Load restoration methods and systems for distribution network fault areas

Abstract

This invention relates to the field of microgrid technology, providing a method and system for load restoration in a distribution network fault area. The method includes: after acquiring disaster information of the distribution network, implementing network reconfiguration and load reduction; modeling the distribution network using a graph representation learning method, abstracting the distribution network into a graph structure; modeling the fault management problem of the distribution network as a Markov decision process model based on the graph structure; and achieving rapid load restoration of the microgrid cluster during distribution network faults through coordinated control based on the Markov decision process model. This invention deeply integrates "graph representation learning + Markov decision process" with microgrid clusters, forming a closed-loop solution of "dynamic topology reconfiguration - source-grid-load-storage coordination - zero power outage for critical loads." It requires no additional hardware and can be embedded into existing distribution automation terminals, significantly improving the resilience of the distribution network under extreme disasters, the renewable energy absorption rate, and the power supply reliability of critical loads.

Description

Technical Field

[0001] This invention relates to the field of microgrid technology, and more specifically to a load restoration method and system for a faulty area of ​​a distribution network. Background Technology

[0002] With the increasing frequency of extreme weather events globally, severe disturbances such as typhoons, blizzards, and thunderstorms pose increasingly serious challenges to the safe operation of power distribution networks. Currently, power distribution networks typically adopt a radial structure of "closed-loop design and open-loop operation," relying on a small number of tie switches and sectionalizing switches to achieve fault isolation and load transfer. However, in extreme disaster scenarios, simultaneous failures of multiple lines, power outages at switching stations, and communication interruptions are common occurrences.

[0003] In related technologies, the restoration of distribution networks treats network reconfiguration, load shedding, and microgrid dispatch as three independent sub-problems. Network reconfiguration is mostly based on offline optimization or static power flow models, which has a long solution time and is difficult to provide a satisfactory reconfiguration solution within seconds. Furthermore, it does not adequately consider the randomness of distributed power generation output, easily leading to reverse power flow overload. Load shedding relies on local overcurrent / undervoltage protection or dispatcher experience, lacking quantitative assessment of load importance, voltage sensitivity, and overall network power supply capacity, resulting in "failure to shed when necessary and excessive shedding when unnecessary," with high risk of power outages for critical loads (hospitals, communication base stations). After disasters, distributed power sources are often forced off the grid due to a lack of reasonable voltage / frequency support paths, leading to a large amount of idle clean power generation capacity. Simultaneously, there is a lack of a unified coordination framework between the distribution network and microgrids (MGs), with microgrids operating independently and unable to form a cluster effect. Moreover, there is a lack of unified modeling methods for uncertainties such as simultaneous multi-line faults and switch failures, resulting in poor strategy robustness and missed opportunities to utilize their black start, islanded operation, and flexible power support to accelerate system recovery. Summary of the Invention

[0004] To address the aforementioned technical problems, the first aspect of this invention provides a method for restoring loads in a faulty area of ​​a power distribution network.

[0005] A second aspect of the present invention provides a load restoration system for a faulty area of ​​a power distribution network.

[0006] The technical solution adopted in this invention is as follows:

[0007] A first aspect of the present invention proposes a load restoration method for a distribution network fault area, comprising the following steps: receiving and parsing meteorological information released by a meteorological center; obtaining disaster information of the distribution network based on the meteorological information, and implementing two control measures for the distribution network: network reconfiguration and load reduction; modeling the distribution network using a graph representation learning method, abstracting the distribution network into a graph structure; modeling the fault management problem of the distribution network into a Markov decision process model based on the graph structure; and achieving rapid load restoration of the microgrid cluster in the distribution network fault through collaborative control based on the Markov decision process model.

[0008] The load restoration method for distribution network fault areas described above in this invention also has the following additional technical features:

[0009] According to one embodiment of the present invention, the control measures for implementing network reconfiguration of a distribution network specifically include: establishing an optimization function with the objectives of minimizing network losses and maximizing distributed power absorption; establishing the relationship between line power adjustment and node injected power based on a sensitivity matrix to assess power flow changes; and implementing network reconfiguration of the distribution network based on the optimization function and the power flow changes, wherein the key operating constraints include: radial topology preservation, node voltage safety constraints, line thermal stability constraints, and robustness assurance in extreme scenarios.

