Alternating current and direct current hybrid power distribution network fault recovery method based on network reconstruction and opportunity constraint

Abstract

The invention discloses an AC / DC hybrid power distribution network fault recovery method based on network reconstruction and opportunity constraint, and belongs to the field of AC / DC power distribution network fault recovery. The method comprises the following steps: step 1, establishing a flexible interconnected power distribution network frame reconstruction model; 2, solving the model constructed in the step 1 based on a binary particle swarm algorithm; 3, constructing an interconnected power distribution network fault recovery model considering opportunity constraints; step 4, carrying out uncertainty scene modeling of wind and light output; 5, converting and solving the opportunity constraint model; and 6, verifying the feasibility of the model through example analysis. According to the method, the flexible internet distribution network rack reconstruction model and the interconnected distribution network fault recovery model considering the chance constraint are established, so that the operation time consumption of the optimization process can be greatly shortened on the basis of effectively ensuring the load recovery effect after the fault, the fault recovery efficiency is remarkably improved, and the fault recovery efficiency is improved. And the method is of great significance to security guarantee of a large-scale power grid.

Description

Technical Field

[0001] The invention belongs to the field of AC-DC distribution network fault restoration, and relates to a flexible interconnected distribution network fault restoration method based on network reconfiguration and chance constraints. Background Art

[0002] With the rapid development of DC power sources such as photovoltaic power and DC loads such as charging piles, DC distribution systems have been widely used. Compared with AC distribution systems, DC distribution systems have the advantages of high transmission power, low network loss, and flexible power control. The AC-DC hybrid distribution system is composed of the interconnection of a DC distribution system and an AC distribution system. By adjusting the power of the voltage source converter (VSC), flexible power transmission can be achieved between the AC and DC systems. It can not only improve the absorption capacity of distributed power sources by the distribution network, but also break through the radial constraint bottleneck of traditional AC distribution networks, providing a new solution for the fault restoration of distribution networks. And with the increasing number of DC loads, AC-DC hybrid distribution networks are receiving more and more attention and will be an important form of future distribution networks. So far, AC-DC hybrid distribution networks have been widely studied and applied.

[0003] Improving the fault restoration ability and ensuring continuous and reliable power supply to loads are of great significance for modern distribution networks. Long-term interruption of loads may cause significant economic losses and affect social production and life. After a fault occurs, the load between lines can be transferred through network reconfiguration, and then the power grid above can restore the lost power loads in the non-fault area. In addition, local resources such as distributed generation (DG) can also be used for load restoration.

[0004] Due to the uncertainty of DG in the distribution network affected by various factors, the existence of its uncertainty may cause large fluctuations in line transmission power and voltage. During the fault restoration process of the distribution network, if deterministic constraints are still used, DG cannot fully exert its capabilities, making the fault restoration process of the distribution network tend to be conservative. Therefore, considering uncertainty is particularly important in the fault restoration of distribution networks. However, the existing methods do not consider the impact of the uncertainty of DG on the fault restoration of AC-DC hybrid distribution networks. Summary of the Invention

[0005] The purpose of the invention is to propose an AC-DC hybrid distribution network fault restoration model with chance constraints, which is reliable, reasonable, and practical. The model fully considers the impact of the uncertainty of DG on fault restoration, and verifies the effectiveness of the model through examples. It can be applied to the fault restoration problem of AC-DC hybrid distribution networks and has practical significance for solving the fault restoration problem of AC-DC hybrid distribution networks with uncertainty.

