Power distribution network fault recovery method based on multi-stage stochastic programming

Abstract

The invention provides a power distribution network fault recovery method based on multi-stage stochastic programming, and the method comprises the steps: generating a fault recovery decision through a target function of a fault recovery strategy, and a multi-period fault recovery model of maintenance strategy constraint, network reconstruction and ad hoc network coordination constraint in the fault recovery strategy; and in combination with a random variable uncertainty implementation condition, implementing a targeted fault recovery action decision. According to the method, load uncertainty and renewable resource uncertainty in a multi-period network recovery process are considered, multi-type flexible resources such as fixed and movable resources, a supply side, a demand side and a grid structure in a network are cooperatively scheduled based on a multi-stage stochastic programming method, the power failure loss is reduced, and the fault recovery speed is increased; the situation that the recovery strategy is not feasible due to inaccurate prediction or measurement of the weak power distribution network is avoided, and a more reliable and more robust fault recovery strategy is provided for the weak power distribution network.

Description

Technical Field

[0001] The present invention relates to the technical field of distribution network fault restoration, and in particular to a distribution network fault restoration method based on multi-stage random programming. Background Art

[0002] As the terminal link in the power system, the stable operation of the distribution network is crucial to ensure the quality and reliability of power supply. As the penetration rate of distributed resources, especially distributed renewable resources, in the distribution network continues to increase, how to effectively use these resources to restore distribution network faults has become a research hotspot.

[0003] However, there are many difficulties in formulating effective and reliable fault recovery strategies. The operating characteristics of various types of distributed resources vary greatly, the emergency power supply capabilities vary greatly, the reconstruction capabilities of weak distribution networks are limited, the perception and prediction capabilities of loads and renewable resources are insufficient, and technologies such as multi-period network reconstruction, partitioned self-organizing networks, and mobile flexible resources are difficult to effectively coordinate. These problems have limited the further improvement of the fault recovery capabilities of weak distribution networks. Summary of the invention

[0004] In order to solve the deficiencies mentioned in the above background technology, the purpose of the present invention is to provide a distribution network fault recovery method based on multi-stage stochastic programming, the basic purpose of which is to achieve fault recovery of weak distribution network considering multi-stage uncertainty.

[0005] During the scheme design process, grid parameters, equipment parameters, load and renewable resource prediction parameters are imported to build a mobile flexible resource operation and control model. According to the characteristics of various distributed flexible resources, a network reconstruction and fault recovery collaborative control model is constructed. Based on the multi-stage random planning method, various uncertainties in multi-period fault recovery are handled, thereby realizing the fault recovery of weak distribution networks considering multi-stage uncertainties.

[0006] The present invention takes into account the load uncertainty and renewable resource uncertainty in the multi-period network recovery process, and coordinates the fixed and mobile, supply-side and demand-side, grid structure and other types of flexible resources in the network based on a multi-stage random programming method to reduce power outage losses and improve fault recovery speed, and avoid the situation where the recovery strategy of the weak distribution network is not feasible due to inaccurate prediction or measurement. It coordinates multiple types and multi-dimensional flexible resources such as source-grid-load-storage, improves the recovery ability of the weak distribution network in response to faults, and provides a reliable multi-period fault recovery strategy in a high uncertainty environment caused by weak measurement and prediction capabilities, providing a more reliable and robust fault recovery means for the weak distribution network.

[0007] The present invention specifically adopts the following technical solutions:

[0008] A distribution network fault recovery method based on multi-stage stochastic programming is proposed. The basic design is to generate a fault recovery decision through a multi-period fault recovery model including the objective function of the fault recovery strategy and the maintenance strategy constraints, network reconstruction and self-organizing network coordination constraints in the fault recovery strategy. Combined with the realization of random variable uncertainty, targeted fault recovery action decisions are implemented.

[0009] Furthermore, the specific implementation process of the fault recovery decision is as follows: under the constraint of the total fault recovery time, consider T time stages: t = 1, 2, ..., T, and set the variable representing the uncertain scenario As the time progresses, it is gradually identified and the following distribution network fault recovery strategies are implemented:

[0010] The goal of the multi-period fault recovery model is to minimize the total amount of load abandonment and the total amount of network loss in the entire recovery process to ensure that the fault recovery strategy is optimal during the entire fault period. The expression of the maintenance strategy is achieved by constraining the opening and closing state of the fault line.

[0011] The dispatch compensation decision includes the coordinated dispatch decision of resources and the adjustment of distribution network state variables.

[0012] Furthermore, the variables representing the uncertain scenarios are generated by sampling the probability distribution of load demand and resource uncertainty.

[0013] Furthermore, the objective function of the multi-period fault recovery model based on multi-stage stochastic programming is:

[0014]

[0015] Among them, x, y, ξ are in multi-stage format, the superscript represents the stage number index, the subscript represents the scenario index, T is the upper limit of the time stage, and represents the total time limit for fault recovery; x represents the fault recovery strategy, and y represents the scheduling compensation decision.

[0016] Furthermore, the objective function expression of the multi-period fault recovery model is as follows:

[0017]

[0018] Where N T Indicates the total number of time steps within the fault period; N bus represents the total number of nodes in the distribution network; Δt represents the time step within the fault period; represents the load shedding rate of node i in time period t; represents the load demand power of node i in time period t; (i, j) represents the branch determined by nodes i and j; Ψ represents the vector set composed of the starting nodes and the end nodes of all branches; represents the branch current size with node i as the starting node and node j as the ending node in time period t, r i,j Indicates the resistance of the branch;

[0019] In the unified fault recovery model of network reconstruction and island partitioning, a series of constraints are added to the variables involving the switch state; during the duration of the fault, the model must ensure that the fault line remains disconnected; the expression of the maintenance strategy in the model is realized by constraining the opening and closing states of the fault line, as follows:

[0020]

[0021] Where F represents the vector set consisting of the start and end nodes of all faulty branches; h represents the maximum number of branches that can be inspected simultaneously within a given period of time t; T f Indicates the time required to repair a faulty branch; represents the switch state of branch (i, j) in the fault recovery period t. If the branch is disconnected, Otherwise, it is 1; in the initial state of the fault, the switch variables of the faulty branches are forced to be 0, and they are all in the disconnected state. During the fault recovery process, at most h faulty branches can be repaired in each period.

