A power distribution network restoration method and device considering switch uncertainty

Abstract

The application discloses a power distribution network recovery method and device considering switch uncertainty, a three-layer robust post-disaster recovery model based on a linearized Distflow model and a traveling salesman model, considers the failure of elements and remote switches in the process of post-disaster recovery, considers distributed generator rescheduling and network reconstruction technology, solves the repair personnel scheduling problem under the condition of uncertain fault type, and obtains a power distribution network recovery scheme applicable to all possible fault scenarios, reduces the maximum load loss that the power distribution network may suffer under the condition of incomplete fault information, in addition, the model only needs the power distribution network operator to predict the maximum number of various types of faults according to historical data, reduces the required information, effectively reduces the added scene and lightens the calculation burden, can find a global optimal solution in a short time, improves the efficiency of post-disaster recovery of the power distribution network, and provides an efficient tool for power distribution network recovery under the condition of incomplete fault information.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of power grid regulation and control, and particularly relates to a distribution network recovery method and device considering switch uncertainty. BACKGROUND

[0002] Extreme weather has caused many power outages, and therefore, improving the ability of the distribution system to withstand natural disasters is crucial for power supply reliability.

[0003] Post-disaster rapid recovery is one of the important features of a resilient distribution network, and reasonable arrangement of recovery strategies can play the role of various resources in the distribution network and greatly reduce the impact of disasters on the distribution network. Network reconfiguration plays an important role in the recovery process, and scholars at home and abroad have done a lot of research on the application of network reconfiguration in post-disaster recovery. On the one hand, network reconfiguration is used to build defensive islands to achieve rapid isolation of faults and improve the ability of the distribution network to withstand disasters. Some research takes into account the difference in operation between manual switches and remote control switches, and proposes a two-stage network reconfiguration method combining planning and operation. The first stage decides the installation position of various switches in advance, and the second stage formulates a network reconfiguration scheme according to the disaster situation of the distribution network to reduce the load loss of the distribution network. On the other hand, network reconfiguration is also used in cooperation with distributed generators, mobile power supply vehicles and repair teams to redivide the distribution network into several microgrids powered by distributed generators to restore critical loads. In these studies, the distribution system operator changes the system topology by installing remote switches in the system, but the post-disaster state of the remote switch is not concerned. In the event of a natural disaster, distribution network components and remote switches may fail within a short period of time, resulting in widespread power outages. The state of the remote switch directly affects the network reconfiguration strategy and thus the recovery process of the distribution network, so the failure of the remote switch needs to be taken into account in post-disaster recovery. Unlike conventional component damage, the failure of the remote switch may exist in the communication equipment or the actuator mechanism. The failure of the actuator mechanism can be regarded as a conventional component failure, such as line damage, but the failure of the communication equipment will cause the action of the remote switch to be uncontrollable and the state to be unobservable. After the communication is interrupted, the remote switch cannot receive the commands sent by the control center, resulting in uncontrollable action of the remote switch. In addition, after the communication is interrupted, the monitoring data of the remote switch cannot be uploaded to the data center, so that the failure type of the switch is unknown. Most existing research assumes that the fault location and fault type are known, and few studies focus on the recovery of the distribution network under incomplete fault information.

[0004] Although the power distribution network operator can learn the fault location through the fault indicator and fault location system, the uncertainty of fault type and repair time greatly increases the complexity of repair personnel scheduling, and how to make a recovery plan suitable for all possible fault scenarios is a problem that the current power distribution network urgently needs to solve. In the face of uncertainty in the recovery process, stochastic programming is often used for decision-making, however, stochastic programming has some practical limitations: 1) due to the small probability of extreme events, the power distribution network is difficult to collect enough data to obtain the accurate probability distribution of all random variables; 2) a large number of scenarios result in high computational burden. How to break through these limitations is the challenge currently faced. SUMMARY

[0005] The purpose of the present application is to provide a power distribution network recovery method and device considering switch uncertainty, to overcome the problem of inaccurate post-disaster recovery of the existing power distribution network under incomplete fault information.

[0006] A power distribution network recovery method considering switch uncertainty, comprising the following steps:

[0007] S1, using a linearized Distflow model to model the operation of the power distribution network to obtain a power distribution network operation model, using 0 / 1 variables to describe the state of the main network elements in the power distribution network operation model;

[0008] S2, using a traveling salesman model to establish a power distribution network repair team dispatching model;

[0009] S3, based on the power distribution network operation model and the power distribution network repair team dispatching model, a three-layer robust post-disaster recovery model is established;

[0010] S4, using the nested column and constraint generation method to decompose the three-layer robust post-disaster recovery model into an upper problem, a lower main problem and a lower sub-problem, using a commercial solver to iteratively solve the three problems, with the minimum system load loss caused by the disaster as the objective function, to obtain the optimal power distribution network recovery scheme.