[0010] According to one embodiment of the present invention, a graph representation learning method is used to model the distribution network, which is abstracted into a graph structure. Specifically, the distribution network is abstracted into a graph structure G=(N,E), where: the node set N includes all bus nodes, load nodes, distributed energy access nodes and zero power injection nodes in the distribution network; the edge set E includes: the physical connection between the distribution lines and the inline transformers; the edge set E is divided into two functional subsets: controllable lines with switching capabilities, including sectionalizing switch lines and tie switch lines, whose on / off states are adjustable; and fixed connection lines whose connection states are immutable.

[0011] According to one embodiment of the present invention, the fault management problem of the distribution network is modeled as a Markov decision process model based on the graph structure, specifically including: generating state K according to the node set N and edge set E of the graph structure; designing all possible combinations of switch and load operations in the distribution network to generate action B; and learning the transition probability P of the distribution network state in a data-driven manner. tr To maximize power supply And minimize the voltage constraint exceedance objective to set the reward function O; according to Generate a Markov decision process model M.

[0012] According to one embodiment of the present invention, based on the Markov decision process model, a microgrid cluster achieves rapid load recovery in a distribution network fault through coordinated control. Specifically, this includes: acquiring fault information in real time to determine fault zones, and controlling the system to operate the tie switch to disconnect the electrical connection on both sides of the fault zone to form physical isolation; generating a distribution network recovery capability assessment S based on the state K of the Markov decision process model; acquiring a load recovery optimization model, solving the load recovery optimization model, and generating a load recovery strategy.

[0013] A second aspect of the present invention proposes a load restoration system for a distribution network fault area, comprising: a parsing module for receiving and parsing meteorological information released by a meteorological center; a control module for implementing two control measures—network reconfiguration and load reduction—on the distribution network after obtaining disaster information based on the meteorological information; a learning module for modeling the distribution network using a graph representation learning method, abstracting the distribution network into a graph structure; a modeling module for modeling the fault management problem of the distribution network into a Markov decision process model based on the graph structure; and a coordination module for achieving rapid load restoration of a microgrid cluster in a distribution network fault through coordinated control based on the Markov decision process model.

[0014] The load restoration system for the fault area of ​​the power distribution network described above in this invention also has the following additional technical features:

[0015] According to one embodiment of the present invention, the control module is specifically used to: establish an optimization function with the objectives of minimizing network losses and maximizing distributed power absorption; establish the relationship between line power adjustment and node injected power based on a sensitivity matrix to evaluate power flow changes; and implement network reconfiguration of the distribution network based on the optimization function and the power flow changes, wherein the key operating constraints include: radial topology maintenance, node voltage safety constraints, line thermal stability constraints, and robustness assurance in extreme scenarios.

[0016] According to an embodiment of the present invention, the learning module is specifically used to: abstract the power distribution network into a graph structure G=(N,E), wherein: the node set N includes all bus nodes, load nodes, distributed energy access nodes and zero power injection nodes in the power distribution network; the edge set E includes: the physical connection between the power distribution lines and the inline transformers; and divide the edge set E into two functional subsets: controllable lines with switching capabilities, including sectionalizing switch lines and tie switch lines, whose on / off states are adjustable; and fixed connection lines whose connection states are immutable.

[0017] According to one embodiment of the present invention, the modeling module is specifically used to: generate state K based on the node set N and edge set E of the graph structure; design all possible combinations of switch and load operations in the distribution network to generate action B; and learn the transition probability P of the distribution network state in a data-driven manner. tr To maximize power supply And minimize the voltage constraint exceedance objective to set the reward function O; according to Generate a Markov decision process model M.

[0018] According to one embodiment of the present invention, the coordination module is specifically used for: acquiring fault information in real time to determine fault zones, and controlling the system to operate the tie switch to disconnect the electrical connection on both sides of the fault zone to form physical isolation; generating a power distribution network recovery capability assessment S based on the state K of the Markov decision process model; acquiring a load recovery optimization model, solving the load recovery optimization model, and generating a load recovery strategy.

[0019] The beneficial effects of this invention are:

[0020] This invention deeply integrates "graph representation learning + Markov decision process" with microgrid clusters to form a closed-loop solution of "topology dynamic reconfiguration - source-grid-load-storage coordination - zero power outage for critical loads". No new hardware is required, and it can be embedded into existing distribution automation terminals, significantly improving the resilience of the distribution network under extreme disasters, the renewable energy absorption rate, and the power supply reliability of critical loads. Attached Figure Description

[0021] Figure 1 This is a flowchart of a load restoration method for a faulty area in a distribution network according to an embodiment of the present invention;

[0022] Figure 2 This is a block diagram of a load restoration system for a faulty area in a power distribution network according to an embodiment of the present invention. Detailed Implementation

[0023] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] Figure 1 This is a flowchart of a load restoration method for a distribution network fault area according to an embodiment of the present invention, the method comprising the following steps:

[0025] Step S1: Receive and parse meteorological information released by the meteorological center.