[0022] Furthermore, the unified fault recovery model includes a network operation model and a network reconstruction model taking into account flexible resources of the grid;

[0023] The network operation model taking into account the flexible resources of the grid includes node power balance constraints and branch voltage drop constraints, specifically:

[0024]

[0025]

[0026] In the formula, They represent the total active power injected into node j, the total active power outflowed, the size of the active load, and the active output of the diesel generator in time period t respectively; They represent the total reactive power injected into node j, the total reactive power outflowed, the size of reactive load, and the reactive output of the diesel generator in time period t respectively; They represent the sum of active line loss and reactive line loss injected into node j in time period t; (1+k)U N 、(1-k)U N Respectively represent the upper and lower limits of the voltage at node i; I maxrepresents the upper limit of the current amplitude on the branch (i, j); for the branch that is in the disconnected state during the period t, The value of is zero; M represents a positive constant;

[0027] The network reconstruction model is based on radial and connectivity constraints, and coordinates multi-type flexible resource operation scheduling models to perform partitioned self-organizing networking after failures, thereby minimizing failure losses and optimizing failure recovery strategies;

[0028] The radial and connectivity constraints are specifically:

[0029]

[0030]

[0031] Among them, the first formula represents the node power balance constraint for the unbalanced node without diesel generators, and does not consider the network loss in the virtual power flow, only focusing on the topological state of the network; the second formula represents the node power balance constraint for the balanced node, represents the power provided by the "source" bus in the virtual network during period t; the third and fourth formulas are virtual power flows Connection status with branch The two constraints in the last formula indicate that as the fault repair work proceeds, more lines can be gradually added, but the upper limit of the total number of lines |γ|-1 cannot be exceeded at any time; where, They represent the power transmitted on line js and line ij in time period t in the virtual network respectively; δ(j) and π(j) represent the set of child nodes and parent nodes of node j; γ represents the set of nodes in the network; {DG} represents the set of diesel generator nodes in the distribution network.

[0032] Furthermore, in the multi-type flexible resource operation scheduling model:

[0033] The output power constraints of gas turbines include: output power constraints, active output, reactive output and apparent power:

[0034]

[0035] in, The minimum prime mover output power considering the efficiency factor of the gas turbine; P MT is the actual output power of the gas turbine prime mover; is the rated output power of the gas turbine prime mover; s MT 0-1 variable indicating the start and stop status of the gas turbine; is the rated apparent power of the gas turbine; is the power generation efficiency of the gas turbine; is the actual thermal power output of the gas turbine; is the energy loss rate of the gas turbine; η HR The efficiency of the waste heat recovery device for the gas turbine; It is the active power of the gas turbine's actual output electrical energy; It is the reactive power of the gas turbine that actually outputs electrical energy;

[0036] The diesel generator operation model is as follows:

[0037] P DG_min ≤P DG ≤P DG_rated

[0038]

[0039] Where P DG The active power output of the diesel generator; The minimum active power limit is to consider the diesel generator economy and parts life; is the rated active power of the diesel generator; is the rated power factor of the diesel generator; Q DG The reactive power output of the diesel generator;

[0040] The operation model of the electric energy storage device ESS is:

[0041] SOC min ≤SOC t ≤SOC max

[0042]

[0043] Among them, the first three formulas are the upper and lower limit constraints of the storage state of charge of the energy storage device, the relationship constraint between energy and state of charge, and the equation relationship constraint between the input and output power of the device and the stored electrical energy. t is the charge state of the EES device during period t; is the amount of electrical energy stored in the EES device during period t; is the rated power capacity of the equipment; σ E is the self-discharge rate of the energy storage device per time period; and are the charging efficiency and discharging efficiency coefficients of the EES device, respectively; and are the charging power and discharging power between the EES device and the distribution network in period t respectively; the last three formulas are the charging and discharging power limit constraints and the logical constraints that a single device cannot charge and discharge at the same time, among which and are the maximum charging power and discharging power of the EES device, respectively; and are 0-1 variables representing the charging state and discharging state respectively.

[0044] Furthermore, in the multi-type flexible resource operation scheduling model:

[0045] When the power grid fails, the mobile energy supply vehicle is used as a mobile flexible resource. The mobile energy supply vehicle operation model of the mobile energy supply vehicle location selection and the on-board equipment operation status is constructed by using high-dimensional 0-1 binary decision variables and the big M method as follows:

[0046]

[0047] Among them, the mobile energy vehicle location variable is a binary decision variable that includes location dimension, time dimension, and the number of mobile energy supply vehicles; When , node i has no mobile energy supply vehicle in time period t; They are the total installed capacity of photovoltaic power source, energy storage and diesel generator in a single mobile energy supply vehicle; They are the total rated charging and discharging power of the power storage device in the mobile energy supply vehicle and the total minimum active power of the diesel generator;

[0048] Among them, the location selection of mobile energy vehicles can also be set as part of the fault recovery strategy rather than a compensation decision.