[0011] Preferably, the linearized DistFlow model is used to model the operation of the power distribution network to obtain a power distribution network operation model as follows:

[0012]

[0013]

[0014]

[0015]

[0016]

[0017]

[0018]

[0019]

[0020]

[0021]

[0022]

[0023] where σ(j) is the set of lines with node j as the starting point, δ(j) is the set of lines with node j as the ending point, P l,t and Q l,t are the active power and reactive power flowing through line l at time t, respectively, and are the active power and reactive power output of the generator at node j at time t, respectively, and represent the active load and reactive load at node j, respectively, and are the active power curtailment and reactive power curtailment at node j at time t, V i,t is the voltage at node i at time t, r l and X l represent the resistance and reactance of line l, respectively, V0 is the standard voltage of the system, M is a constant, q l,t is the state of line l at time t, which is a 0 / 1 variable; N L is the set of all lines, V min,j and V max,j are the minimum voltage and maximum voltage allowed at node j, respectively, and are the maximum active power and maximum reactive power output of the generator at node j, is the capacity of line l, pf j is the power factor at node j.

[0024] Preferably, the radial distribution network needs to remain radial during operation, and the distribution network topology constraint is set as follows:

[0025]

[0026]

[0027]

[0028]

[0029] where n b is the number of nodes, the 0 / 1 variable ξ j,t is 1 if node j becomes a root node and 0 otherwise, f l,t is the virtual flow passing through line l at time t, and R is the set of potential root nodes.

[0030] 0 / 1 variable x and y to describe the path selection of the repair team:

[0031]

[0032]

[0033] Preferably, the specific expression of the repair team dispatch model is as follows:

[0034]

[0035]

[0036]

[0037]

[0038]

[0039]

[0040]

[0041]

[0042]

[0043]

[0044]

[0045]

[0046]

[0047]

[0048]

[0049]

[0050]

[0051]

[0052]

[0053]

[0054]

[0055] where N F is the set of all failures and repair stations, dp represents a repair station, N C is the set of repair crews, is the time when repair crew c arrives at failure m, is the time when repair crew c repairs element failure m, is the travel time of repair crew from failure m to failure n, N FS is the set of telecontrol switch failures, is the repair time of telecontrol switch failure I, where failure I is the communication device failure of the telecontrol switch and the corresponding repair process is repair I, is the repair time of telecontrol switch failure II, where failure II is the communication device and actuating mechanism failure of the telecontrol switch and the corresponding repair process is repair II, 0 / 1 variable λ I,m,c is 1 if repair crew c performs repair I at telecontrol switch failure m, otherwise it is 0, 0 / 1 variable λ II,m,c is 1 if repair crew c performs repair II at telecontrol switch failure m, otherwise it is 0, z m,t is an auxiliary variable used to record the time when repair crew arrives at failure m, τ m,t , μ m,t is an auxiliary variable used to record the repair time of failure m, N S is the set of all telecontrol switches, a I,m is a 0 / 1 variable, which is 1 if telecontrol switch m has failure I, otherwise it is 0, a II,m is a 0 / 1 variable, which is 1 if telecontrol switch m has failure II, otherwise it is 0, is a 0 / 1 variable representing whether telecontrol switch I is actuated at time t, which is 1 if the switch is actuated, otherwise it is 0.

[0056] Preferably, in the first layer model of the three-layer robust post-disaster recovery model, the distribution network operator formulates a dispatch plan for the repair team for all scenarios in the set of fault scenarios, determines the repair sequence of each damaged element and damaged remote switch, and minimizes the loss of load of the distribution network; the second layer model of the three-layer robust post-disaster recovery model considers the network reconfiguration strategy of the distribution network operator and the dispatch plan formulated by the first layer model to find the worst fault scenario to maximize the loss of load of the distribution network; the third layer model of the three-layer robust post-disaster recovery model formulates a distributed generator dispatch scheme and a network reconfiguration strategy according to the dispatch plan formulated by the first layer model under the fault scenario in which the loss of load of the distribution network is maximum calculated by the second layer model, and minimizes the loss of load of the distribution network.