[0026] Specifically, when extreme disasters strike the power distribution network, a multi-source data fusion mechanism is first activated under the "cloud-edge-device" collaborative perception framework: the high-performance server cluster in the cloud receives and analyzes the kilometer-level grid forecast data released by the meteorological center in real time through the power dedicated communication network. Its data dimensions not only cover traditional meteorological elements such as typhoon path, wind speed, and ice thickness, but also extend to derived parameters such as precipitation probability, temperature gradient, and solar radiation intensity, providing macroscopic and a priori environmental situation perception for predicting the physical impact of disasters on power distribution lines, towers, and substation facilities.

[0027] Step S2: After obtaining disaster information of the distribution network based on meteorological information, implement two control measures for the distribution network: network reconfiguration and load reduction.

[0028] Specifically, as emergency control measures, network reconfiguration (DNR) and load shedding are two key control measures to prevent voltage collapse and widespread power outages when the distribution network encounters extreme disasters (such as typhoons and blizzards).

[0029] Furthermore, in distribution network fault scenarios caused by extreme disasters, network reconfiguration, as a key control measure, dynamically alters the network topology by adjusting the states of sectionalizing switches and tie switches in the distribution network, thereby optimizing power flow distribution. Its core objectives include three aspects: First, by transferring loads from overloaded lines to lightly loaded lines, it effectively reduces the risk of line overload and avoids cascading faults caused by localized overloads; second, after the faulty components are effectively isolated, rapid network topology reconfiguration maximizes the restoration of power supply to non-faulty areas, significantly improving the system's power supply reliability; and finally, by optimizing the grid connection paths of distributed power sources, it promotes the efficient local consumption of renewable energy while effectively suppressing the impact of reverse power flow on the main grid operation.

[0030] In one embodiment of the present invention, the control measures for implementing network reconfiguration of the distribution network specifically include the following steps S21-S23:

[0031] S21. Establish an optimization function with the goal of minimizing network loss and maximizing distributed power absorption.

[0032] Specifically, the optimization function can be established using the following formula:

[0033] ;

[0034] in, , These are the line's self-conductance and mutual conductance, respectively. , For nodes Active and reactive power, , For nodes Active and reactive power; This is the status variable for the tie switch, and its value can be 0 (open) or 1 (closed). This is the distributed power generation absorption penalty coefficient, used to balance absorption level and economic efficiency.

[0035] S22, based on the sensitivity matrix, establishes the relationship between line power adjustment and node injected power to assess power flow changes.

[0036] Specifically, the relationship between line power adjustment and node injected power can be expressed as:

[0037]

[0038] in, Inject power into the node, This is the line power adjustment amount. This is a sensitivity matrix, reflecting the node... The effect of injection power changes on the line The linear influence of active power provides a theoretical basis for the rapid generation of reconfiguration strategies.

[0039] S23. Based on the optimization function and power flow changes, network reconfiguration is implemented for the distribution network based on key operational constraints, including: radial topology preservation, node voltage safety constraints, line thermal stability constraints, and robustness assurance in extreme scenarios.

[0040] It is important to emphasize that the formulation and implementation of distribution network reconfiguration strategies must meet strict operational constraints. Besides the power balance and equipment operation limits implicit in the aforementioned objective function, it is also essential to ensure that the reconfigured network maintains a radial topology, i.e., no closed-loop operation. Simultaneously, all node voltages must be maintained within permissible safe ranges to avoid voltage exceedances. Furthermore, the power flow of all lines after reconfiguration must not exceed their thermal stability limits to prevent equipment overload. In scenarios where extreme disasters lead to communication outages or partial switch failures, the reconfiguration model must also possess robustness to cope with incomplete topology information or partial control command failures. This typically requires introducing stochastic optimization or robust optimization methods to model the uncertainty of switch states, thereby ensuring the effectiveness of the reconfiguration strategy in harsh real-world environments.

[0041] Specifically, network reconfiguration must ensure that the following key operational constraints are met:

[0042] Preservation of Radial Topology: The reconstructed network must maintain a strictly radial topology, meaning there are no closed-loop operating paths. This is achieved through the following mechanisms: A graph-theory-based spanning tree constraint modeling method is used to construct a node-branch association matrix to describe network connectivity. By introducing binary switch state variables and applying connectivity constraints, it is ensured that there is exactly one connected path between any node and the root node (substation). The system is configured with a dynamic ring network detection module to calculate the basic loop matrix of the network in real time. When a non-zero loop vector is detected, a topology correction mechanism is automatically triggered, prioritizing the disconnection of redundant tie switches to eliminate closed loops. This mechanism works in deep collaboration with the switch change SOE messages uploaded from the edge layer to ensure the real-time accuracy of the topology model, providing a reliable network structure foundation for reconstructing decisions.