[0006] The object of the present invention is achieved as follows: A fault recovery method for an AC-DC hybrid distribution network based on network reconstruction and chance-constrained programming, which comprises the following steps:

[0007] Step 1: Establish a flexible interconnected distribution network framework reconstruction model;

[0008] Step 2: Solve the model constructed in Step 1 based on the binary particle swarm optimization algorithm;

[0009] Step 3: Construct a fault recovery model for the interconnected distribution network considering chance constraints;

[0010] Step 4: Conduct uncertainty scenario modeling for the wind and light power output;

[0011] Step 5: Transformation and solution of the chance-constrained model;

[0012] Step 6: Verify the feasibility of the model through case analysis;

[0013] The process of reconstructing the interconnected distribution network framework with flexible interconnected devices as the core in Step 1 is as follows:

[0014] (1) First, determine the DC power supplyable node set, and establish the DC power supplyable node set through breadth-first search, representing all possible connection relationships between the power source node set and the VSC node set, that is:

[0015] N DS ={N S ,N VSC ,D S-VSC} (1)

[0016] In the formula: N S 、N VSC are the power source node and the VSC node set respectively; D S-VSC is the set of all possible connection paths between N S and N VSC ; N DS is the DC power supplyable node set, and its elements refer to all load nodes included in the D S-VSC path.

[0017] (2) Then determine the set of possible paths between the load node set and the DC power supplyable node set. Starting from the elements in the system load node set N L and ending with the elements in N DS , establish the possible paths between any load node and the DC power supplyable node, that is:

[0018]

[0019] (3) Finally, establish the objective of the grid framework reconstruction. Since the system operation loss is positively correlated with the network impedance, according to the principle of the shortest path (minimum impedance), the objective of the distribution network grid framework reconstruction with flexible interconnection equipment as the core is constructed, which can be specifically expressed as:

[0020]

[0021] In the formula: is the branch between nodes i and j in the wth AC subsystem, is the total number of nodes included in the wth AC subsystem; W ac is the total number of AC subsystems.

[0022] In the interconnected distribution network fault recovery model in step 3, the objective is to minimize the load shedding amount after the fault occurs, and to ensure the reliable power supply to important loads as much as possible. Therefore, the objective function of the fault recovery model can be expressed as:

[0023]

[0024] In the formula: P i L , P i R are the total load and the load recovery amount carried by node i respectively; N B is the total number of all nodes included in the distribution network; ξ i is the grade weight of the load carried by node i. The weight coefficients of the first, second, and third grade loads are taken as 100, 10, and 1 respectively. The constraint conditions mainly include: AC and DC network power flow constraints; VSC steady-state operation constraints; transmission capacity constraints; voltage security constraints; chance constraints; here the transmission capacity constraints mainly include the transmission capacity constraints of AC and DC distribution lines and VSCs. The steady-state operation constraints of VSCs will be explained in detail in combination with the diagram.

[0025] 1) AC and DC network power flow constraints:

[0026]

[0027] In the formula: Formulas (5), (6) and (8), (9) are the Dist-Flow power flow constraints of the AC and DC networks respectively, and formulas (7) and (10) are their standard second-order cone forms. are the active and reactive powers transmitted by the AC distribution line between nodes i and j respectively; is the active power transmitted by the DC distribution line between nodes i and j; and are the active and reactive powers injected by node j respectively; are the square terms of the voltage amplitudes of node i respectively; They are the square terms of the currents flowing through the AC and DC distribution lines between nodes i and j; R ij , X ij are the resistance and reactance of the distribution line between nodes i and j respectively; N B,ac , N B,dc are the total numbers of AC and DC nodes in the system respectively.

[0028] The following formula is for the chance constraint:

[0029] First, relax the power balance equations shown in Eqs. (5) and (8) into the following inequality constraint form:

[0030]

[0031] On this basis, transform the inequality constraints shown in Eqs. (11) and (12) into a probability constraint problem that satisfies a certain confidence level, that is:

[0032]

[0033] where: Pr{z} represents the probability that event z holds; β P is the confidence level of the inequality constraints shown in Eqs. (11) and (12), and 0 ≤ β P ≤ 1.