[0057] Preferably, the three problems are solved iteratively using a commercial solver, and the specific steps are as follows:

[0058] a. Set the upper bound UB = +∞, the lower bound LB = -∞, the upper layer iteration number k = 1, and an arbitrary fault scenario is given;

[0059] b. Solve the upper layer problem to obtain the dispatch scheme of the repair team, the repair time x * 、 z * 、τ * 、μ * and the objective function β1, update LB = max{LB, β1}, and pass the results z * 、τ * 、μ * to the lower layer problem;

[0060] c. Set LUB = +∞, LLB = -∞, and set the lower layer iteration number m = 1;

[0061] d. Solve the lower layer main problem to obtain the fault combination and the objective function value β2, update LUB = min{LUB, β2}, and pass and to the lower layer sub-problem;

[0062] e. Solve the lower layer sub-problem to obtain the switch action q ch,* and the objective function value update m = m + 1;

[0063] f. If LUB ≠ LLB, pass q ch,* to the lower layer main problem, and return to step 4; if LUB = LLB, pass the fault combination and Add to the upper layer of the failure scenario set, k=k+1, update UB=min{UB, LUB};

[0064] If UB≠LB, return to step b, if UB=LB, output the maintenance team scheduling scheme x * .

[0065] A power distribution network restoration system considering switch uncertainty, comprising an initialization module and a power distribution network restoration model;

[0066] The initialization module is configured to model the operation of the power distribution network to obtain an operation model of the power distribution network, use 0 / 1 variables to describe the state of main network elements in the operation model of the power distribution network, use a traveling salesman model to establish a power distribution network maintenance team dispatching model, and establish a three-layer robust post-disaster restoration model based on the operation model of the power distribution network and the power distribution network maintenance team dispatching model;

[0067] The power distribution network restoration model is configured to use a nested column and constraint generation method to decompose the three-layer robust post-disaster restoration model into an upper layer problem, a lower layer main problem and a lower layer sub-problem, use a commercial solver to iteratively solve the three problems, and obtain an optimal power distribution network restoration scheme with the minimum system load loss caused by disasters as an objective function.

[0068] Preferably, the operation of the power distribution network is modeled using a linearized DistFlow model, and the obtained operation model of the power distribution network is as follows:

[0069]

[0070]

[0071]

[0072]

[0073]

[0074]

[0075]

[0076]

[0077]

[0078]

[0079]

[0080] where σ(j) is the set of lines with node j as the starting point, δ(j) is the set of lines with node j as the ending point, P l,t and Q l,t are the active power and reactive power flowing through line l at time t, respectively, and are the active power and reactive power output of the generator at node j at time t, respectively, and are the active load and reactive load at node j, respectively, and are the active power and reactive power curtailment at node j at time t, V i,t is the voltage at node i at time t, r l and X l are the resistance and reactance of line l, respectively, V0 is the standard voltage of the system, M is a constant, q l,t is the state of line l at time t, which is a 0 / 1 variable; N L is the set of all lines, V min,j and V max,j are the minimum voltage and maximum voltage allowed at node j, respectively, and are the maximum active power and maximum reactive power output of the generator at node j, is the capacity of line l, pf j is the power factor at node j.

[0081] The three problems are solved iteratively using a commercial solver, and the specific steps are as follows:

[0082] a. Set the upper bound UB = +∞, the lower bound LB = -∞, the upper iteration number k = 1, and give an arbitrary fault scenario;

[0083] b. Solve the upper problem to obtain the dispatching scheme of the repair team, the repair time x * , z * , τ * , μ * , and the objective function β1, update LB = max{LB, β1}, and pass the results z * , τ * , μ * to the lower problem;

[0084] c. Set LUB = +∞, LLB = -∞, and set the lower iteration number m = 1;

[0085] d. Solve the lower main problem to obtain the fault combination and the objective function value β2, update LUB = min{LUB, β2}, and pass and to the lower sub-problem;

[0086] e, solve the lower sub-problem to obtain the switch action q ch,* and the objective function value update

[0087] f, if LUB ≠ LLB, pass q ch,* to the lower main problem, return to step 4, and if LUB = LLB, add the fault combination and to the upper fault scenario set, k = k + 1, and update UB = min{UB, LUB};

[0088] if UB ≠ LB, return to step b, and if UB = LB, output the maintenance team scheduling scheme x * .