[0049] The flexible resources of the grid topology are used for load transfer in fault self-organizing network and fault recovery. The model is as follows:

[0050] For a line with a flexibly controllable switching state, the binary switching state variable α ij The large-M method makes the voltage balance equation and power flow equation applicable to distribution networks with flexible resources in topological structures:

[0051] m=(1-α ij )M

[0052]

[0053] Where V i 、V j Represents the square of the voltage amplitude at node i and node j respectively; M represents a positive constant. ij = 1, branch ij is put into operation, and the branch voltage drop relationship and branch power flow relationship constraints take effect; when α ij =0, the constraint does not work, that is, the voltage amplitudes at both ends of the disconnected branch are irrelevant, and there is no power flow on the line;

[0054] The composition of the demand side node electric load is shown as follows:

[0055]

[0056] In the formula, represents the active / reactive load at node i in time period t, represents the predicted electric load at node i in time period t, represents the size of the flexible response electric load at node i in time period t; the demand response scheduling constraints of the flexible response electric load include the flexible response conservation constraint of electric energy and the upper and lower limit constraints of the flexible load in a single time period:

[0057]

[0058] in, is the actual flexible response load size of the node moving in or out during period t, is the maximum flexible response load size that can be moved in or out of a node during period t.

[0059] And, a terminal device includes a memory, a processor, and a computer program stored in the memory and capable of running on the processor, wherein the memory stores a computer program capable of running on the processor, and when the processor loads and executes the computer program, a distribution network fault recovery method based on multi-stage random planning as described above is adopted.

[0060] And, a distribution network fault recovery device based on multi-stage random programming, comprising:

[0061] Generation module: used to generate fault recovery decisions through a multi-period fault recovery model including the objective function of the fault recovery strategy and the maintenance strategy constraints, network reconstruction and self-organizing network coordination constraints in the fault recovery strategy

[0062] and an execution module, which is used to call the generation module to generate a fault recovery decision in combination with the realization of random variable uncertainty, and implement a targeted fault recovery action decision.

[0063] Furthermore, the execution module combines the realization of random variable uncertainty, calls the generation module to generate a fault recovery decision, and implements the specific implementation process of the targeted fault recovery action decision as follows:

[0064] Under the constraint of the total failure recovery time, consider T time stages: t = 1, 2, ..., T, and let the variable representing the uncertain scenario be As the time progresses, it is gradually identified and the following distribution network fault recovery strategies are implemented:

[0065] The goal of the multi-period fault recovery model is to minimize the total amount of load abandonment and the total amount of network loss in the entire recovery process to ensure that the fault recovery strategy is optimal during the entire fault period. The expression of the maintenance strategy is achieved by constraining the opening and closing state of the fault line.

[0066] The dispatch compensation decision includes the coordinated dispatch decision of resources and the adjustment of distribution network state variables.

[0067] Compared with the prior art, the present invention and its preferred solution take into account the uncertainty in the fault recovery of the distribution network, characterize the relationship between random scenarios and multi-stage recovery decisions in the multi-period fault recovery process, thereby forming a basic strategy for the fault recovery of the distribution network based on multi-stage random planning, and further consider the load uncertainty and renewable resource uncertainty in the multi-period network recovery process, and coordinate the dispatch of multiple types of flexible resources such as fixed and mobile, supply side and demand side, and grid structure in the network based on the multi-stage random planning method, reduce power outage losses and improve fault recovery speed, avoid the situation where the recovery strategy is not feasible due to inaccurate prediction or measurement of weak distribution networks, and provide a more reliable and robust fault recovery strategy for weak distribution networks. It constructs a mobile flexible resource operation control model, constructs a network reconstruction and fault recovery collaborative control model according to the characteristics of multiple distributed flexible resources, and handles multiple uncertainties in multi-period fault recovery based on the multi-stage random planning method, thereby realizing the fault recovery of weak distribution networks considering multi-stage uncertainties. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments:

[0069] Figure 1 It is a schematic diagram of a flow chart of constructing a distribution network fault recovery model based on multi-stage stochastic programming according to an embodiment of the present invention;

[0070] Figure 2 It is a schematic diagram of the relationship between multi-period random scenarios and multi-stage recovery decisions in the method of the embodiment of the present invention;

[0071] Figure 3 Schematic diagram of the structure of the distribution network model in this embodiment;

[0072] Figure 4 is a curve diagram of average load loss power during the fault recovery period in this embodiment;

[0073] Figure 5 is a schematic diagram of the fault recovery strategy for period 1 in this embodiment;

[0074] Figure 6 is a schematic diagram of the fault recovery strategy for period 2 in this embodiment;

[0075] Figure 7 is a schematic diagram of the fault recovery strategy for period 3 in this embodiment;

[0076] Figure 8 is a schematic diagram of the fault recovery strategy for period 4 in this embodiment;

[0077] Fig. 9 is a load demand and load supply curve of the distribution network throughout the day under normal working conditions in this embodiment;

[0078] Fig.10 It is a graph of average network loss during the fault period in this embodiment. DETAILED DESCRIPTION

[0079] Hereinafter, the specific embodiments of the present application will be described in detail with reference to the accompanying drawings. According to these detailed descriptions, those skilled in the art can clearly understand the present application and can implement the present application. Without violating the principles of the present application, the features in the various embodiments can be combined to obtain new implementations, or certain features in certain embodiments can be replaced to obtain other preferred implementations.

[0080] In order to make the features and advantages of this patent more obvious and easy to understand, the following embodiments are specifically described in detail with the accompanying drawings as follows:

[0081] like Figure 1 As shown, this embodiment first provides an introduction to the process and overall idea of ​​constructing a distribution network fault recovery method based on multi-stage random programming in the present invention, including the following steps:

[0082] S101: Construct an operation model of multi-type spatiotemporal flexible resources of source-grid-load-storage;

[0083] In this embodiment, first, grid parameters, equipment parameters, and load and renewable resource prediction parameters are imported; Figure 3 As shown, the system structure is an adapted IEEE 33 system.

[0084] Then, illustratively, according to the imported example system parameters, the operation model of source-grid-load-storage multi-type spatiotemporal flexible resources is constructed as follows:

[0085] 1) Build a supply-side fixed flexible resource operation model

[0086] The constructed supply-side fixed flexible resource operation model includes a variety of distributed power sources, such as micro gas turbines, diesel generators, and electric energy storage devices.