[0034]

[0035] Voltage constraint:

[0036]

[0037] are the active and reactive powers transmitted by the AC distribution line between nodes i and j respectively; is the active power transmitted by the DC distribution line between nodes i and j; and are the upper limits of the transmission capacities of the AC and DC distribution lines and the VSC respectively. Eqs. (15) and (16) are the transmission capacity constraints.

[0038] In step 4, the uncertainty scenario modeling of the wind and light output is carried out. Since the established fault recovery model of the flexible interconnected distribution network is a stochastic programming problem containing uncertain variables of distributed photovoltaic and wind power, this scheme adopts a method of stochastic optimization based on probability scenarios. By using the uncertainty probability distribution, typical photovoltaic and wind power output states are generated, and then the original model is converted into a stochastic optimization model described by the uncertainty states and their corresponding probabilities.

[0039] To analyze the uncertainty of the photovoltaic and wind turbine output powers P PV and P WT of, PPV and P WT are respectively divided into M and Y states, and there are upper and lower power limits for any state m (m = 1, 2, …, M) and there are upper and lower power limits for any state y (y = 1, 2, …, Y) Then, under the limit constraints determined by the probability density functions f(P PV ), f(P WT ), the weighted average of the output powers of the photovoltaic and wind turbines within the limits of each state can be respectively expressed as:

[0040]

[0041] Furthermore, the probabilities corresponding to any states m and y can be represented by the probability distribution functions F(P PV ), F(P WT ) obtained by integrating the probability density functions f(P PV ), f(P WT ) from the lower limit to the upper limit, that is:

[0042]

[0043] In summary, using the average output powers of the photovoltaic and wind turbines in the above different states and their corresponding probabilities, N un (N un = M × Y) uncertain state combinations are generated, and the general determination steps for this combination are 3 steps:

[0044] Construct the average output power scenario sets ψ PV and ψ WT of the photovoltaic and wind turbines respectively, that is:

[0045]

[0046] Pair and group the average output powers of the photovoltaic and wind turbines using the Cartesian product of the two sets shown in Equation (22):

[0047]

[0048] Calculate the probabilities corresponding to each uncertain state combination:

[0049]

[0050] For the transformation and solution of the chance-constrained model in step 5, this solution uses the sample average approximation method to transform and solve the distribution network fault restoration model considering chance constraints. Its principle is to use the sampling idea to represent random variables with samples, thereby transforming the stochastic optimization problem into a deterministic problem for solution.

[0051] The unified form of the chance constraint can be expressed as:

[0052] Pr{F(x,ζ)>0}≥β P (25)

[0053] In the formula: F(x,ζ)>0 represents the constraint condition; x and ζ represent ordinary variables and random variables respectively.

[0054] Performing SAA approximation on formula (25) is converted to:

[0055]

[0056] In the formula: E(F(x,ζ i )) is an indicator function for judging whether the constraint condition holds. It takes 1 when the constraint condition holds and 0 otherwise; ζ i is a certain sample obtained by sampling; γ represents the confidence level when the condition holds for sample average approximation.

[0057] Therefore, according to the SAA method, the chance constraints of the distribution network fault restoration models shown in formulas (13) and (14) can be respectively converted to:

[0058]

[0059]

[0060] It should be noted that the indicator function E(·) in the above 4 formulas is still a non-convex function and cannot be directly solved. Therefore, binary variables and are introduced to represent the sum of the indicator functions as the sum of the binary variables of all samples in different scenarios. The above process can be described by formulas (32) - (35), that is:

[0061]

[0062] In formulas (32) and (33): K is a very large positive number, which can always ensure that when and take 1, formulas (32) and (33) definitely hold; when or take 0, it means that the corresponding l-th sampling state is taken into account, that is, the determined fault restoration plan can ensure the power balance and safe operation of the system under state l; when Or When taking 1, it means that state l is not considered, that is, the determined fault recovery plan cannot guarantee the power balance and safe operation of the system under state l.