[0043] Node voltage safety constraints: All node voltages must be strictly maintained within the permissible safe operating range.

[0044] Line thermal stability constraints: power of all lines It must be below its thermal stability limit. Based on the line power transmission distribution factor, a power flow security assessment model is established:

[0045]

[0046] Among them, thermal stability limit power The factors are determined by the wire type, ambient temperature, and cooling conditions. The current power of the line is given. The system constructs a triple protection mechanism of "pre-verification - real-time monitoring - dynamic correction": before optimization, sensitivity analysis is used to predict overload risk; during operation, the line current is monitored in real time through FTU (Feeder Terminal Unit, distribution switch monitoring terminal), based on... Calculate the actual power. Line voltage, For line current, For phase; when the overload probability exceeds the threshold, power flow reconfiguration is automatically initiated, and the overload is eliminated through load transfer or DG output adjustment.

[0047] Robustness Guarantee in Extreme Scenarios: To address complex scenarios such as communication interruptions and switch malfunctions under extreme disasters, the system establishes a multi-layered robust response mechanism to ensure the reliability of the reconstruction strategy. This mechanism first constructs a switch malfunction probability model based on historical fault data, and then uses scenario analysis to generate a typical scenario set covering over 90% of possible failure modes, providing a data foundation for uncertain decision-making. Based on this, a distributed robust optimization framework is established, with a core min-max optimization model and an objective function that guarantees performance under the worst-case scenario.

[0048]

[0049] in Let X be the set of uncertainties, and let X be the set of handover switches. To constrain violations and penalties, The distributed power absorption penalty coefficient is used to ensure the policy operates within the uncertainty set by minimizing the system performance loss and constraint violation penalty term in the worst-case scenario. Effectiveness within the scope. When communication interruption occurs, edge computing nodes can autonomously activate a pre-stored library of typical reconfiguration schemes and execute preset backup recovery strategies to maintain basic power supply capabilities. Simultaneously, the system strengthens the network topology design through k-edge connectivity analysis in graph theory, ensuring the network still possesses necessary reconfiguration capabilities even in the event of partial switch failure. This establishes a resilient architecture that conforms to the N-1 safety criterion, thus forming a comprehensive robustness guarantee system from predictive optimization to emergency execution, and from algorithm design to topology reinforcement.

[0050] In this invention, load shedding measures refer to key control methods that proactively disconnect some non-critical loads to maintain stable system operation when extreme disasters cause insufficient power grid supply capacity. Its core objectives include: maintaining system voltage stability and preventing node voltage exceedances; prioritizing power supply to critical loads, ensuring the continuous operation of important loads such as hospitals, communication base stations, and emergency facilities.

[0051] To achieve the above objectives, this invention constructs an optimization function aimed at minimizing load reduction costs, the mathematical expression of which is as follows:

[0052]

[0053] in, Indicates load The unit cost reduction is used to reflect its power supply priority; and Representing loads respectively Active power values ​​before and after reduction.

[0054] The optimization process must meet the following constraints: power balance, voltage constraints, and line capacity constraints.

[0055] By coordinating and optimizing load reduction with the pre-deployment of mobile energy storage, and formulating reasonable load control strategies in the pre-disaster stage, the total operating cost of the system can be effectively reduced while ensuring system voltage stability and power supply to critical loads, and the resilience of the power system under extreme disasters can be significantly improved.

[0056] Step S3: The distribution network is modeled using a graph representation learning method, and the distribution network is abstracted into a graph structure.

[0057] In one embodiment of the present invention, a graph representation learning method is used to model the power distribution network, abstracting the power distribution network into a graph structure, specifically including the following steps S31-S32:

[0058] S31, the power distribution network is abstracted as a graph structure G=(N,E), where: the node set N includes all bus nodes, load nodes, distributed energy access nodes and zero power injection nodes in the power distribution network; the edge set E includes: the physical connection between the power distribution lines and the inline transformers, which fully reflects the spatial connection relationship of the power distribution network.

[0059] S32 divides the edge set E into two functional subsets: controllable lines with switching capabilities, including segmented switch lines and tie switch lines, whose on / off states are adjustable; and fixed connection lines, whose connection states are immutable.

