Power distribution network power supply restoration method considering node importance and restoration dependence

Abstract

The invention provides a power distribution network power supply recovery method considering node importance and recovery dependence. The method comprises the following steps: 1) obtaining fault information of a power distribution network in real time; 2) comprehensively evaluating the importance weight of each node by adopting a comprehensive weight method combining an analytic hierarchy process and an entropy weight method; 3) using a reverse breadth-first search algorithm from the fault points to identify a repair dependency relationship among the fault points, and determining an upstream fault set which needs to be repaired firstly when any fault point is repaired; 4) quantifying the reduction amount of the ice melting device deployed before the disaster on the fault recovery time, and correcting the fault recovery time of each fault point in the fault set; 5) constructing a repair sequence optimization model with the goal of minimizing the total weighted load loss amount; 6) solving the repair sequence optimization model by the scheduling control center by adopting an improved genetic algorithm, and generating an optimal sequence for fault repair; and 7) dynamically adjusting the optimal sequence of fault repair.

Description

Technical Field

[0001] This invention relates to a power supply restoration method for distribution networks that takes into account node importance and repair dependence, and belongs to the field of power system restoration control and emergency management technology. Background Technology

[0002] As the "last mile" of the power system, the distribution network directly faces a wide range of users. The reliability of power supply from the distribution network is crucial to the operation of the social economy and the protection of people's livelihoods. Extreme weather events such as snow and typhoons can easily cause multiple faults in the distribution network, resulting in large-scale and long-term power outages. Traditional post-disaster power restoration strategies rely heavily on the experience of dispatchers and lack systematic quantitative decision support, making it difficult to quickly formulate the optimal sequence for fault repair in complex fault scenarios.

[0003] Existing research has made some progress in power restoration, but the following shortcomings still exist:

[0004] 1) Most recovery strategies only aim to restore the maximum load or minimize the number of switching operations, failing to fully consider the differences in importance of different users / loads, which may lead to delays in the recovery of critical loads;

[0005] 2) Failure to fully consider the repair dependencies between faults; the power grid has a radial structure, and the restoration of power supply to a node often depends on the repair of its upstream lines and nodes; ignoring this dependency may result in the repair team making ineffective trips or the repaired node still being unable to supply power due to the failure of the upstream to be restored;

[0006] 3) Failure to coordinate and optimize the deployment of pre-disaster de-icing devices with post-disaster fault repair and power restoration operations; De-icing can prevent faults from worsening or restore power supply capacity, and its deployment should be integrated with the repair and power restoration plan;

[0007] 4) There is a lack of comprehensive consideration of the impact mechanism among the deployment of pre-disaster ice-melting devices, emergency repair, and power restoration in ice disaster scenarios.

[0008] Therefore, there is an urgent need for a rapid post-disaster recovery method for distribution networks that can comprehensively consider node importance, repair dependencies, the disaster mitigation effect of de-icing devices, and coordinate and optimize emergency repair and power restoration operations, so as to minimize the overall power outage losses after a disaster. Summary of the Invention

[0009] This invention proposes a power supply restoration method for distribution networks that considers node importance and repair dependence. Its purpose is to comprehensively consider node importance, repair dependence, and the disaster reduction effect of de-icing devices, and to minimize the overall power outage loss after a disaster by establishing a repair sequence optimization model with the goal of minimizing the total weighted load loss.

[0010] The technical solution of this invention: a power supply restoration method for distribution networks that considers node importance and repair dependency, the method comprising:

[0011] Step 1) Obtain real-time distribution network fault information to obtain detailed information on each fault point after the disaster and the deployment information of the ice melting device before the disaster;

[0012] Step 2) Select three characteristics—load level, number of key users, and network centrality—as evaluation indicators, and use a comprehensive weighting method combining the analytic hierarchy process (AHP) and entropy weighting to comprehensively evaluate the importance weight of each node.

[0013] Step 3) Use the reverse breadth-first search algorithm to identify the repair dependencies between each fault point and determine the set of upstream faults that need to be repaired first when repairing any fault point;

[0014] Step 4) Based on the coverage of the pre-disaster deployed de-icing devices, quantify the reduction in fault repair time caused by the pre-disaster deployed de-icing devices, and correct the fault repair time for each fault point in the fault set.

[0015] Step 5) Construct a repair sequence optimization model with the objective of minimizing the total weighted load loss;

[0016] Step 6) The scheduling and control center uses an improved genetic algorithm to solve the repair sequence optimization model and generate the optimal sequence for fault repair.

[0017] Step 7) During the recovery process, after the scheduling center receives information about newly discovered faults or changes in existing faults, it re-executes steps 3) to 6) to dynamically adjust the optimal sequence of fault repair until no new faults occur and the existing fault information remains unchanged.