[0063] After the above deterministic transformation of the chance constraint, the flexible interconnected distribution network fault recovery model considering the chance constraint proposed in this solution has been transformed into a mixed-integer second-order cone programming model, which can be efficiently solved by relying on Yamlip to call the CPLEX commercial solver.

[0064] A fault recovery method for AC-DC hybrid distribution network based on network reconfiguration and chance constraint of the present invention is applied to the load fault recovery scenario with uncertainty, and a flexible interconnected distribution network fault recovery strategy considering network reconfiguration and chance constraint is proposed. First, a network reconfiguration method for interconnected distribution network with flexible interconnected devices as the core is proposed, and the system topology is generated by the binary particle swarm optimization (Binary-PSO) algorithm. Then, considering the uncertainty of DG output, by constructing an uncertainty probability scenario set, an interconnected distribution network load recovery chance constraint model with the goal of maximizing load hierarchical recovery is proposed, and the sample average approximation (SAA) method is used to transform and solve the proposed model. Finally, based on the interconnected IEEE 33-node distribution network test case, the effectiveness and superiority of the fault recovery strategy proposed in the present invention are verified. By establishing a flexible interconnected distribution network framework reconfiguration model and constructing an interconnected distribution network fault recovery model considering chance constraint, the present invention can significantly shorten the operation time of the optimization process and significantly improve the fault recovery efficiency on the basis of effectively guaranteeing the load recovery effect after the fault, which is of great significance for the security guarantee of large-scale power grids. Description of the Drawings

[0065] Figure 1 It is a schematic diagram of the reconfiguration of a flexible interconnected distribution network.

[0066] Figure 2 It is a framework diagram of the flexible interconnected distribution network fault recovery model.

[0067] Figure 3 It is a flow chart of the binary PSO algorithm for solving.

[0068] Figure 4 It is an equivalent circuit diagram of the VSC.

[0069] Figure 5 It is a structure diagram of a multi-region flexible interconnected distribution network.

[0070] Figure 6 It is a distribution diagram of the active load of the node.

[0071] Figure 7 It is the probability density state division of photovoltaic and wind power and the corresponding probability graph.

[0072] Figure 8 It is the VSC transmission power graph before and after reconstruction.

[0073] Figure 9 It is the load restoration result graph before and after reconstruction.

[0074] Figure 10 It is the load restoration result graph of different reconstruction methods.

[0075] Figure 11 It is the load restoration result graph considering the chance constraint or not.

[0076] Figure 12 It is the load restoration result under different confidence levels;

[0077] Figure 13 It is the proportion graph of power balance violation scenarios under different confidence levels. Specific implementation manners

[0078] The preferred embodiments will be described in detail below with reference to the accompanying drawings.

[0079] Figure 1 It is a schematic diagram of the flexible interconnected distribution network reconstruction. Taking the three-terminal flexible interconnected distribution network as an example, after a fault occurs, the system forms several island regions IR, which are specifically described as follows:

[0080] ① First, through network reconstruction with the flexible DC interconnected link based on VSC as the core, by switching the distribution lines and tie lines, and the flexible power regulation function of the flexible interconnected device, a power supply restoration area for multiple power sources to multiple load nodes is constructed to achieve power mutual assistance and load restoration between each feeder or different power supply areas.

[0081] ② To ensure the operation economy of the system, the above reconstruction network aims at the minimum total impedance of the distribution lines and is solved by the Binary-PSO algorithm. Of course, for the island IR2 that cannot be connected to the VSC node, it can only rely on the DG in the island to supply power to the load, and when the DG output is insufficient, the load can only be shed.

[0082] Figure 2 It is a framework diagram of the flexible interconnected distribution network fault recovery model. The proposed flexible interconnected distribution network fault recovery strategy mainly includes two stages, which are specifically described as follows.

[0083] ① Network reconstruction with flexible interconnected devices as the core. According to Figure 1 the reconstruction process of the distribution network shown, network reconstruction with flexible interconnected devices as the core is carried out.