[0060] In power distribution networks with switching control capabilities, the core of their interruption management mechanism can be formalized as a dynamic learning and proactive reconstruction process of the network topology. Based on this, the present invention further divides the edge set E into two functional subsets: controllable lines with switching capabilities, including sectionalizing switching lines and tie switching lines, whose on / off states are adjustable, forming the operational basis for network reconstruction; and fixed connection lines, whose connection states are immutable, forming the network's topological skeleton.

[0061] The node set N adopts a three-type data fusion mechanism: predictive variables, including the predicted values ​​of active / reactive power demand or power generation of each node; estimated variables, power-related parameters obtained based on state estimation; and measured variables, real-time measurement data of the three-phase voltage of each node.

[0062] Regarding the edge set E, a power flow simulator is introduced as a simulation tool to obtain measured values ​​of branch power flow that match the actual operating characteristics of the system, providing data support for system state analysis.

[0063] From a graph theory perspective, the distribution network reconfiguration problem can be transformed into an optimization decision-making problem for controllable edge states. At the main distribution network level, the emergency load shedding control strategy is mathematically represented by binary variables associated with switchable load nodes, enabling rapid recovery and stable operation of the system's supply and demand balance under fault or extreme conditions.

[0064] This invention constructs a graph representation model of the power distribution system, establishes a classification system for nodes and edges, and a variable representation mechanism, forming a well-structured, data-driven, and highly scalable analytical framework. This method fully explores the inherent correlation between the topological features and operational data of the power distribution system, providing a model foundation for the subsequent integration of intelligent algorithms such as graph neural networks, and significantly improving the optimization efficiency and decision-making accuracy of power distribution network reconfiguration and emergency control strategies.

[0065] From a graph domain perspective, the reconfiguration of a distribution network is essentially a problem of solving the state (closed or open) of switchable edges (i.e., lines with on / off control capabilities). At the main distribution network level, emergency load shedding control strategies can be mathematically represented by binary variables associated with switchable load nodes.

[0066] Step S4: Based on the Markov decision process model, the microgrid cluster achieves rapid load recovery during distribution network faults through coordinated control.

[0067] In one embodiment of the present invention, the fault management problem of the distribution network is modeled as a Markov decision process model based on a graph structure, specifically including the following steps S41-S45:

[0068] S41, generate state K based on the set of nodes N and the set of edges E of the graph structure.

[0069] (1) Specifically, state K consists of distribution network related observations that reflect the current operating state of the network, including node set N, edge set E and network topology.

[0070]

[0071] in, PNd , QNd The estimated or predicted active and reactive power demand of the node. PNg , QNg The active and reactive power generation at the corresponding nodes. The three-phase voltage measured at the bus is expressed as... V N , V viol This represents a voltage discrepancy in the network. Edge nodes include: power flows through network branches, denoted as... l The network's operating topology is as follows t The total energy supplied in the network E supp The variable O in state K encapsulates the power outage scenario, i.e., multi-line failures in the network, including switch power outages. The inoperability of a downed switch is addressed using a shielding mechanism that inhibits the corresponding handover action, denoted by μ.

[0072] S42, design all possible combinations of switch and load operations in the distribution network to generate action B.

[0073] Specifically, Action B is the control action for emergency response, including switching operations and load shedding, including the number of switchable lines (including segmented lines and tie lines) and the number of switchable loads in the network. The status of line switches is represented by binary variables (0 for open, 1 for closed).

[0074] S43, learn the distribution network state transition probability P in a data-driven manner. tr .

[0075] Transition probability P tr Capture the dynamic properties of networks with emergency response capabilities, denoted as: This represents the transition from the network state at time step t to the network state at step (t+1). The network status at time step (t+1) is represented. This represents the network state at time step t. This represents the action at time step t. Transition probabilities are learned by the agent from its interactions with the environment.

[0076] S44, to maximize power supply And set the reward function O to minimize the voltage constraint overshoot.

[0077] Specifically, the reward function O employs the GRL (Graph Representation Learning) algorithm to take optimal control actions to mitigate power distribution network faults, and its form is as follows:

[0078]

[0079] Indicates the amount of electricity supplied. V viol This indicates a voltage discrepancy in the network, and the reward reflects the maximization of power supply. The goal is to improve the resilience of the distribution network by minimizing voltage constraint exceedances. This is a correction measure to account for specific fault conditions and poor network status under switching operations.

[0080] S45, according to Generate a Markov decision process model M.

[0081] Step S5: Based on the Markov decision process model, the microgrid cluster achieves rapid load recovery during distribution network faults through coordinated control.