[0089] Compared with the prior art, the present application has the following beneficial technical effects:

[0090] The power distribution network restoration method considering switch uncertainty provided by the present application is a three-layer robust post-disaster restoration model based on a linearized Distflow model and a traveling salesman model, considers the faults of elements and remote switches in the process of post-disaster restoration, considers distributed generator rescheduling and network reconstruction technology, solves the maintenance personnel scheduling problem under the condition of uncertain fault types, and obtains a power distribution network restoration scheme applicable to all possible fault scenarios, reduces the maximum load loss that the power distribution network may suffer under the condition of incomplete fault information, and in addition, the model only needs the power distribution network operator to predict the maximum number of various types of faults according to historical data, thereby reducing the required information requirements.

[0091] The present application uses nested columns and constraint generation algorithms to iteratively solve the established robust optimization model, effectively reduces the added scenarios and reduces the computational burden, can find a global optimal solution in a short time, improves the efficiency of post-disaster restoration of the power distribution network, and provides an efficient tool for the restoration of the power distribution network under the condition of incomplete fault information. BRIEF DESCRIPTION OF DRAWINGS

[0092] Figure 1 The constraints contained in the three-layer robust post-disaster restoration model.

[0093] Figure 2 The remote switch state diagram under different fault conditions.

[0094] Figure 3 The composition of the decomposed problem.

[0095] Figure 4 Solve the model flow chart. DETAILED DESCRIPTION

[0096] In order to make the person skilled in the art better understand the present application, the technical solutions in the embodiments of the present application will be described clearly and completely in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor should belong to the scope of protection of the present application.

[0097] It should be noted that the terms "first", "second" and the like in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and do not necessarily indicate a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the application described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device including a series of steps or units does not necessarily limit to those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0098] The power distribution network restoration method considering switch uncertainty of the present application considers the uncertainty of the fault and fault type of the power distribution network elements and remote switch, takes the minimization of the maximum system load loss as the objective function, establishes a three-layer robust post-disaster restoration model based on the linearized Distflow model and the Traveling Salesman Model, and obtains the most robust power distribution network restoration scheme under the condition of incomplete fault information; Specifically, the following steps are included:

[0099] S1, using the linearized Distflow model to model the operation of the power distribution network to obtain the operation model of the power distribution network, using 0 / 1 variable to describe the state of the main network element in the operation model of the power distribution network;

[0100] S2, using the Traveling Salesman Model to establish a power distribution network repair team dispatching model;

[0101] S3, based on the operation model of the distribution network and the distribution network maintenance team dispatching model, a three-layer robust post-disaster recovery model is established, in the first layer model of the three-layer robust post-disaster recovery model, the distribution network operator formulates a maintenance team scheduling plan for all scenarios in the fault scenario set, determines the repair sequence of each damaged element and damaged remote switch, and minimizes the loss of load of the distribution network; the second layer model of the three-layer robust post-disaster recovery model considers the network reconfiguration strategy of the distribution network operator and the scheduling plan formulated in the first layer model to find the worst fault scenario to maximize the loss of load of the distribution network; in the third layer model of the three-layer robust post-disaster recovery model, under the fault scenario in which the loss of load of the distribution network is maximum calculated in the second layer model, a distributed generator scheduling scheme and a network reconfiguration strategy are formulated according to the scheduling plan formulated in the first layer model, and the loss of load of the distribution network is minimized;

[0102] S4, the three-layer robust post-disaster recovery model is decomposed into an upper layer problem, a lower layer main problem and a lower layer sub-problem using a nested column and constraint generation method, and a commercial solver is used to iteratively solve the three problems, so that the optimal distribution network recovery scheme can be obtained with the minimum loss of system load caused by disasters as the objective function.

[0103] In S1, the present application is directed to a radial distribution network, a linearized DistFlow model is used to model the operation of the distribution network, and the obtained operation model of the distribution network is as follows:

[0104]

[0105]

[0106]

[0107]

[0108]

[0109]

[0110]

[0111]

[0112]

[0113]

[0114]