[0084] ②Load restoration considering the uncertainty of distributed photovoltaic and wind power. The load restoration model in this part transforms the proposed model into a deterministic model through a stochastic optimization method based on uncertainty scenarios and is solved by a solver. The two stages work together to effectively ensure the efficient power supply restoration to the load after a fault.

[0085] Figure 3 It is the flow chart of the binary PSO algorithm for solving the network reconfiguration model in Step 1. The specific algorithm flow is described as follows.

[0086] ①Each particle is encoded with binary variables 0 - 1, where 1 represents the switch is closed and 0 represents the switch is open; the length of the particle is the total number of switches in the network, and the number of particle groups is the number of loops in the network; the initial state of the particle is that all switches in the network are closed. The update rule of the particle is as follows. represent the individual and global best positions respectively; and are the velocity and position of the i - th particle at the s - th iteration respectively;

[0087]

[0088] ②The specific steps of the solution process of the binary PSO (Particle Swarm Optimization) algorithm are as follows:

[0089] Step 1: Initialize the distribution network parameters and particles.

[0090] Step 2: Verify whether the network is radial and whether there are islands through the parameters input in Step 1. If there are islands or it is not radial, return to Step 1; otherwise, go to the next step.

[0091] Step 3: Calculate the fitness of the particle swarm and sort it to obtain the individual and global optimal values.

[0092] Step 4: Update the position and velocity of the particles and calculate the fitness value of the new particles.

[0093] Step 5: Compare the fitness of the old and new particles and calculate the position of the particles with worse fitness.

[0094] Step 6: Update the historical individual optimal and global optimal values of the particles.

[0095] Step 7: Judge whether the iteration convergence condition is satisfied. If it is satisfied, output the final network reconfiguration result; otherwise, return to Step 1.

[0096] Figure 4 It is the equivalent circuit diagram of the VSC, and the specific description is as follows:

[0097] ①As Figure 4 is the equivalent circuit diagram of the VSC. In the figure, and respectively exchange the active and reactive powers input by the lateral VSC; is the active power output by the VSC on the DC side; is the equivalent reactive power inside the VSC; and are the voltages on the AC and DC sides of the VSC respectively; is the internal voltage of the VSC; I n 、R n and X n are the current, resistance and reactance of the equivalent branch of the VSC respectively.

[0098] ② The steady-state operation constraints of the VSC are:

[0099]

[0100] In the formula: l n is the square term of I n . Additionally, assuming that the VSC adopts the SPWM modulation method, the intermediate variable in Equation 37, i.e., the DC voltage utilization rate of the VSC, takes The modulation ratio of the VSC takes 0 ≤ M n ≤ 1.

[0101] Figure 5 is the structure diagram of the multi-region flexible interconnected distribution network, and the specific description is as follows;

[0102] To verify the effectiveness and superiority of the flexible interconnected distribution network fault recovery strategy considering network reconfiguration and chance constraints proposed in this embodiment, a hybrid AC-DC distribution network example with a 3-standard IEEE 33-node distribution network interconnected by VSCs is established, Figure 5 which shows its grid structure, PV and wind power access points and fault disconnection positions.

[0103] Figure 6 is the active load distribution diagram of the nodes, and the specific description is as follows.

[0104] ① The base voltage of the example system is taken as 12.66 kV; the upper and lower limits of the allowable operating voltage of the system are 1.05 p.u. and 0.95 p.u.;

[0105] ② Empirical mode decomposition is used to filter the original data samples. The figure shows the active load and corresponding importance degree of each node in each IEEE 33-node system. The reactive load is proportional to the active load according to the standard example parameters.

[0106] Figure 7It is the probability density state division of photovoltaic and wind power and their corresponding probability diagrams. Since the distribution network covers a relatively small geographical area, it is assumed that the photovoltaic and wind power outputs connected to each node in the system have consistent fluctuation states respectively. The probability density curves of photovoltaic and load power can be obtained through the proposed probability scenario generation method. The final 50 representative scenarios of photovoltaic and wind power are obtained through the following formula and applied to solve the fault recovery model of the flexible interconnected distribution network considering chance constraints.