[0082] In a specific embodiment of the present invention, based on a Markov decision process model, rapid load recovery of a microgrid cluster during distribution network faults is achieved through coordinated control, specifically including the following steps S51-S53:

[0083] S51 acquires fault information in real time to determine the fault zone and controls the system to operate the interconnection switch to disconnect the electrical connection on both sides of the fault zone, forming physical isolation.

[0084] Specifically, fault information is obtained using a distribution network monitoring system, and the fault region is determined by combining it with a graph theory model. Let N be the set of distribution network nodes and E be the set of lines; the fault region can be represented as a subgraph. ,in, Represents the faulty node. Represents the faulty circuit. , Fault isolation is achieved through the operation of the interconnection switch; isolation time... Must meet , This is the critical power outage time.

[0085] S52, based on the state K of the Markov decision process model, generates the recovery capability assessment S of the distribution network.

[0086] Specifically, the dispatchable capacity of each microgrid is calculated. : ,in Let i be the power generation capacity of microgrid i. This is for local load. The total recoverable capacity of the cluster is... (n is the number of microgrids in the cluster), and the recovery capability assessment S of the distribution network is generated based on the total recoverable capacity of the cluster.

[0087] The resilience assessment S is actually a state processing function: S = f(K). It transforms the raw, multidimensional state observation data K into a highly condensed indicator that is crucial for decision-making.

[0088] S53, obtain the load recovery optimization model, solve the load recovery optimization model based on action B of the Markov decision process model, and generate a load recovery strategy.

[0089] Specifically, the load recovery optimization model is obtained according to the following formula:

[0090] ;

[0091] in, For load priority weights, For load power, To restore the state variables.

[0092] Constraints include power balance and trend constraints , For line current, This is the rated current.

[0093] By solving the model using a distributed algorithm, orderly load recovery can be achieved.

[0094] The load restoration strategy is based on a generator of action B from a Markov decision process model, which is both an optimization model and a solver. It connects the state K and action B of the Markov decision process model. Specifically, it runs the load restoration optimization model, which solves for a specific and optimal load restoration scheme, i.e., determining which loads to restore and which not to restore. The solution of this optimization scheme directly corresponds to a specific action in action B.

[0095] In summary, the load restoration method for distribution network fault areas according to embodiments of the present invention deeply integrates "graph representation learning + Markov decision process" with microgrid clusters to form a closed-loop solution of "topology dynamic reconfiguration - source-grid-load-storage coordination - zero power outage for critical loads". It does not require additional hardware and can be embedded in existing distribution automation terminals, significantly improving the resilience of the distribution network under extreme disasters, the renewable energy absorption rate, and the power supply reliability of critical loads.

[0096] Corresponding to the load restoration method for distribution network fault areas described above, this invention also proposes a load restoration system for distribution network fault areas. Since the system embodiments of this invention correspond to the method embodiments described above, details not disclosed in the system embodiments can be found in the method embodiments described above, and will not be repeated here.

[0097] Figure 2 This is a block diagram of a load restoration system for a distribution network fault area according to an embodiment of the present invention, as shown below. Figure 2 As shown, the system includes: a parsing module, a control module, a learning module, a modeling module, and a collaboration module.

[0098] The system comprises the following modules: a parsing module for receiving and parsing meteorological information from the meteorological center; a control module for implementing network reconfiguration and load reduction measures on the distribution network after obtaining disaster information based on the meteorological information; a learning module for modeling the distribution network using graph representation learning methods, abstracting the distribution network into a graph structure; a modeling module for modeling the fault management problem of the distribution network into a Markov decision process model based on the graph structure; and a coordination module for achieving rapid load recovery of the microgrid cluster during distribution network faults through coordinated control based on the Markov decision process model.

[0099] According to one embodiment of the present invention, the control module is specifically used to: establish an optimization function with the objectives of minimizing network losses and maximizing distributed power absorption; establish the relationship between line power adjustment and node injected power based on the sensitivity matrix to evaluate power flow changes; and implement network reconfiguration of the distribution network based on the optimization function and power flow changes, wherein the key operating constraints include: radial topology maintenance, node voltage safety constraints, line thermal stability constraints, and robustness assurance in extreme scenarios.

[0100] According to one embodiment of the present invention, the learning module is specifically used to: abstract the power distribution network into a graph structure G=(N,E), wherein: the node set N includes all bus nodes, load nodes, distributed energy access nodes and zero power injection nodes in the power distribution network; the edge set E includes: the physical connection between the power distribution lines and the inline transformers; and divide the edge set E into two functional subsets: controllable lines with switching capabilities, including sectionalizing switch lines and tie switch lines, whose on / off states are adjustable; and fixed connection lines whose connection states are immutable.