[0115] Wherein, σ(j) is a line set with node j as the starting point, δ(j) is a line set with node j as the ending point, P l,t and Ql,t are the active power and reactive power flowing through line l at time t, and are the active and reactive outputs of the generator at node j at time t, and represent the active load and reactive load at node j respectively, and is the active power reduction and reactive power reduction at node j at time t, V i,t represents the voltage at node i at time t, r l and X l Represent the resistance and reactance of line l respectively, V0 is the standard voltage of the system, M is a constant, q l,t is the state of line l at time t, which is a 0 / 1 variable. If it is 0, it means the line is out of operation, and if it is 1, it means the line is operating normally. L is the set of all circuits, V min,j and V max,j are the minimum and maximum voltages allowed at node j, respectively. and is the maximum active output and maximum reactive output of the generator at node j, is the capacity of line l, pf j is the power factor at node j; constraints (2) and (3) are active power and reactive power balance constraints, ensuring that the net injected power of the node is 0; constraints (4) and (5) represent the relationship between the voltage at both ends of the line and the active and reactive power flowing through the line; constraint (6) ensures that the node voltage will not exceed the limit, ensuring the voltage safety of the distribution network; constraints (7) and (8) limit the active and reactive output of the generator; constraints (9) and (10) ensure that the active and reactive power flowing through the line will not exceed the capacity of the line; constraint (11) ensures that the load shedding amount of node j will not exceed the original load amount of node j; constraint (12) ensures that the power factor of each node remains unchanged.

[0116] The distribution network maintenance team dispatch model is essentially a maintenance team dispatch problem, that is, a vehicle routing problem. This paper uses the traveling salesman model to model it and introduces 0 / 1 variables. x and y to describe the path selection of the maintenance team:

[0117]

[0118]

[0119] The specific expression of the maintenance team dispatch model is as follows:

[0120]

[0121]

[0122]

[0123]

[0124]

[0125]

[0126]

[0127]

[0128]

[0129]

[0130]

[0131]

[0132]

[0133]

[0134]

[0135]

[0136]

[0137]

[0138]

[0139]

[0140]

[0141]

[0142] where N F is the set of all failures and repair stations, dp represents a repair station, N C is the set of repair crews, is the time at which repair crew c arrives at failure m, is the time required for repair crew c to repair element failure m, is the travel time for repair crew to travel from failure m to failure n, N FS is the set of remote switch failures, is the repair time needed after failure I of the remote switch, failure I is the failure of the communication device of the remote switch, and its corresponding repair process is repair I, is the repair time needed after failure II of the remote switch, failure II is the failure of both the communication device and the action mechanism of the remote switch, and its repair process is repair II, 0 / 1 variable λ I,m,c is 1 if repair team c performs repair I at failure m of the remote switch, otherwise it is 0, 0 / 1 variable λ II,m,c is 1 if repair team c performs repair II at failure m of the remote switch, otherwise it is 0, z m,t is an auxiliary variable used to record the time when the repair team arrives at failure m, τ m,t , μ m,t is an auxiliary variable used to record the repair time of failure m, N S is the set of all remote switches, a I,m is a 0 / 1 variable, which is 1 if failure I occurs in remote switch m, otherwise it is 0, a II,m is a 0 / 1 variable, which is 1 if failure II occurs in remote switch m, otherwise it is 0, is a 0 / 1 variable representing whether the remote switch l acts at time t, if it is 1, it means that the switch acts, otherwise it is 0. Constraint (15) indicates that the repair team must start from the repair station, constraint (16) ensures that the repair team will leave the failure point after repairing the failure, constraint (17) ensures that the repair team will return to the repair station after the entire recovery process is completed, constraints (18) and (19) ensure that each failure point must be repaired, constraints (20) and (21) are used to calculate the time when the repair team arrives at each failure point, constraint (22) ensures that the repair team will only perform one repair process at each failure switch, constraints (23)-(26) calculate the time when the repair team arrives at the failure, constraints (27)-(31) are used to calculate the repair status of each failure at each time, constraint (32) indicates that the failed component will be put into operation immediately after repair, constraint (33) indicates that the undamaged line is always in normal operation state, constraints (34)-(35) describe the relationship between the repair process selection of the repair team and the state of the remote switch, as shown in detail in Figure 2 .

[0143] The radial distribution network needs to maintain the radial state during operation. A radial graph must satisfy two conditions at the same time: (1) the number of edges in the graph must be equal to the number of nodes minus the number of subgraphs; (2) each subgraph must be connected; therefore, the distribution network topology constraints are set as follows:

[0144]

[0145]

[0146]

[0147]

[0148] where n b is the number of nodes, the 0 / 1 variable ξ j,t is 1 if node j becomes a root node and 0 otherwise, f l,t is the virtual flow passing through line l at time t, and R is the set of potential root nodes. Constraints (36) satisfy condition (1), and constraints (37)-(39) guarantee the connectivity of each subgraph of the distribution network, satisfying condition (2).

