltage constraint is expressed as:

[0040] Ui,min ≤U i ≤U i,max

[0041] In the above formula, U i,min and U i,max are the minimum and maximum values ​​of the voltage at node i, respectively;

[0042] The branch capacity constraint is expressed as:

[0043] |P i |≤P i,max

[0044] In the above formula, P i is the active power flowing through branch i, P i,max is the maximum capacity of the branch;

[0045] The ring network restriction is expressed as:

[0046]

[0047] In the above formula, T i is the closed state of the i-th switch, It is a collection of switches in a closed circuit, and there is no ring network structure in the distribution network;

[0048] The power flow constraint is expressed as:

[0049]

[0050] In the above formula, P i and Q i is the output power of the distribution network to node i;

[0051] P DGi and Q DGi is the injected power of node i to the distributed generation;

[0052] P Di and Q Di is the load power of node i;

[0053] U i and U j is the voltage of nodes i and j;

[0054] G ij , B ij is the branch conductance and susceptance;

[0055] α ij is the voltage phase angle difference between nodes i and j.

[0056] In a preferred example, the present invention can be further configured as follows: based on the fault recovery model, solving and calculating to obtain the best fault recovery solution, integrating the improved binary particle swarm algorithm and the genetic algorithm, the specific method is:

[0057] S1, initialize the number of particles N, particle dimension D, maximum number of iterations S, crossover probability P c , mutation probability P m ;

[0058] S2, calculate the fitness of each particle, search the feasible region of the island, and take the current particle position as the historical optimal solution;

[0059] S3, update the velocity and position of the particle;

[0060] S4, perform a crossover operation on two random particles in the population to generate new particles P and Q, calculate their fitness, and select the new particle with the largest fitness;

[0061] S5, perform mutation operation on the new particles, select a certain number of particles using the mutation probability, calculate the fitness value, and select the new particles with larger values;

[0062] S6, perform branch capacity, node voltage, distribution network structure, and power flow calculations, solve the load recovery amount, save the better particles, and complete an iteration;

[0063] S7, determining whether the maximum number of iterations has been reached, if not, repeating S3 to S6 until the maximum number of iterations has been reached;

[0064] S8. Output the optimal solution for fault recovery and system network loss, and the calculation ends.

[0065] In a preferred example, the present invention can be further configured as follows: In the improved binary particle swarm algorithm, the position and velocity of each particle represent the probability of 0 and 1, and are updated through a specific iteration rule. Assuming that the current particle swarm has a b-dimensional space, the position of particle i is x i =(K i1 ,K i2 ,…,K ib ), whose velocity is represented by v i =(v i1 ,v i2 ,…,v ib ), the speed and position update rules of each particle iteration are as follows:

[0066]

[0067] In the above formula, ω is the inertia factor, which is a positive integer;

[0068] c1 and c2 are learning factors;

[0069] r1 and r2 are random numbers between [0,1];

[0070] rand() is a random number on [0, 1];

[0071] S(v ib ) is the sigmoid function.

[0072] In a preferred example, the present invention can be further configured as follows: the genetic algorithm uses the variance σ 2 Adjust the fitness value of the population, use crossover mutation to form a new population, enhance the global search capability, and thus solve the local optimal problem;

[0073]

[0074] f s =max{1,max|f i -f avg |

[0075] In the above formula, f i is the solution of iteration i;

[0076] f avg is the average iteration value;

[0077] f s is the normalization factor of fitness.

[0078] In a preferred example, the present invention can be further configured as follows: α1, α2, α3, and α4 are 0.2, 0.3, 0.2, and 0.3 respectively;

[0079] F R =min(0.2F1+0.3F2+0.3F4)+max0.2F3.

[0080] In a preferred example, the present invention can be further configured as follows: the maximum number of iterations S is 100 times.