[0087] Furthermore, for the constructed micro gas turbine (MT), it has good power response characteristics and can adjust active and reactive power according to the operating state of the power grid. The output power constraint of the micro gas turbine prime mover is the formula, the output power constraint, the active output, reactive output and apparent power relationship constraint are as follows:

[0088] P MT_min ·s MT ≤P MT ≤P MT_rated ·s MT

[0089]

[0090] in, The minimum prime mover output power considering the efficiency factor of the micro gas turbine; P MT is the actual output power of the micro gas turbine prime mover; is the rated output power of the micro gas turbine prime mover; s MT 0-1 variable indicating the start and stop status of the gas turbine; is the rated apparent power of the micro gas turbine; The power generation efficiency of the micro gas turbine; is the actual thermal power output of the micro gas turbine; is the energy loss rate of the micro gas turbine; η HR The efficiency of the waste heat recovery device for the micro gas turbine; It is the active power of the actual output electric energy of the micro gas turbine; It is the reactive power of the actual electrical energy output by the micro gas turbine.

[0091] For the constructed diesel generator (DG) operation model, it can be regarded as the power factor angle Fixed distributed flexible resources. The diesel generator operation model is as follows:

[0092] P DG_min ≤P DG ≤P DG_rated

[0093]

[0094] Where P DG The active power output of the diesel generator; The minimum active power limit is to consider the diesel generator economy and parts life; is the rated active power of the diesel generator; is the rated power factor of the diesel generator, which is a constant; Q DGIt is the reactive power output of the diesel generator.

[0095] For the constructed ESS operation model:

[0096] SOC min ≤SOC t ≤SOC max

[0097]

[0098]

[0099] Among them, the first three formulas are the upper and lower limit constraints of the storage state of charge of the energy storage device, the relationship constraint between energy and state of charge, and the equation relationship constraint between the input and output power of the device and the stored electrical energy. t is the charge state of the EES device during period t; is the amount of electrical energy stored in the EES device during period t; is the rated power capacity of the equipment; σ E is the self-discharge rate of the energy storage device per time period; and is the charging efficiency and discharging efficiency coefficient of the EES device; and is the charging power and discharging power between the EES device and the distribution network during period t. The last three formulas are the charging and discharging power limit constraints and the logical constraints that a single device cannot charge and discharge at the same time. and is the maximum charging power and discharging power of the EES device; and is a 0-1 variable representing the charging state and discharging state.

[0100] 2) Build a supply-side mobile and flexible resource operation model

[0101] The constructed operation model of the integrated mobile energy supply vehicle (IMES): The mobile energy supply vehicle consists of a small diesel generator, a photovoltaic power source, and an electrical energy storage device. The output models of the three types of equipment are no different from the fixed ones, and only the mobile dispatching method needs to be separately modeled and optimized. When a power grid fails, the mobile energy supply vehicle can be used as a mobile flexible resource, allowing us to use smaller-scale power generation resources to achieve better grid operation and dispatch. Furthermore, in the following formula, a mobile energy supply vehicle operation model of the mobile energy supply vehicle location selection and the operating status of the on-board equipment is constructed using high-dimensional 0-1 binary decision variables and the Big M method:

[0102]

[0103]

[0104] The above formula takes into account the capacity and power limit constraints of the mobile energy supply vehicle with flexible and movable characteristics. The operation model of its equipment output is similar to the fixed flexible resource model pre-built above. The difference is that there is a mobile energy supply vehicle location variable This variable is a binary decision variable that includes location dimension, time dimension, and the number of mobile energy supply vehicles. , node i has no mobile energy supply vehicle in time period t. The total installed capacity of photovoltaic power source, energy storage and diesel generator in a single mobile energy supply vehicle; It is the total rated charging and discharging power of the power storage device in the mobile energy supply vehicle and the total minimum active power of the diesel generator.

[0105] 3) Build a flexible resource model for the grid topology

[0106] The constructed grid topology flexible resource scheduling model: The existing distribution network has some redundant lines as backup, that is, closed-loop design and open-loop operation. Therefore, there are certain grid topology flexible resources that can be used for load transfer in fault self-organizing network and fault recovery. Further, the grid topology flexible resource model is as follows:

[0107] For a line with a flexible and controllable switching state, the binary switching state variable α can be used ij The big-M method makes the voltage balance equation and power flow equation applicable to distribution networks with flexible topological resources.

[0108] m=(1-α ij )M

[0109]

[0110] Where V i 、V j Represents the square of the voltage amplitude at node i and node j respectively; the capital letter M represents a large positive number. ij = 1, branch ij is put into operation, and the branch voltage drop relationship and branch power flow relationship constraints take effect; when α ij =0, the constraint does not work, that is, the voltage amplitudes at both ends of the disconnected branch are irrelevant and there is no power flow on the line.

[0111] 4) Build a demand-side responsive flexible resource scheduling model

[0112] The demand-side response flexible resource scheduling model constructed is as follows: the node's electric load is divided into fixed electric load and flexible electric load. On the premise that the total amount of the day remains unchanged, the flexible electric load can be scheduled and allocated on the daily time scale, thereby reasonably reducing the loss of load during faults and accelerating fault recovery.

[0113] Furthermore, the electrical load composition at a certain point is shown as follows:

[0114]

[0115] In the formula, represents the active / reactive load at node i in time period t, represents the predicted electric load at node i in time period t, represents the size of the flexible response electric load at node i in time period t. Furthermore, the demand response scheduling constraint of the flexible electric load can be obtained:

[0116]

[0117] The above formula is the energy conservation constraint of flexible response and the upper and lower limit constraints of flexible load in a single period, where P / is the actual flexible response load size that moves in or out of the node during period t, is the maximum flexible response load size that moves in or out of the node during period t.