[0149] The robust optimization model established by the present application is a three-layer mixed integer linear programming model, which cannot be directly solved by existing commercial solvers. In addition, the large number of possible fault scenarios will increase the computational burden, therefore, the present application uses a nested column and constraint generation algorithm to decompose the established model into an upper level problem (UP), a lower level master problem (LMP), and a lower level sub-problem (LSP) for iterative solving, reducing the number of added scenarios. Each problem is as shown in Figure 3

[0150] Upper level problem: the distribution network operator dispatches the repair team to repair the fault, minimizes the load loss of the system, and transmits the optimal repair team dispatching scheme to the lower level problem;

[0151] Lower level master problem: given the repair team dispatching scheme, the lower level master problem solves the worst switch fault scenario, maximizes the system load loss, and transmits the obtained fault scenario to the lower level sub-problem;

[0152] Lower level sub-problem: based on the repair team dispatching plan provided by the upper level problem and the fault scenario provided by the lower level master problem, the lower level sub-problem dispatches the distributed generator and formulates a network reconfiguration strategy to minimize the load loss.

[0153] The specific mathematical forms of each problem are as follows:

[0154] Upper level problem:

[0155] Obj: min β1

[0156] s.t.

[0157] Equations (15)-(31)

[0158] ​

[0159]

[0160]

[0161]

[0162]

[0163]

[0164]

[0165]

[0166]

[0167]

[0168]

[0169]

[0170]

[0171]

[0172]

[0173] where β1is the objective function value of the upper problem, the superscript k represents the number of iterations of the outer column and constraint generation algorithm, and the symbol "^" represents that the variable is a constant value.

[0174] Lower main problem:

[0175] According to the strong duality theory of linear problems, the lower main problem is rewritten in the following form:

[0176] Obj: max β2

[0177]

[0178]

[0179]

[0180]

[0181]

[0182]

[0183]

[0184]

[0185]

[0186]

[0187]

[0188]

[0189]

[0190]

[0191]

[0192] where β2is the objective function value of the lower-level master problem, subscript m is the iteration number of the inner column and constraint generation algorithm, η1-η 13 are the dual variables of constraints (2)-(12), constraint (55) indicates that only one type of failure can occur for a failed switch, constraint (56) limits the maximum number of failures II. Constraints (57)-(65) are the specific forms of the dual problem.

[0193] Lower-level sub-problem

[0194] Obj:

[0195] s.t.(2)-(12)(70)

[0196]

[0197]

[0198]

[0199]

[0200] The specific process of iterative solution of the three problems using a commercial solver is shown in Figure 4 , and the solution steps are as follows:

[0201] 1) Set the upper bound UB = +∞, LB = -∞, the upper-level iteration number k = 1, and give an arbitrary failure scenario;

[0202] 2) Solve the upper-level problem to obtain the dispatching scheme of the repair team, the repair time x * , z * , τ * , μ* and the objective function β1, update LB=max{LB, β1}, the result z * , τ * , μ * is passed to the lower problem;

[0203] 3) Set LUB=+∞, LLB=-∞, set the number of lower iteration m=1;

[0204] 4) Solve the lower main problem, obtain the fault combination and the objective function value β2, update LUB=min{LUB, β2}, the result and is passed to the lower sub-problem;

[0205] 5) Solve the lower sub-problem, obtain the switch action q ch,* and the objective function value update m=m+1;

[0206] 6) If LUB≠LLB, pass q ch,* to the lower main problem, return to step 4, if LUB=LLB, add the fault combination and to the upper fault scene set, k=k+1, update UB=min{UB, LUB};

[0207] If UB≠LB, return to step 2), if UB=LB, output the maintenance team scheduling scheme x * .

[0208] The three-layer robust post-disaster recovery model is proposed based on the linearized Distflow model and the traveling salesman model, the faults of elements and remote switches are considered in the process of post-disaster recovery, the distributed generator rescheduling and network reconstruction technology are considered, the difficult problem of maintenance personnel scheduling under the condition of uncertain fault type is solved, the obtained power distribution network recovery scheme is suitable for all possible fault scenes, the maximum load loss of the power distribution network under the condition of incomplete fault information is reduced, in addition, the model only needs the power distribution network operator to predict the maximum number of various types of faults according to historical data, the requirement of the required information is reduced;

[0209] The established robust optimization model is iteratively solved by using the nested column and constraint generation algorithm, the required added scene is effectively reduced, the calculation burden is reduced, the global optimal solution can be found in a short time, the efficiency of post-disaster recovery of the power distribution network is improved, and an efficient tool is provided for the recovery of the power distribution network under the condition of incomplete fault information.