[0081] In summary, the present invention includes at least one of the following beneficial technical effects:

[0082] Starting from the four key goals of minimizing grid loss F1, the number of switch operations F2, the operation of controllable DG F3 and the average user downtime F4, the present invention follows the system operation safety constraints under multi-source load access, establishes a fault recovery model, and adopts a combination method of improved binary particle swarm algorithm and genetic algorithm to find the best fault recovery plan, so as to fully coordinate various controllable DGs, improve power supply reliability, and meet the distribution network fault recovery requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0083] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0084] Figure 1 It is a system diagram of a coordinated fault recovery system of a main power grid and a distribution network of the present invention. DETAILED DESCRIPTION

[0085] The present invention is further described in detail below in conjunction with the accompanying drawings.

[0086] Embodiment 1:

[0087] Reference Figure 1 , a coordinated fault recovery system of a main power grid and a distribution network disclosed in the present invention includes a fault recovery model building module for building a fault recovery model based on maximizing the number of random operations of controllable distributed power sources and minimizing power losses, and following system operation safety constraints under multi-source load access;

[0088] A model solving module, used to solve and calculate the best fault recovery solution based on the fault recovery model;

[0089] The power generation calculation module is used to send fault recovery control instructions and perform fault recovery based on the best fault recovery plan.

[0090] The fault recovery model includes establishing an objective function and fault recovery constraints, and performing normalization processing based on the objective function and the constraints to further optimize the fault recovery model;

[0091] The objective function includes minimization of grid loss F1, number of switch operations F2, controllable DG operation F3 and average user downtime F4. The fault recovery model is as follows:

[0092] F R =min(α1F1+α2F2+α4F4)+maxα3F3

[0093] Among them, α1, α2, α3, and α4 are weight coefficients, and their sum is 1;

[0094] The objective function includes minimization of grid loss, number of switch operations, controllable DG operation and average user downtime, wherein the grid loss minimization is expressed as:

[0095]

[0096] In the above formula, N is the total number of branches in the distribution network;

[0097] S i is the position of the switch on branch i, which is 0 when not connected and 1 when actively connected;

[0098] P i and Q i are the active power and reactive power losses on branch i respectively;

[0099] U i is the voltage of its i-node, R i is the resistance of the i-th branch;

[0100] The number of switch operations is expressed as:

[0101]

[0102] In the above formula, S s Indicates the initial closed state of the Sth branch switch;

[0103] T s and O s They represent the pre-fault section switches and the pre-fault tie switch set respectively;

[0104] The controllable DG operation is expressed as:

[0105]

[0106] In the above formula, N is the total number of DGs put into operation;

[0107] m represents the mth unit in operation;

[0108] Represents the output power of unit m;

[0109] The average user downtime is expressed as:

[0110]

[0111] The fault recovery constraints include node voltage constraints, branch capacity constraints, ring network constraints and power flow constraints.

[0112] The fault recovery constraint conditions include node voltage constraint, branch capacity constraint, ring network constraint and power flow constraint, wherein the node voltage constraint is expressed as:

[0113] U i,min ≤U i ≤U i,max

[0114] In the above formula, U i,min and U i,max are the minimum and maximum values ​​of the voltage at node i, respectively;

[0115] The branch capacity constraint is expressed as:

[0116] |P i |≤P i,max

[0117] In the above formula, P i is the active power flowing through branch i, P i,max is the maximum capacity of the branch;

[0118] The ring network restriction is expressed as:

[0119]

[0120] In the above formula, T i is the closed state of the i-th switch, It is a collection of switches in a closed circuit, and there is no ring network structure in the distribution network;

[0121] The power flow constraint is expressed as:

[0122]

[0123] In the above formula, P i and Q i is the output power of the distribution network to node i;

[0124] P DGi and Q DGi is the injected power of node i to the distributed generation;

[0125] P Di and Q Di is the load power of node i;

[0126] U i and U j is the voltage of nodes i and j;

[0127] G ij , B ij is the branch conductance and susceptance;

[0128] α ij is the voltage phase angle difference between nodes i and j.

[0129] The normalization process is performed based on the objective function and the constraint conditions to further optimize the fault recovery model, including minimum network loss and maximum power failure recovery;

[0130] The calculation method of the minimum network loss is:

[0131]

[0132] In the above formula, E is the total network loss after the power grid fault is restored, E s The total network loss before the active distribution network is restored;

[0133] The calculation method of the maximum power failure recovery is:

[0134]

[0135] In the above formula, P is the restored power supply load power, P s is the total power of the power-off load.