[0118] S102: Constructing a post-fault network reconstruction and multi-period fault recovery model based on the multi-type flexible resource operation model; illustratively, constructing a network reconstruction and multi-period fault recovery model based on the multi-type flexible resource operation model in step S101, coordinating the dispatch of multiple types of flexible resources such as fixed and mobile, supply side and demand side, and grid structure in the network, reducing power outage losses and improving fault recovery speed, and the detailed implementation process is as follows:

[0119] 1) Build a network operation model that takes into account flexible grid resources

[0120] Construct a network operation model that takes into account the flexible resources of the grid, and construct node power balance constraints and branch voltage drop constraints, specifically:

[0121]

[0122] In the formula, They represent the total active power injected into node j, the total active power outflowed, the size of the active load, and the active output of the diesel generator in time period t respectively; They represent the total reactive power injected into node j, the total reactive power outflowed, the size of reactive load, and the reactive output of the diesel generator in time period t respectively; They represent the sum of active line loss and reactive line loss injected into node j in time period t; (1+k)U N 、(1-k)U N Respectively represent the upper and lower limits of the voltage at node i; I max Represents the upper limit of the current amplitude on branch (i, j). For a branch that is disconnected during time period t, The values ​​of are all zero, which means that the constraint of disconnecting the line is automatically removed in the model; otherwise, the power balance and line voltage drop equations are both effective.

[0123] Then, using the second-order cone relaxation method, the quadratic terms in the above model are transformed into the following convex model:

[0124]

[0125]

[0126] After the transformation, all decision variables involved in the model meet the application conditions of the second-order cone algorithm, and its solution space has been constrained within the convex cone structure. Therefore, the existing efficient mathematical programming solution tools can be used to solve it.

[0127] 2) Build a network reconstruction model

[0128] The constructed network reconstruction model, based on radial and connectivity constraints, cooperates with the multi-type flexible resource operation scheduling model described in the above step S101 to perform partitioned self-organizing networking after a fault, thereby minimizing fault losses and optimizing fault recovery strategies.

[0129] Furthermore, the radial and connectivity constraints are as follows:

[0130]

[0131] Among them, the first formula represents the node power balance constraint for the unbalanced node without diesel generators, and does not consider the network loss in the virtual power flow, only focusing on the topological state of the network; the second formula represents the node power balance constraint for the balanced node, The third and fourth formulas are the redefined virtual power flow Connection status with branch The two constraints in the last formula indicate that as the fault repair work progresses, more lines are allowed to be gradually added, but the upper limit of the total number of lines |γ|-1 cannot be exceeded at any time. This design helps to simulate the process of gradually restoring power supply in the actual power grid and follow certain maintenance and restoration strategies. In the formula, Denote the power transmitted on line js and line ij in time period t in the virtual network; δ(j) and π(j) denote the set of child nodes and parent nodes of node j; γ denotes the set of nodes in the network; {DG} denotes the set of nodes containing diesel generators in the distribution network; M denotes a sufficiently large number. This model assumes that each island contains only one diesel generator, and therefore the number of islands is the same as the number of diesel generators, so |DG| is a predetermined parameter.

[0132] 3) Multi-period fault recovery strategy model

[0133] A multi-period fault recovery strategy model is constructed. The model mainly includes the objective function of the fault recovery strategy and the maintenance strategy constraints, network reconstruction and self-organizing network coordination constraints in the fault recovery strategy. Specifically:

[0134] The objectives of the fault recovery model are divided into two parts: one is to minimize the total amount of load abandonment, and the other is to minimize the total amount of network loss during the entire recovery process to ensure that the fault recovery strategy is optimal during the entire fault period. Furthermore, the mathematical expression of the objective function is as follows:

[0135]

[0136] Where N T Indicates the total number of time steps within the fault period; N bus represents the total number of nodes in the distribution network; Δt represents the time step within the fault period; represents the load shedding rate of node i in time period t; represents the load demand power of node i in time period t; (i, j) represents the branch determined by nodes i and j; Ψ represents the vector set composed of the starting nodes and the end nodes of all branches; represents the branch current size with node i as the starting node and node j as the ending node in time period t, r i,j Represents the resistance of the branch.

[0137] The expression of the maintenance strategy in the model is realized by constraining the opening and closing states of the fault line. The specific method is as follows:

[0138]

[0139] Where F represents the vector set consisting of the start and end nodes of all faulty branches; h represents the maximum number of branches that can be inspected simultaneously within a given period of time t; T f Indicates the time required to repair a faulty branch; represents the switch state of branch (i, j) in the fault recovery period t. If the branch is disconnected, Otherwise, it is 1. The first formula above indicates that the switch variable of the faulty branch is forced to be 0 in the initial state of the fault, and all are in the disconnected state; the second formula is the maintenance and recovery speed limit. During the fault recovery process, at most h faulty branches can be repaired in each period.

[0140] S103: Considering the uncertainty of load and renewable resources, a multi-stage scenario tree is constructed, and then a network reconstruction and multi-period fault recovery model based on a multi-stage stochastic programming method is constructed to provide a reliable fault recovery strategy for weak distribution networks.

[0141] Specifically, in this embodiment, the specific method of constructing a multi-stage scenario tree and a multi-stage reliable recovery strategy considering the uncertainty of load and renewable resources is as follows:

[0142] First, according to the probability distribution of load demand and renewable resource uncertainty, a multi-stage scenario tree is sampled and constructed to characterize its uncertainty. The scenario tree is as follows: Figure 2 shown.