[0107] The results of the probability density state division of photovoltaic and wind power and their corresponding scenario probabilities are specifically as Figure 7 shown.

[0108]

[0109]

[0110] The average output power scenario sets ψ PV and ψ WT of photovoltaic and wind turbines are constructed respectively by Equation (38). Equation (39) is the formula for pairing and grouping the average output powers of photovoltaic and wind turbines using the Cartesian product of the two sets shown in Equation (38). and are the weighted average values of the output powers of photovoltaic and wind turbines within their respective state limits respectively. Equation (40) calculates the probabilities corresponding to each uncertain state combination.

[0111] Figure 8 and Figure 9 are the VSC transmission power diagrams before and after reconstruction and the load recovery result diagrams before and after reconstruction. Under the fault conditions shown in Figure 4 , the fault recovery model is used to perform fault recovery on the interconnected distribution network with and without considering network reconstruction respectively. It should be noted that the photovoltaic and wind power in the two comparison schemes adopted in this embodiment are determined values weighted by scenarios and their corresponding probabilities, and the chance constraints of uncertainty are not considered. The network reconstruction results of the method proposed in this embodiment are shown in the following table. The transmission powers of the VSC and the load recovery results of various types of systems under the two schemes are specifically as Figure 8 and Figure 9 shown.

[0112]

[0113] Reference Figure 10 is to verify the superiority of the network reconstruction method proposed in this embodiment compared with the existing reconstruction methods. By comparing the fault recovery strategies with the optimization objectives of minimizing network loss, minimizing voltage deviation, combining minimizing network loss and voltage deviation, and maximizing load recovery amount, and the load recovery effects of the method in this embodiment, the statistical results of the operation time consumption of load recovery under each reconstruction scheme are shown in the following table:

[0114]

[0115] It can be seen from Figure 10 that the influence of the above different reconstruction schemes on the load recovery effect is not obvious, and the deviation of the total load recovery capacity of each scheme is within 1.59%. In addition, it can be seen from the operation time-consuming statistical results shown in the above table that under the network reconstruction strategy with flexible interconnection equipment as the core proposed in this embodiment, the total operation time-consuming for load recovery is significantly reduced, and the maximum reduction rate reaches 99.54%. This is because the method proposed in this embodiment constructs the network reconstruction with flexible DC equipment as the core only based on the network impedance relationship, without involving the optimization of the operation state of the distribution network. The dimension of the optimization solution space is greatly reduced, and the operation complexity is significantly decreased. The above analysis shows that the method proposed in this embodiment can effectively ensure the load recovery effect after a fault, greatly shorten the operation time-consuming of the optimization process, and significantly improve the fault recovery efficiency.

[0116] Figure 11 are the load recovery result diagrams with and without considering chance constraints, and the specific descriptions are as follows:

[0117] Considering the uncertainties of photovoltaic and wind power, chance constraints are used to characterize and transform the uncertainties of the proposed fault recovery model. Based on the 50 typical scenarios of wind and light power generated in the above result diagrams, the initial confidence level is set to 100%, that is, the chance constraint model is solved under the condition that the uncertainty state fully meets the system operation constraints, and the load recovery result under 100% confidence level is obtained. Compared with the result of the deterministic model, considering the uncertainties of photovoltaic and wind power, the total load survival of the system after the fault has decreased, from 8.57 + j4.86 MW of the deterministic model to 7.81 + j4.37 MW, with a decrease of about 11%. This is because under the 100% confidence level considering the uncertainties of wind and light, the load recovery decision needs to ensure that the system can operate safely and reliably in any scenario, avoiding possible extreme fluctuations in wind and light power. It cannot achieve the optimal decision for any scenario, and there is a certain conservatism in the load recovery scheme made.