[0136] Based on the fault recovery model, the optimal fault recovery solution is solved and calculated, which combines the improved binary particle swarm algorithm and genetic algorithm. The specific method is as follows:

[0137] S1, initialize the number of particles N, particle dimension D, maximum number of iterations S, crossover probability P c , mutation probability P m ;

[0138] S2, calculate the fitness of each particle, search the feasible region of the island, and take the current particle position as the historical optimal solution;

[0139] S3, update the velocity and position of the particle;

[0140] S4, perform a crossover operation on two random particles in the population to generate new particles P and Q, calculate their fitness, and select the new particle with the largest fitness;

[0141] S5, perform mutation operation on the new particles, select a certain number of particles using the mutation probability, calculate the fitness value, and select the new particles with larger values;

[0142] S6, perform branch capacity, node voltage, distribution network structure, and power flow calculations, solve the load recovery amount, save the better particles, and complete an iteration;

[0143] S7, determining whether the maximum number of iterations has been reached, if not, repeating S3 to S6 until the maximum number of iterations has been reached;

[0144] S8. Output the optimal solution for fault recovery and system network loss, and the calculation ends.

[0145] The maximum number of iterations S is 100.

[0146] In the improved binary particle swarm algorithm, the position and velocity of each particle represent the probability of 0 and 1, and are updated through specific iteration rules. Assuming that the current particle swarm has a b-dimensional space, the position of particle i is X i =(K i1 ,K i2 ,…,K ib), whose velocity is represented by v i =(v i1 ,v i2 ,…,v ib ), the speed and position update rules of each particle iteration are as follows:

[0147]

[0148] In the above formula, ω is the inertia factor, which is a positive integer;

[0149] c1 and c2 are learning factors;

[0150] r1 and r2 are random numbers between [0,1];

[0151] rand() is a random number on [0, 1];

[0152] S(v ib ) is the sigmoid function.

[0153] The genetic algorithm uses the variance σ 2 Adjust the fitness value of the population, use crossover mutation to form a new population, enhance the global search capability, and thus solve the local optimal problem;

[0154]

[0155] f s =max{1,max|f i -f avg |

[0156] In the above formula, f i is the solution of iteration i;

[0157] f avg is the average iteration value;

[0158] f s is the normalization factor of fitness.

[0159] The α1, α2, α3, and α4 are 0.2, 0.3, 0.2, and 0.3 respectively;

[0160] F R =min(0.2F1+0.3F2+0.3F4)+max0.2F3.

[0161] Starting from the four key goals of minimizing grid loss F1, the number of switch operations F2, the operation of controllable DG F3 and the average user downtime F4, the present invention follows the system operation safety constraints under multi-source load access, establishes a fault recovery model, and adopts a combination method of improved binary particle swarm algorithm and genetic algorithm to find the best fault recovery plan, so as to fully coordinate various controllable DGs, improve power supply reliability, and meet the distribution network fault recovery requirements.

[0162] Embodiment 2:

[0163] This embodiment selects the IEEE 33-node distribution network basic example as an experimental example to verify the effectiveness of the fault recovery system described in Example 1.

[0164] The rated voltage of the system is 12.66 kV, the reference power is 10 MVA, and the locations of distributed power sources accessible in the system are node 4, node 15, and node 31, including wind power, photovoltaic power, and energy storage. α1, α2, α3, and α4 are taken as 0.2, 0.3, 0.2, and 0.3 respectively. This embodiment considers that the fault occurs in the lower-level power grid area, so the system side maintains stable power supply.

[0165] F R =min(0.2F1+0.3F2+0.3F4)+max0.2F3

[0166] The grid loss minimization is expressed as:

[0167]

[0168] In the above formula, N is the total number of branches in the distribution network;

[0169] S i is the position of the switch on branch i, which is 0 when not connected and 1 when actively connected;

[0170] P i and Q i are the active power and reactive power losses on branch i respectively;

[0171] U i is the voltage of its i-node, R i is the resistance of the i-th branch;

[0172] The number of switch operations is expressed as:

[0173]

[0174] In the above formula, S s Indicates the initial closed state of the Sth branch switch;

[0175] Ts and O s They represent the pre-fault section switches and the pre-fault tie switch set respectively;

[0176] The controllable DG operation is expressed as:

[0177]

[0178] In the above formula, N is the total number of DGs put into operation;

[0179] m represents the mth unit in operation;

[0180] Represents the output power of unit m;

[0181] The average user downtime is expressed as:

[0182]

[0183] The fault recovery constraints include node voltage constraints, branch capacity constraints, ring network constraints and power flow constraints.