[0143] Figure 2 The multi-stage scenario tree briefly describes the relationship between random scenarios and multi-stage recovery decisions in the multi-period fault recovery process. For different time stages t = 1, 2, ..., T of a scenario i, due to the random variable It is gradually realized and found out as time goes by, and it is not generated at the same time. Therefore, the corresponding fault recovery decision variables It should also be achieved step by step, that is, observe

[0144] Furthermore, according to the multi-stage stochastic programming method described above, the objective function of the multi-period fault recovery model based on multi-stage stochastic programming can be transcribed into the following mathematical form:

[0145]

[0146] Among them, x, y, and ξ are all expanded into a multi-stage format, the superscript represents the stage number index, the subscript represents the scenario index, and T represents the total time limit for fault recovery. x represents the fault recovery strategies such as network reconstruction and line maintenance and recovery decisions, and y represents the coordinated scheduling decisions of flexible resources such as gas turbines and fixed energy storage, as well as state variables such as distribution network branch power and node voltage.

[0147] Distribution network decision makers make fault recovery decisions at each stage , the final realization result of the uncertain scenario at the current stage t has not been revealed, so for each scenario at the current stage, the planning decision are all unique and identical, so that the model satisfies the unexpected constraints. At the same time, the decision of the tth stage must be completely feasible for all subsequent random scenarios, which is also the source of the reliability of the fault recovery strategy of this method. At the same time, the fault recovery decision of the subsequent stage of multi-stage random planning For the uncertain scenario in the previous stage It is adaptive and its status is equivalent to the operation scheduling compensation decision in the previous stage. Therefore, while ensuring its robustness in dealing with uncertainty, multi-stage stochastic programming also has better economy and flexibility than two-stage stochastic programming or robust optimization.

[0148] Based on the model in this method, distribution network decision makers can coordinate the dispatch of multiple types of flexible resources in the network, such as fixed and mobile, supply side and demand side, and grid structure, when responding to faults, reduce power outage losses and improve fault recovery speed, avoid situations where the recovery strategy of weak distribution networks is not feasible due to inaccurate predictions or measurements, and provide more reliable and robust fault recovery strategies for weak distribution networks.

[0149] According to the above model, a multi-stage weak distribution network robust fault recovery strategy is finally obtained. Distribution network decision makers can implement targeted fault recovery action decisions based on this strategy and the actual realization of random variable uncertainty.

[0150] In this embodiment, within the preset fault recovery time, the corresponding time period is selected, and the power supply demand of each load node in the distribution network during the fault period is obtained according to the load time-varying model and the daily demand curve, and the fault recovery is performed according to the diesel generator output and the load demand of each node. Considering that the fault of the distribution network in an extreme disaster scenario may be in multiple lines, causing power outages in multiple parts of the distribution network, this embodiment sets the distribution network to have a sudden natural disaster at 8:00 in the morning, and a multi-line circuit breaker fault occurs (lines 6, 12, 18, 21, 24, 32). The fault time is 4h, and the processing time for each fault is 1h. At the same time, it is set to overhaul a maximum of 2 branches at the same time (h=2).

[0151] For example, the corresponding fault recovery strategy and fault recovery effect are as follows: Figures 4 to 10 shown.

[0152] Specifically, Figure 5-Figure 8The fault recovery strategy in this embodiment is shown in FIG. After the fault occurs, in period 1, the section switch 26 is disconnected, and the connecting switches 33, 34, 36, and 37 are closed, forming a multi-island operation of G1-G4; in period 2, the fault line 6 and line 24 are overhauled, the section switch 14 is disconnected, and the connecting switches 33, 34, 35, and 37 are closed, forming a single island operation mode; in period 3, the fault lines 12 and 18 are overhauled, and the connecting switches 33 and 37 are closed to reconstruct the network operation; in period 4, all fault branches are overhauled, and the connecting switches are disconnected, and the original network topology is restored, and the active power loss of the entire network is minimized.

[0153] Figure 4 and Fig. 9 , Fig.10 The fault recovery effect and distribution network loss in this embodiment are demonstrated. Specifically, Figure 4 The average load loss power curve during the fault recovery period under uncertain conditions is shown. Fig. 9 The load demand and load supply curves of the distribution network under normal working conditions are shown. At 8:00, the power system failed, the total load supply dropped to 1033kW, unable to meet the load demand of 2775kW, and the load abandonment rate reached 63%; under the strategy of this embodiment, after the first time step (i.e., 9:00), the distribution network was reconstructed and divided into 4 islands, and the corresponding main power supply and diesel generators were used to supply power to each area separately, the total load supply reached 2807kW, the total load demand was 3514kW, and the load abandonment rate dropped to 20%, effectively alleviating the power shortage problem; at 10:00, with the completion of the maintenance of the two faulty lines, the distribution network structure was reconstructed again, the power supply capacity was further enhanced, the total load supply was 3567kW, the total load demand was 3696kW, and the load abandonment rate was reduced to 3.5%; finally, at 11:00, the distribution network restored the original topology, met the supply of all loads, there was no load loss, and the power system operated normally. Therefore, this embodiment fully illustrates the effectiveness of the fault recovery strategy obtained by the method.

[0154] Based on the same inventive concept, the present invention also provides a computer device, which includes: one or more processors, and a memory for storing one or more computer programs; the program includes program instructions, and the processor is used to execute the program instructions stored in the memory. The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, which is used to implement one or more instructions, specifically for loading and executing one or more instructions in a computer storage medium to implement the above method.

[0155] It needs to be further explained that, based on the same inventive concept, the present invention also provides a computer storage medium, on which a computer program is stored, and the computer program is executed by a processor to execute the above method. The storage medium can adopt any combination of one or more computer-readable media. The computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electrical, magnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples (non-exhaustive list) of computer-readable storage media include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present invention, a computer-readable storage medium can be any tangible medium containing or storing a program, which can be used by an instruction execution system, device or device or used in combination with it.