[0101] According to one embodiment of the present invention, the modeling module is specifically used to: generate state K based on the node set N and edge set E of the graph structure; design all possible combinations of switch and load operations in the distribution network to generate action B; and learn the transition probability P of the distribution network state in a data-driven manner. tr To maximize power supply And minimize the voltage constraint exceedance objective to set the reward function O; according to Generate a Markov decision process model M.

[0102] According to one embodiment of the present invention, the coordination module is specifically used for: acquiring fault information in real time to determine fault zones, and controlling the system to operate the tie switch to disconnect the electrical connection on both sides of the fault zone to form physical isolation; generating a power distribution network recovery capability assessment S based on the state K of the Markov decision process model; acquiring a load recovery optimization model, solving the load recovery optimization model, and generating a load recovery strategy.

[0103] The load restoration system for the fault area of ​​the distribution network according to the present invention deeply integrates "graph representation learning + Markov decision process" with microgrid clusters to form a closed-loop solution of "topology dynamic reconfiguration - source-grid-load-storage coordination - zero power outage for important loads". It does not require new hardware and can be embedded in existing distribution automation terminals, which significantly improves the resilience of the distribution network under extreme disasters, the renewable energy absorption rate and the power supply reliability of critical loads.

[0104] In the description of this invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. "A plurality of" means two or more, unless otherwise explicitly specified.

[0105] In this specification, the terms "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, those skilled in the art can combine and integrate the different embodiments or examples and features described in this specification without contradiction. Furthermore, those skilled in the art can combine and integrate the different embodiments or examples described herein, as well as the features of the different embodiments or examples, without contradiction.

[0106] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing custom logic functions or processes, and the scope of preferred embodiments of the invention includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of the invention pertain.

[0107] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the computer-readable medium may be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0108] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware as in another embodiment, it can be implemented using any of the following techniques known in the art, or a combination thereof: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0109] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0110] Furthermore, the functional units in the various embodiments of the present invention can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0111] The storage medium mentioned above can be a read-only memory, a disk, or an optical disk, etc. Although embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention.

[0112] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for load restoration of a faulted area of an electric power distribution network, characterized by, The method comprises the following steps: receiving and analyzing meteorological information released by a meteorological center; after obtaining disaster information of the power distribution network according to the meteorological information, implementing two control measures of network reconstruction and load reduction on the power distribution network; modeling the power distribution network by using a graph representation learning method, and abstracting the power distribution network into a graph structure; modeling a fault management problem of the power distribution network into a Markov decision process model based on the graph structure; based on the Markov decision process model, realizing rapid load recovery of the micro-grid cluster in the power distribution network fault through collaborative control; wherein, based on the Markov decision process model, rapid load recovery of the micro-grid cluster in the power distribution network fault is realized through collaborative control, specifically including: The fault information is obtained in real time by using a power distribution network monitoring system to determine a fault partition, a fault area is determined by combining a graph theory model, and a system operation tie switch is controlled to disconnect electrical connections on both sides of the fault area to form physical isolation, a power distribution network node set is N, a line set is E, and a fault area is represented as a subgraph wherein, represents a fault node, represents a fault line, , Fault isolation is achieved by tie switch action, and the isolation time needs to meet , is a critical power-off time; Generate the restoration capability evaluation S of the power distribution network based on the state K of the Markov decision process model, and calculate the schedulable capacity of each micro-grid : wherein, , wherein is the power generation of the micro-grid i, is the local load, and the total recoverable capacity of the cluster is n is the number of micro-grids in the cluster, and the restoration capability evaluation S of the power distribution network is generated according to the total recoverable capacity of the cluster; An action B based on a Markov decision process model is used to solve the load restoration optimization model by a distributed algorithm to generate a load restoration strategy and realize orderly load restoration, and the load restoration optimization model is obtained according to the following formula: ; wherein, is a load priority weight, is a load power, is a restoration state variable; constraint conditions include power balance and power flow constraints , is a line current, is a rated current.

2. The method of claim 1, wherein, the control measure of implementing network reconstruction on the power distribution network, specifically including: establishing an optimization function with the objective of minimizing network loss and maximizing distributed power consumption; based on a sensitivity matrix, establishing a relationship between line power adjustment and node injection power to evaluate power flow changes; based on the optimization function and the power flow changes, implementing network reconstruction on the power distribution network based on key operation constraints, wherein the key operation constraints include: radial topology structure maintenance, node voltage safety constraint, line thermal stability constraint and extreme scenario robustness guarantee.