[0184] The fault recovery constraint conditions include node voltage constraint, branch capacity constraint, ring network constraint and power flow constraint, wherein the node voltage constraint is expressed as:

[0185] U i,min ≤U i ≤U i,max

[0186] In the above formula, U i,min and U i,max are the minimum and maximum values ​​of the voltage at node i, respectively;

[0187] The branch capacity constraint is expressed as:

[0188] |P i |≤P i,max

[0189] In the above formula, P i is the active power flowing through branch i, P i,max is the maximum capacity of the branch;

[0190] The ring network restriction is expressed as:

[0191]

[0192] In the above formula, T i is the closed state of the i-th switch, It is a collection of switches in a closed circuit, and there is no ring network structure in the distribution network;

[0193] The power flow constraint is expressed as:

[0194]

[0195] In the above formula, P i and Q i is the output power of the distribution network to node i;

[0196] P DGi and Q DGi is the injected power of node i to the distributed generation;

[0197] P Di and Q Di is the load power of node i;

[0198] U i and U j is the voltage of nodes i and j;

[0199] G ij , B ij is the branch conductance and susceptance;

[0200] α ij is the voltage phase angle difference between nodes i and j.

[0201] The normalization process is performed based on the objective function and the constraint conditions to further optimize the fault recovery model, including minimum network loss and maximum power failure recovery;

[0202] The calculation method of the minimum network loss is:

[0203]

[0204] In the above formula, E is the total network loss after the power grid fault is restored, E s The total network loss before the active distribution network is restored;

[0205] The calculation method of the maximum power failure recovery is:

[0206]

[0207] In the above formula, P is the restored power supply load power, P s is the total power of the power-off load.

[0208] Based on the fault recovery model, the optimal fault recovery solution is solved and calculated, which combines the improved binary particle swarm algorithm and genetic algorithm. The specific method is as follows:

[0209] S1, initialize the number of particles N, particle dimension D, maximum number of iterations S, crossover probability P c , mutation probability P m ;

[0210] S2, calculate the fitness of each particle, search the feasible region of the island, and take the current particle position as the historical optimal solution;

[0211] S3, update the velocity and position of the particle;

[0212] S4, perform a crossover operation on two random particles in the population to generate new particles P and Q, calculate their fitness, and select the new particle with the largest fitness;

[0213] S5, perform mutation operation on the new particles, select a certain number of particles using the mutation probability, calculate the fitness value, and select the new particles with larger values;

[0214] S6, perform branch capacity, node voltage, distribution network structure, and power flow calculations, solve the load recovery amount, save the better particles, and complete an iteration;

[0215] S7, determining whether the maximum number of iterations has been reached, if not, repeating S3 to S6 until the maximum number of iterations has been reached;

[0216] S8. Output the optimal solution for fault recovery and system network loss, and the calculation ends.

[0217] The maximum number of iterations S is 100.

[0218] Assume that a permanent fault occurs on branch 2-3 at 8 am and it takes 6 hours to restore power supply. During this time period, in an active distribution network with controllable loads, fault recovery can be divided into two cases:

[0219] (1) Without considering the coordination between various energy sources, the fault recovery of the power-lost load branch is carried out according to the traditional distribution network fault recovery method;

[0220] (2) Considering the coordinated interaction of multiple sources, controllable distributed power sources and energy storage devices are used to realize regional island division, dynamically complete the power supply of some nodes, and the ADN determines the recovery path of the load of the node where the power supply has not been restored.

[0221] Based on the ADN fault recovery research, the fault recovery in the first case achieved power supply restoration of the power-lost load in the non-fault area, with a minimum network loss of 173.65kW.