Claims

1. A distribution network restoration method considering switch uncertainty, characterized in that: The following steps are involved: S1, use the linearized Distflow model to model the operation of the distribution network to obtain the distribution network operation model, and use 0 / 1 variables to describe the status of the main network components in the distribution network operation model; S2, using the traveling salesman model to establish a distribution network maintenance team dispatch model; S3, a three-layer robust post-disaster recovery model is established based on the distribution network operation model and the distribution network maintenance team dispatch model; S4, using nested columns and constraint generation methods, decomposes the three-layer robust post-disaster recovery model into an upper-layer problem, a lower-layer main problem, and lower-layer sub-problems. A commercial solver is used to iteratively solve the three problems, with the objective function of minimizing the maximum system load loss that may be caused by the disaster, to obtain the optimal distribution network restoration plan; In the first layer of the three-layer robust disaster recovery model, the distribution network operator develops a maintenance team scheduling plan for all scenarios in the fault scenario set, determines the repair sequence for each damaged component and damaged teleswitch, and minimizes the load loss of the distribution network. The second layer of the three-layer robust disaster recovery model considers the network reconstruction strategy of the distribution network operator and the dispatch plan formulated by the first layer model to find the worst fault scenario to maximize the load loss of the distribution network. The third layer of the three-layer robust disaster recovery model formulates a distributed generator dispatch plan and network reconstruction strategy based on the dispatch plan formulated by the first layer model under the fault scenario with the maximum load loss of the distribution network calculated by the second layer model, so as to minimize the load loss of the distribution network. The specific steps for iteratively solving the three problems using a commercial solver are as follows: a. Set the upper bound UB = +∞, LB = -∞, the number of upper layer iterations k = 1, and give an arbitrary fault scenario; b. Solve the upper-level problem to obtain the maintenance team scheduling plan and the repair time x for each fault * 、 z * , τ * 、μ * And the objective function β1, update LB = max{LB, β1}, and the result z * , τ * 、μ * Passing on the problem to the lower level; c. Set LUB = +∞, LLB = -∞, and set the number of lower layer iterations m = 1; d. Solve the lower-level main problem and obtain the fault combination And the objective function value β2, update LUB = min{LUB, β2}, and set a I * and Pass to the lower level sub-problem; e. Solve the lower-level subproblem and obtain the switch action q ch,* And the objective function value renew f, if LUB≠LLB, set q ch,* Pass to the lower master problem, return to step 4, if LUB = LLB, the fault combination and Add to the upper layer fault scenario set, k = k + 1, update UB = min{UB, LUB}; If UB≠LB, return to step b. If UB=LB, output the maintenance team scheduling plan x. * .

2. A distribution network restoration method considering switch uncertainty according to claim 1, characterized in that: The linearized DistFlow model is used to model the operation of the distribution network. The resulting distribution network operation model is as follows: Among them, σ(j) is the set of lines starting from node j, δ(j) is the set of lines ending at node j, and P l,t and Q l,t are the active power and reactive power flowing through line l at time t, and are the active and reactive outputs of the generator at node j at time t, and represent the active load and reactive load at node j respectively, and is the active power reduction and reactive power reduction at node j at time t, V i,t represents the voltage at node i at time t, r l and X l Represent the resistance and reactance of line l respectively, V0 is the standard voltage of the system, M is a constant, q l,t is the state of line l at time t, which is a 0 / 1 variable; N L is the set of all circuits, V min,j and V max,j are the minimum and maximum voltages allowed at node j, respectively. and is the maximum active output and maximum reactive output of the generator at node j, is the capacity of line l, pf j is the power factor at node j.

3. A distribution network restoration method considering switch uncertainty according to claim 2, characterized in that: The radial distribution network needs to maintain its radial shape during operation. The distribution network topology constraints are set as follows: Among them, n b is the number of nodes, if node j becomes the root node, then the 0 / 1 variable ξ j,t is 1, otherwise it is 0, f l,t is the virtual traffic flowing on line l at time t, and R is the set of potential root nodes.