3. The method of claim 1, wherein, modeling the power distribution network by using a graph representation learning method, and abstracting the power distribution network into a graph structure, specifically including: abstracting the power distribution network into a graph structure G=(N,E), wherein: the node set N includes all bus nodes, load nodes, distributed energy access nodes and zero-power injection nodes in the power distribution network; the edge set E includes: physical connections of power distribution lines and interlinked transformers; dividing the edge set E into two functional subsets: controllable lines with switching capability, including sectionalizing switch lines and tie switch lines, whose on-off state is adjustable; fixed connection lines, whose connection state is invariable.

4. The method of claim 1, wherein, modeling the fault management problem of the power distribution network into a Markov decision process model based on the graph structure, specifically including: generating a state K according to the node set N and the edge set E of the graph structure; designing a combination of all possible switch and load operations in the power distribution network to generate an action B; Learning transition probabilities P of a power distribution grid in a data-driven manner tr ; to maximize power supply and minimize voltage constraint excursions. According to A Markov decision process model M is generated.

5. A load restoration system for a power distribution network failure area, characterized by, including: the analysis module is used for receiving and analyzing meteorological information released by a meteorological center; the control module is used for implementing two control measures of network reconstruction and load reduction on the power distribution network after obtaining disaster information of the power distribution network according to the meteorological information; the learning module is used for modeling the power distribution network by using a graph representation learning method, and abstracting the power distribution network into a graph structure; the modeling module is used for modeling a fault management problem of the power distribution network into a Markov decision process model based on the graph structure; the collaborative module is used for realizing rapid load recovery of the micro-grid cluster in the power distribution network fault through collaborative control based on the Markov decision process model; wherein, the collaborative module is specifically used for: The fault information is obtained in real time by using a power distribution network monitoring system to determine a fault partition, a fault area is determined by combining a graph theory model, and a system operation tie switch is controlled to disconnect electrical connections on both sides of the fault area to form physical isolation, a power distribution network node set is N, a line set is E, and a fault area is represented as a subgraph wherein, represents a fault node, represents a fault line, , Fault isolation is achieved by tie switch action, and the isolation time needs to meet , is a critical power-off time; Generate the restoration capability evaluation S of the power distribution network based on the state K of the Markov decision process model, and calculate the schedulable capacity of each micro-grid : wherein, , wherein is the power generation of the micro-grid i, is the local load, the total recoverable capacity of the cluster is n is the number of micro-grids in the cluster, and the restoration capability evaluation S of the power distribution network is generated according to the total recoverable capacity of the cluster; An action B based on a Markov decision process model is used to solve the load restoration optimization model by a distributed algorithm to generate a load restoration strategy and realize orderly load restoration, and the load restoration optimization model is obtained according to the following formula: ; wherein, is a load priority weight, is a load power, is a restoration state variable; constraint conditions include power balance and power flow constraints , is a line current, is a rated current.

6. The power distribution network faulted zone load restoration system of claim 5 wherein, the control module is specifically used for: establishing an optimization function with the objective of minimizing network loss and maximizing distributed power consumption; based on a sensitivity matrix, establishing a relationship between line power adjustment and node injection power to evaluate power flow changes; According to the optimization function and the power flow change, network reconfiguration is implemented on the power distribution network based on key operation constraints, wherein the key operation constraints include: radial topology structure maintenance, node voltage safety constraint, line thermal stability constraint and extreme scenario robustness guarantee.

7. The power distribution network faulted zone load restoration system of claim 5, wherein, The learning module is specifically used for: The power distribution network is abstracted as a graph structure G=(N, E), wherein: the node set N includes all bus nodes, load nodes, distributed energy access nodes and zero-power injection nodes in the power distribution network; and the edge set E includes: physical connections of power distribution lines and inner-link transformers; The edge set E is divided into two functional subsets: controllable lines with switching capability, including sectionalizing switch lines and tie switch lines, whose on-off states are adjustable; and fixed connection lines, whose connection states are unchangeable.

8. The power distribution network faulted zone load restoration system of claim 5, wherein, The modeling module is specifically used for: According to the node set N and the edge set E of the graph structure, a state K is generated; All possible combinations of switch and load operations in the power distribution network are designed to generate actions B; Learning transition probabilities P of a power distribution grid in a data-driven manner tr ; to maximize power supply and minimize voltage constraint excursions. According to A Markov decision process model M is generated.

Images