[0118] Figure 12 and Figure 13 are the load recovery results under different confidence levels and the proportion diagrams of power balance violation scenarios under different confidence levels, and the specific descriptions are as follows:

[0119] To further analyze the influence of the chance constraint model proposed in this embodiment on the system load recovery and safe operation when setting different confidence levels, this embodiment gives the statistical results of the load recovery results and the proportion of system power balance violation scenarios when the chance constraint confidence levels are 100%, 90%, 80%, 70%, 60%, 50% and 40% respectively. FromFigure 12 From the load recovery result diagrams at different confidence levels in Figure 13 it can be seen that as the confidence level gradually decreases, the load recovery amount of the system increases accordingly. During the process of the confidence level decreasing from 100% to 40%, the total capacity of the surviving load increases from 8.9518 MVA to 9.9138 MVA, an increase of 10.75%. However, as the confidence level decreases, the ability of the system to cope with uncertain factors decreases accordingly, resulting in a significant increase in the risk that the fault recovery strategy considering the chance constraint does not satisfy the power balance. The number of scenarios not satisfying the power balance constraint increases from 0 violations at 100% confidence level to 59.42% at 40% confidence level, as shown by the proportion of power balance violation scenarios at different confidence levels in

[0120] As mentioned above, the above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed by the present invention should be covered by the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. A fault recovery method for AC-DC hybrid distribution network based on network reconstruction and chance-constrained, characterized in that, It includes the following steps: Step 1: Establish a flexible interconnected distribution network framework reconstruction model; Step 2: Solve the model constructed in Step 1 based on the binary particle swarm optimization algorithm; Step 3: Construct an interconnected distribution network fault recovery model considering chance constraints; Step 4: Conduct uncertainty scenario modeling of the wind and light output; Step 5: Transformation and solution of the chance constraint model; Step 6: Verify the feasibility of the model through case analysis.

2. A fault recovery method for an AC / DC hybrid distribution network based on network reconstruction and chance-constrained, as claimed in claim 1, wherein: The process of reconstructing the interconnected distribution network framework with flexible interconnected devices as the core in Step 1 is as follows: (1) First, determine the set of DC power supplyable nodes, and establish the set of DC power supplyable nodes through breadth-first search, representing all possible connection relationships between the power source node set and the VSC node set, that is: Where: N S and N VSC are the power source node set and the VSC node set respectively; D S-VSC is the set of all possible connection paths between N S and N VSC ; N DS is the set of DC power supply nodes, and the elements therein refer to all load nodes included in the D S-VSC path; (2) Then determine the set of possible paths between the load node set and the DC power supplyable node set. Starting from the elements in the system load node set N L as the starting points and taking the elements in N DS as the ending points, establish the possible paths between any load node and the DC power supplyable node, that is: (3) Finally, establish the framework reconstruction objective. Since the system operation loss is positively correlated with the network impedance, according to the shortest path principle, construct the distribution network framework reconstruction objective with flexible interconnected devices as the core, which can be specifically expressed as: In the formula: is the branch between nodes i and j in the w-th AC subsystem; is the total number of nodes included in the w-th AC subsystem; W ac is the total number of AC subsystems.

3. A fault recovery method for an AC-DC hybrid distribution network based on network reconstruction and chance-constrained, as claimed in claim 1, wherein: The process of solving the model constructed in Step 1 based on the binary particle swarm optimization algorithm in Step 2 is as follows: Step 1: Initialize the distribution network parameters and particles; Step 2: Verify whether the network is radial and whether there are islands through the parameters input in Step 1. If there are islands or it is not radial, return to Step 1, otherwise proceed to the next step; Step 3: Calculate the fitness of the particle swarm and sort it to obtain the individual and global optimal values; Step 4: Update the position and velocity of the particles, and calculate the fitness value of the new particles; Step 5: Compare the fitness of the new and old particles, and calculate the position of the particles with worse fitness; Step 6: Update the historical individual optimal and global optimal values of the particles; Step 7: Judge whether the iterative convergence condition is satisfied. If it is satisfied, output the final framework reconstruction result, otherwise return to Step 1.