4. A distribution network restoration method considering switch uncertainty according to claim 1, characterized in that: 0 / 1 variables x and y to describe the path selection of the maintenance team:

5. A distribution network restoration method considering switch uncertainty according to claim 4, characterized in that: The specific expression of the maintenance team dispatch model is as follows: where N F is the set of all faults and repair stations, dp represents the repair station, N C For the maintenance team to gather, is the time it takes for the maintenance team c to arrive at the fault m, The time required for maintenance team c to repair component failure m, N is the travel time of the maintenance team from fault m to fault n, FS is the set of remote switch faults, is the time required to repair the remote switch after fault I occurs. Fault I is a failure of the communication device of the remote switch, and its corresponding repair process is maintenance I. is the time required to repair the remote switch after Fault II occurs. Fault II is the simultaneous failure of the remote switch communication device and the action mechanism. Its repair process is Maintenance II. 0 / 1 variable λ I,m,c If it is 1, it means that the maintenance team c is performing maintenance I at the remote switch fault m, otherwise it is 0, and the 0 / 1 variable λ II,m,c If it is 1, it means that the maintenance team c is performing maintenance II at the remote switch fault m, otherwise it is 0. z m,t is an auxiliary variable used to record the time when the maintenance team arrives at fault m, τ m,t 、μ m,t is an auxiliary variable used to record the repair time of fault m, N S is the set of all remote switches, a I,m It is a 0 / 1 variable. If the remote switch m fails, it is 1. Otherwise, it is 0. II,m It is a 0 / 1 variable. If the remote switch m fails II, it is 1, otherwise it is 0. The 0 / 1 variable represents whether the remote switch l is activated at time t. If it is 1, it means the switch is activated, otherwise it is 0.

6. A distribution network restoration system considering switching uncertainty based on the distribution network restoration method considering switching uncertainty according to claim 1, characterized in that: Includes initialization module and distribution network restoration model; An initialization module is used to model the operation of the distribution network and obtain an operation model of the distribution network, using 0 / 1 variables to describe the status of the main network components in the operation model of the distribution network; A distribution network maintenance team dispatch model is established using the traveling salesman model; a three-layer robust post-disaster recovery model is established based on the distribution network operation model and the distribution network maintenance team dispatch model; The distribution network restoration model uses nested columns and constraint generation methods to decompose the three-layer robust post-disaster recovery model into an upper-layer problem, a lower-layer main problem, and a lower-layer sub-problem. The three problems are iteratively solved using a commercial solver, and the objective function is to minimize the maximum system load loss that may be caused by the disaster, thereby obtaining the optimal distribution network restoration plan.

7. A distribution network restoration system considering switch uncertainty according to claim 6, characterized in that: The linearized DistFlow model is used to model the operation of the distribution network. The resulting distribution network operation model is as follows: Among them, σ(j) is the set of lines starting from node j, δ(j) is the set of lines ending at node j, and P l,t and Q l,t are the active power and reactive power flowing through line l at time t, and are the active and reactive outputs of the generator at node j at time t, and represent the active load and reactive load at node j respectively, and is the active power reduction and reactive power reduction at node j at time t, V i,t represents the voltage at node i at time t, r l and X l Represent the resistance and reactance of line l respectively, V0 is the standard voltage of the system, M is a constant, q l,t is the state of line l at time t, which is a 0 / 1 variable; N L is the set of all circuits, V min,j and V max,j are the minimum and maximum voltages allowed at node j, respectively. and is the maximum active output and maximum reactive output of the generator at node j, is the capacity of line l, pf j is the power factor at node j.

8. The distribution network restoration system considering switch uncertainty according to claim 6, characterized in that: The specific steps for iteratively solving the three problems using a commercial solver are as follows: a. Set the upper bound UB = +∞, LB = -∞, the number of upper layer iterations k = 1, and give an arbitrary fault scenario; b. Solve the upper-level problem to obtain the maintenance team scheduling plan and the repair time x for each fault * 、 z * , τ * 、μ * And the objective function β1, update LB = max{LB, β1}, and the result z * , τ * 、μ * Passing on the problem to the lower level; c. Set LUB = +∞, LLB = -∞, and set the number of lower layer iterations m = 1; d. Solve the lower-level main problem and obtain the fault combination And the objective function value β2, update LUB = min{LUB, β2}, and Pass to the lower level sub-problem; e. Solve the lower-level subproblem and obtain the switch action q ch,* And the objective function value renew f, if LUB≠LLB, set q ch,* Pass to the lower master problem, return to step 4, if LUB = LLB, the fault combination and Add to the upper layer fault scenario set, k = k + 1, update UB = min{UB, LUB}; If UB≠LB, return to step b. If UB=LB, output the maintenance team scheduling plan x. * .

Images