[0156] In the description of this specification, the description with reference to the terms "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present disclosure. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0157] The above shows and describes the basic principles, main features and advantages of the present disclosure. Those skilled in the art should understand that the present disclosure is not limited by the above embodiments, and the above embodiments and descriptions are only for explaining the principles of the present disclosure. Without departing from the spirit and scope of the present disclosure, the present disclosure may have various changes and improvements, and these changes and improvements fall within the scope of the present disclosure to be protected.

[0158] This patent is not limited to the above-mentioned optimal implementation mode. Anyone can derive various other forms of a distribution network fault recovery method based on multi-stage random planning under the inspiration of this patent. All equal changes and modifications made according to the scope of the patent application of the present invention should be covered by this patent.

Claims

1. A distribution network fault recovery method based on multi-stage stochastic programming, characterized in that: Generate fault recovery decisions through a multi-period fault recovery model that includes the objective function of the fault recovery strategy and the maintenance strategy constraints, network reconstruction and self-organizing network coordination constraints in the fault recovery strategy. Combined with the realization of random variable uncertainty, targeted fault recovery action decisions are implemented.

2. A distribution network fault recovery method based on multi-stage stochastic programming according to claim 1, characterized in that: The specific implementation process of the fault recovery decision is as follows: Under the constraint of the total failure recovery time, consider T time stages: t = 1, 2, ..., T, and let the variable representing the uncertain scenario be As the time phase progresses, it is gradually identified and the following distribution network fault recovery strategy is implemented: Observe the variable realization at the time of the fault →Failure recovery decision → Observed scene And make operation scheduling compensation decisions →Failure recovery decision → Observed And make operation scheduling compensation decisions →Failure recovery decision → Observed and make compensation decisions The goal of the multi-period fault recovery model is to minimize the total amount of load abandonment and the total amount of network loss in the entire recovery process to ensure that the fault recovery strategy is optimal during the entire fault period. The expression of the maintenance strategy is achieved by constraining the opening and closing state of the fault line. The dispatch compensation decision includes the coordinated dispatch decision of resources and the adjustment of distribution network state variables.

3. A distribution network fault recovery method based on multi-stage stochastic programming according to claim 2, characterized in that: The variables representing the uncertain scenarios are generated based on probability distribution sampling of load demand and resource uncertainty.

4. The method for distribution network fault recovery based on multi-stage stochastic programming according to claim 2 is characterized in that: The objective function of the multi-period fault recovery model based on multi-stage stochastic programming is: Among them, x, y, ξ are in multi-stage format, the superscript represents the stage number index, the subscript represents the scenario index, T is the upper limit of the time stage, and represents the total time limit for fault recovery; x represents the fault recovery strategy, and y represents the scheduling compensation decision.

5. The method for distribution network fault recovery based on multi-stage stochastic programming according to claim 2 is characterized in that: The objective function expression of the multi-period fault recovery model is as follows: Where N T Indicates the total number of time steps within the fault period; n bus represents the total number of nodes in the distribution network; Δt represents the time step within the fault period; represents the load shedding rate of node i in time period t; represents the load demand power of node i in time period t; (i, j) represents the branch determined by nodes i and j; Ψ represents the vector set composed of the starting nodes and the end nodes of all branches; represents the branch current size with node i as the starting node and node j as the ending node in time period t, r i，j Indicates the resistance of the branch; In the unified fault recovery model of network reconstruction and island partitioning, a series of constraints are added to the variables involving the switch state; during the duration of the fault, the model must ensure that the fault line remains disconnected; the expression of the maintenance strategy in the model is realized by constraining the opening and closing states of the fault line, as follows: Where F represents the vector set consisting of the start and end nodes of all faulty branches; h represents the maximum number of branches that can be inspected simultaneously within a given period of time t; T f Indicates the time required to repair a faulty branch; represents the switch state of branch (i, j) in the fault recovery period t. If the branch is disconnected, Otherwise, 1; In the initial state of the fault, the switch variables of the faulty branches are forced to be 0, and they are all in the disconnected state. During the fault recovery process, at most h faulty branches can be repaired in each period.

6. A distribution network fault recovery method based on multi-stage stochastic programming according to claim 5, characterized in that: The unified fault recovery model includes a network operation model and a network reconstruction model taking into account flexible resources of the grid; The network operation model taking into account the flexible resources of the grid includes node power balance constraints and branch voltage drop constraints, specifically: In the formula, They represent the total active power injected into node j, the total active power outflowed, the size of the active load, and the active output of the diesel generator in time period t respectively; They represent the total reactive power injected into node j, the total reactive power outflowed, the size of reactive load, and the reactive output of the diesel generator in time period t respectively; They represent the sum of active line loss and reactive line loss injected into node j in time period t; (1+k)U N 、(1-k)U N Respectively represent the upper and lower limits of the voltage at node i; I max Indicates the upper limit of the current amplitude on branch (i, j); For the branch that is disconnected during the period t, The value of is zero; M represents a positive constant; The network reconstruction model is based on radial and connectivity constraints, and coordinates multi-type flexible resource operation scheduling models to perform partitioned self-organizing networking after failures, thereby minimizing failure losses and optimizing failure recovery strategies; The radial and connectivity constraints are specifically: Among them, the first formula represents the node power balance constraint for the unbalanced node without diesel generators, and does not consider the network loss in the virtual power flow, only focusing on the topological state of the network; the second formula represents the node power balance constraint for the balanced node, represents the power provided by the "source" bus in the virtual network during period t; The third and fourth formulas are virtual power flows Connection status with branch The two constraints in the last formula indicate that as the fault repair work proceeds, more lines can be gradually added, but the upper limit of the total number of lines |γ|-1 cannot be exceeded at any time; where, They respectively represent the power transmitted on line js and line ij in time period t in the virtual network; δ(j) and π(j) represent the set of child nodes and parent nodes of node j; γ represents the set of nodes in the network; {DG} represents the set of diesel generator nodes in the distribution network.