4. A fault recovery method for an AC / DC hybrid distribution network based on network reconstruction and chance-constrained, as claimed in claim 1, wherein: The objective function of the interconnected distribution network fault recovery model considering chance constraints in Step 3 is: Where: P i L and P i R are the total load and the load restoration amount carried by node i respectively; N B is the total number of all nodes included in the distribution network; ξ i is the grade weight of the load carried by node i. The weight coefficients of the first, second, and third-level loads are taken as 100, 10, and 1 respectively.

5. A fault recovery method for an AC-DC hybrid distribution network based on network reconstruction and chance-constrained, as claimed in claim 1, wherein: The constraint conditions in Step 3 include: AC and DC network power flow constraints; VSC steady-state operation constraints; transmission capacity constraints; voltage security constraints; chance constraints.

6. A fault recovery method for an AC-DC hybrid distribution network based on network reconstruction and chance-constrained, as claimed in claim 1, wherein: The specific process of conducting uncertainty scenario modeling of the wind and light output in Step 4 is as follows: Divide the photovoltaic output power P PV and the wind turbine output power P WT into M and Y states respectively, and there are upper and lower power limits for any state m (m = 1, 2,..., M) and there are upper and lower power limits for any state y (y = 1, 2,..., Y) Then, under the limit constraints determined by the probability density functions f(P PV ), f(P WT ), the weighted average of the output powers of the photovoltaic and the wind turbine within the limits of each state can be respectively expressed as: Furthermore, the probability of any state m and y corresponding and holding can be obtained by integrating the probability density functions f(P PV ), f(P WT ) from the lower limit to the upper limit, and the probability distribution functions F(P PV ), F(P WT ), that is: In summary, by using the average output power and its corresponding probability of the photovoltaic and wind turbines in the above different states, N un uncertainty state combinations are generated, where N un = M × Y, and the steps for this combination are as follows: Construct the average output power scenario sets ψ PV and ψ WT for photovoltaic and wind turbines respectively, i.e.: Use the Cartesian product of the two sets shown in Equation (22) to pair and group the average output powers of photovoltaic and wind turbines: Calculate the probabilities corresponding to each uncertain state combination:

7. A fault recovery method for an AC / DC hybrid distribution network based on network reconstruction and chance-constrained, as claimed in claim 1, wherein: The process of transformation and solution of the chance constraint model in Step 5 is: The unified form of the chance constraint can be expressed as: In the formula: F(x, ζ) > 0 represents the constraint condition; x and ζ represent ordinary variables and random variables respectively; Perform SAA approximation transformation on Equation (25) to obtain: where: E(F(x, ζ i )) is an indicator function for judging whether the constraint condition holds. It takes 1 when the constraint condition holds, and 0 otherwise; ζ i is a certain sample obtained by sampling; γ represents the confidence level when the condition holds for the average approximation of the sample; According to the SAA method, transform the chance constraints of the distribution network fault recovery models shown in Equations (13) and (14) respectively into: The indicator function E(·) in the above four equations of equations (28)-(31) is still a non-convex function and cannot be directly solved. Therefore, binary variables are introduced and The sum of the indicator functions is expressed as the sum of the binary variables of all samples in different scenarios. The above process is described by equations (32)-(35), that is: In Equations (32) and (33): K is an extremely large positive number, which can always ensure that when and take 1, Equations (32) and (33) hold absolutely; when or take 0, it means that the corresponding l-th sampling state is taken into account, that is, the determined fault recovery scheme can ensure the power balance and safe operation of the system in state l; when or take 1, it means that state l is not taken into account, that is, the determined fault recovery scheme cannot ensure the power balance and safe operation of the system in state l.

Images