[0222] In the second case, the fault recovery scenario considers the coordinated interaction of multiple sources and adjusts different fault recovery structures according to the power output in different time periods to achieve the optimal fault recovery solution.

[0223] In the fault recovery scheme under the multi-source interaction scenario, the isolated power supply domain in each period changes over time. The fault recovery mathematical model of ADN is solved to obtain the fault recovery results in each period. The results show that the power-off load node can be restored.

[0224] In the multi-source interaction scenario, the fault recovery scheme considering the controllable DG output performs better than the scheme not considering the controllable DG output, which is manifested as lower system network loss. The simulation results show that after the fault occurs, all power-off loads can successfully regain power, and the system network loss also decreases slightly, which means that the overall node voltage is effectively improved. This finding illustrates the effectiveness of the recovery model proposed in Example 1, and provides important support for the robustness and reliability of the power system.

[0225] The implementation principle of the above embodiment is:

[0226] The fault recovery model building module is used to build a fault recovery model based on maximizing the number of random operations of controllable distributed generation and minimizing power loss, and following the system operation safety constraints under multi-source load access; the objective function includes minimization of grid loss F1, number of switch operations F2, controllable DG operation F3 and average user downtime F4. Fault recovery constraints include node voltage constraints, branch capacity constraints, ring network restrictions and power flow constraints.

[0227] A model solving module, used to solve and calculate the best fault recovery solution based on the fault recovery model, integrating an improved binary particle swarm algorithm and a genetic algorithm;

[0228] The power generation calculation module is used to send fault recovery control instructions and perform fault recovery based on the best fault recovery plan.

[0229] Starting from the four key goals of minimizing grid loss F1, the number of switch operations F2, the operation of controllable DG F3 and the average user downtime F4, the present invention follows the system operation safety constraints under multi-source load access, establishes a fault recovery model, and adopts a combination method of improved binary particle swarm algorithm and genetic algorithm to find the best fault recovery plan, so as to fully coordinate various controllable DGs, improve power supply reliability, and meet the distribution network fault recovery requirements.

[0230] The embodiments of this specific implementation method are all preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Therefore, any equivalent changes made based on the structure, shape, and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A coordinated fault recovery system for a main power grid and a distribution network, characterized in that: It includes a fault recovery model building module, which is used to build a fault recovery model based on maximizing the number of random operations of controllable distributed power sources and minimizing power loss, and following the system operation safety constraints under multi-source load access; A model solving module, used to solve and calculate the best fault recovery solution based on the fault recovery model; The power generation calculation module is used to send fault recovery control instructions and perform fault recovery based on the best fault recovery plan.

2. A coordinated fault recovery system for a main power grid and a distribution network according to claim 1, characterized in that: The fault recovery model includes establishing an objective function and fault recovery constraints, and performing normalization processing based on the objective function and the constraints to further optimize the fault recovery model; The objective function includes minimization of grid loss F1, number of switch operations F2, controllable DG operation F3 and average user downtime F4. The fault recovery model is as follows: F R =min(α1F1+α2F2+α4F4)+maxα3F3 Among them, α1, α2, α3, and α4 are weight coefficients, and their sum is 1; The fault recovery constraints include node voltage constraints, branch capacity constraints, ring network constraints and power flow constraints.

3. A coordinated fault recovery system for a main power grid and a distribution network according to claim 2, characterized in that: The normalization process is performed based on the objective function and the constraint conditions to further optimize the fault recovery model, including minimum network loss and maximum power failure recovery; The calculation method of the minimum network loss is: In the above formula, E is the total network loss after the power grid fault is restored, E s The total network loss before the active distribution network is restored; The calculation method of the maximum power failure recovery is: In the above formula, P is the restored power supply load power, P s is the total power of the power-off load.