7. A distribution network fault recovery method based on multi-stage stochastic programming according to claim 6, characterized in that: In the multi-type flexible resource operation scheduling model: The output power constraints of gas turbines include: output power constraints, active output, reactive output and apparent power: P MT_min ·s MT ≤P MT ≤P MT_rated ·s MT in, The minimum prime mover output power considering the efficiency factor of the gas turbine; P MT is the actual output power of the gas turbine prime mover; is the rated output power of the gas turbine prime mover; s MT 0-1 variable indicating the start and stop status of the gas turbine; is the rated apparent power of the gas turbine; is the power generation efficiency of the gas turbine; is the actual thermal power output of the gas turbine; is the energy loss rate of the gas turbine; η HR The efficiency of the waste heat recovery device for the gas turbine; It is the active power of the gas turbine's actual output electrical energy; It is the reactive power of the gas turbine that actually outputs electrical energy; The diesel generator operation model is as follows: P DG_min ≤P DG ≤P DG_rated Where P DG The active power output of the diesel generator; The minimum active power limit is to consider the diesel generator economy and parts life; is the rated active power of the diesel generator; is the rated power factor of the diesel generator; Q DG The reactive power output of the diesel generator; The operation model of the electric energy storage device ESS is: SOC min ≤SOC t ≤SOC max Among them, the first three formulas are the upper and lower limit constraints of the storage state of charge of the energy storage device, the relationship constraint between energy and state of charge, and the equation relationship constraint between the input and output power of the device and the stored electrical energy. t is the charge state of the EES device during period t; is the amount of electrical energy stored in the EES device during period t; is the rated power capacity of the equipment; σ E is the self-discharge rate of the energy storage device per time period; and are the charging efficiency and discharging efficiency coefficients of the EES device, respectively; and are the charging power and discharging power between the EES device and the distribution network in period t respectively; the last three formulas are the charging and discharging power limit constraints and the logical constraints that a single device cannot charge and discharge at the same time, among which and are the maximum charging power and discharging power of the EES device, respectively; and are 0-1 variables representing the charging state and discharging state respectively.

8. A distribution network fault recovery method based on multi-stage stochastic programming according to claim 7, characterized in that: In the multi-type flexible resource operation scheduling model: When the power grid fails, the mobile energy supply vehicle is used as a mobile flexible resource. The mobile energy supply vehicle operation model of the mobile energy supply vehicle location selection and the on-board equipment operation status is constructed by using high-dimensional 0-1 binary decision variables and the big M method as follows: Among them, the mobile energy vehicle location variable is a binary decision variable that includes location dimension, time dimension, and the number of mobile energy supply vehicles; When , node i has no mobile energy supply vehicle in time period t; They are the total installed capacity of photovoltaic power source, energy storage and diesel generator in a single mobile energy supply vehicle; They are the total rated charging and discharging power of the power storage device in the mobile energy supply vehicle and the total minimum active power of the diesel generator; The flexible resources of the grid topology are used for load transfer in fault self-organizing network and fault recovery. The model is as follows: For a line with a flexibly controllable switching state, the binary switching state variable α ij The large-M method makes the voltage balance equation and power flow equation applicable to distribution networks with flexible resources in topological structures: m=(1-a ij )M Where V i 、V j Represents the square of the voltage amplitude at node i and node j respectively; M represents a positive constant. ij = 1, branch ij is put into operation, and the branch voltage drop relationship and branch power flow relationship constraints take effect; when α ij = 0, the constraint does not work, that is, the voltage amplitudes at both ends of the disconnected branch are irrelevant, and there is no power flow on the line; The composition of the demand side node electric load is shown as follows: In the formula, represents the active / reactive load at node i in time period t, represents the predicted electric load at node i in time period t, represents the size of the flexible response electric load at node i in time period t; the demand response scheduling constraints of the flexible response electric load include the flexible response conservation constraint of electric energy and the upper and lower limit constraints of the flexible load in a single time period: in, is the actual flexible response load size of the node moving in or out during period t, is the maximum flexible response load size that can be moved in or out of a node during period t.

9. A terminal device, comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, wherein the memory stores a computer program capable of running on the processor, and when the processor loads and executes the computer program, a distribution network fault recovery method based on multi-stage random planning as described in any one of claims 1 to 8 is adopted.

10. A distribution network fault recovery device based on multi-stage random programming, characterized in that: include: Generation module: used to generate fault recovery decisions through a multi-period fault recovery model including the objective function of the fault recovery strategy and the maintenance strategy constraints, network reconstruction and self-organizing network coordination constraints in the fault recovery strategy and an execution module, which is used to call the generation module to generate a fault recovery decision in combination with the realization of random variable uncertainty, and implement a targeted fault recovery action decision.

11. A distribution network fault recovery device based on multi-stage random programming according to claim 10, characterized in that: The execution module combines the realization of random variable uncertainty, calls the generation module to generate a fault recovery decision, and implements the specific implementation process of the targeted fault recovery action decision as follows: Under the constraint of the total failure recovery time, consider T time stages: t = 1, 2, ..., T, and let the variable representing the uncertain scenario be As the time phase progresses, it is gradually identified and the following distribution network fault recovery strategy is implemented: Observe the variable realization at the time of the fault →Failure recovery decision → Observed scene And make operation scheduling compensation decisions →Failure recovery decision → Observed And make operation scheduling compensation decisions →Failure recovery decision → Observed and make compensation decisions The goal of the multi-period fault recovery model is to minimize the total amount of load abandonment and the total amount of network loss in the entire recovery process to ensure that the fault recovery strategy is optimal during the entire fault period. The expression of the maintenance strategy is achieved by constraining the opening and closing state of the fault line. The dispatch compensation decision includes the coordinated dispatch decision of resources and the adjustment of distribution network state variables.

Images