4. A coordinated fault recovery system for a main power grid and a distribution network according to claim 2, characterized in that: The objective function includes minimization of grid loss, number of switch operations, controllable DG operation and average user downtime, wherein the grid loss minimization is expressed as: In the above formula, N is the total number of branches in the distribution network; S i is the position of the switch on branch i, which is 0 when not connected and 1 when actively connected; P i and Q i are the active power and reactive power losses on branch i respectively; U i is the voltage of its i-node, R i is the resistance of the i-th branch; The number of switch operations is expressed as: In the above formula, S s Indicates the initial closed state of the Sth branch switch; T s and O s They represent the pre-fault section switches and the pre-fault tie switch set respectively; The controllable DG operation is expressed as: In the above formula, N is the total number of DGs put into operation; m represents the mth unit in operation; Represents the output power of unit m; The average user downtime is expressed as:

5. A coordinated fault recovery system for a main power grid and a distribution network according to claim 2, characterized in that: The fault recovery constraint conditions include node voltage constraint, branch capacity constraint, ring network constraint and power flow constraint, wherein the node voltage constraint is expressed as: IN i,min ≤U i ≤U i,max In the above formula, U i,min and U i,max are the minimum and maximum values ​​of the voltage at node i, respectively; The branch capacity constraint is expressed as: |P i |≤P i,max In the above formula, P i is the active power flowing through branch i, P i,max is the maximum capacity of the branch; The ring network restriction is expressed as: In the above formula, T i is the closed state of the i-th switch, It is a collection of switches in a closed circuit, and there is no ring network structure in the distribution network; The power flow constraint is expressed as: In the above formula, P i and Q i is the output power of the distribution network to node i; P DGi and Q DGi is the injected power of node i to the distributed generation; P Di and Q Di is the load power of node i; U i and U j is the voltage of nodes i and j; G ij , B ij is the branch conductance and susceptance; α ij is the voltage phase angle difference between nodes i and j.

6. A coordinated fault recovery system for a main power grid and a distribution network according to claim 1, characterized in that: Based on the fault recovery model, the optimal fault recovery solution is solved and calculated, which combines the improved binary particle swarm algorithm and genetic algorithm. The specific method is as follows: S1, initialize the number of particles N, particle dimension D, maximum number of iterations S, crossover probability P c , mutation probability P m ; S2, calculate the fitness of each particle, search the feasible region of the island, and take the current particle position as the historical optimal solution; S3, update the velocity and position of the particle; S4, perform a crossover operation on two random particles in the population to generate new particles P and Q, calculate their fitness, and select the new particle with the largest fitness; S5, perform mutation operation on the new particles, select a certain number of particles using the mutation probability, calculate the fitness value, and select the new particles with larger values; S6, perform branch capacity, node voltage, distribution network structure, and power flow calculations, solve the load recovery amount, save the better particles, and complete one iteration; S7, determining whether the maximum number of iterations has been reached, if not, repeating S3 to S6 until the maximum number of iterations has been reached; S8. Output the optimal solution for fault recovery and system network loss, and the calculation ends.

7. A coordinated fault recovery system for a main power grid and a distribution network according to claim 6, characterized in that: In the improved binary particle swarm algorithm, the position and velocity of each particle represent the probability of 0 and 1, and are updated through specific iteration rules. Assuming that the current particle swarm has a b-dimensional space, the position of particle i is X i =(K i1 ,K i2 ,…,K ib ), whose velocity is represented by v i =(v i1 ,v i2 ,…,v ib ), the speed and position update rules of each particle iteration are as follows: In the above formula, ω is the inertia factor, which is a positive integer; c1 and c2 are learning factors; r1 and r2 are random numbers between [0,1]; rand() is a random number on [0, 1]; S(v ib ) is the sigmoid function.

8. A coordinated fault recovery system for a main power grid and a distribution network according to claim 7, characterized in that: The genetic algorithm uses the variance σ 2 Adjust the fitness value of the population, use crossover mutation to form a new population, enhance the global search capability, and thus solve the local optimal problem; f s =max{1,max|f i -f avg | In the above formula, f i is the solution of iteration i; f avg is the average iteration value; f s is the normalization factor of fitness.

9. A coordinated fault recovery system for a main power grid and a distribution network according to claim 2, characterized in that: The α1, α2, α3, and α4 are 0.2, 0.3, 0.2, and 0.3 respectively; F R =min(0.2F1+0.3F2+0.3F4)+max0.2F3。 10. A coordinated fault recovery system for a main power grid and a distribution network according to claim 6, characterized in that: The maximum number of iterations S is 100.

Images