[0018] Furthermore, obtaining detailed information on each post-disaster fault point and pre-disaster de-icing device deployment information specifically includes:

[0019] Step 1-1): Input the post-disaster distribution network topology. Fault set Pre-disaster ice-melting device deployment assembly ,in: Represents the set of nodes in the power grid. Represents the set of power grid lines, and ;

[0020] Steps 1-2) Define the fault set For each fault point , The information is represented by a tuple: , These represent the fault points. Location coordinates, fault type, and fault start time; for each de-icing device , The information is represented by a tuple: ,in, Indicates the deployment location. Indicates the coverage area.

[0021] Furthermore, the comprehensive weighting method combining the analytic hierarchy process (AHP) and entropy weighting to comprehensively evaluate the importance weight of each node specifically includes:

[0022] Step 2-1) Construct the evaluation index matrix: Assume there are a total of nodes to be evaluated. There are 1,000 evaluation indicators. One, forming the original evaluation indicator matrix. , It is the first The distribution network node at the first Evaluation values ​​under each indicator;

[0023] Step 2-2): Use the extreme value method to analyze the original evaluation index matrix. Standardization is performed to obtain the standardization matrix. : ,in They are the first The maximum and minimum values ​​of each indicator across all nodes. For the first The distribution network node at the first The standardized evaluation value under each indicator;

[0024] Steps 2-3): Calculate the combined weights;

[0025] Steps 2-4) Calculate nodes Importance score : , .

[0026] Furthermore, the load level refers to the load carried by the node; the number of critical users refers to the number of critical users within the power supply range of the node; and the network centrality refers to the degree centrality of the node, reflecting its connectivity importance in the network topology.

[0027] Furthermore, the calculation of the combined weights specifically includes:

[0028] Step 2-3-1) Constructing the subjective weights of the analytic hierarchy process A judgment matrix is ​​constructed through expert scoring, and the subjective weight vector is obtained after calculating the eigenvectors and passing a consistency check. ;

[0029] Step 2-3-2) Constructing the objective weights of the entropy weight method ;

[0030] Step 2-3-3) Calculate the combined weights , By objective weight and subjective weight We get the weighted sum. The Each component , ,in, for The One portion, This is the preference coefficient.

[0031] Furthermore, the objective weights constructed using the entropy weight method... Specifically, it includes:

[0032] Step 2-3-2-1) Calculate the first... The node at the th The proportion under each indicator : ;

[0033] Step 2-3-2-2), calculate the first... Entropy value of the item index : It is a constant. , The goal is to standardize the entropy value to the [0, 1] interval;

[0034] Step 2-3-2-3), calculate the first... Information utility value of the indicator : , ;

[0035] Steps 2-3-2-4) Calculate the objective weights The Each component : , .

[0036] Furthermore, the step of identifying the repair dependencies between faults using a reverse breadth-first search algorithm from the fault points, and determining the set of upstream faults that need to be repaired first when repairing any fault point, specifically includes: identifying each fault point based on the radial topology of the distribution network. The repair dependency relationship means that a fault can only be truly restored through a power restoration operation after all its upstream faults have been repaired and the path to the power source is clear; (Definition) Fault point The set of all upstream faults is used to start from the fault point using a reverse breadth-first search algorithm. Traverse in reverse order, recording all reachable parent nodes that have failed, and for all failed nodes... Perform a reverse breadth-first search algorithm to obtain the global dependency repair relationships: .

[0037] Furthermore, the reduction in fault repair time due to the pre-disaster deployment of de-icing devices specifically includes:

[0038] Step 4-1) Based on each fault point of Each ice-melting device of and Calculate each fault point Was it caused by a certain ice-melting device? Override; Define an indicator function:

[0039]

[0040] Step 4-2) Assume that the de-icing device can reduce the repair time of faults in the covered area by a certain percentage. ,0< If the value is less than 1, the corrected fault repair time is:

[0041]

[0042] in, For indicator functions, Emergency repair of fault points when no de-icing device is available Time required This is the corrected fault repair time.

[0043] Furthermore, the total weighted unload is: ,in, Represents nodes per unit of time The load loss, if there is a power outage, is Otherwise, it is 0; For time step, For nodes Importance score, the objective function is defined as: ;

[0044] The decision variable of the repair sequence optimization model is each fault point. Repair start time and repair completion time The constraints for fixing the sequence optimization model include:

[0045] (1) Power restoration dependency constraints: fault It must fail in all its upstream areas. Power restoration can only begin after the repairs are completed;

[0046] (2) Resource constraints: The number of faults that can be repaired in parallel at the same time does not exceed the total number of emergency repair teams; Assuming there are R emergency repair teams, at most R faults can be repaired at the same time.

[0047] (3) Repair time constraint: the repair completion time for each fault point Equal to repair start time Adding the corrected fault repair time , ;

[0048] (4) Power supply restoration constraints: After a fault is repaired, the dispatch center can immediately restore the power supply to the downstream non-faulty areas through remote operation.

[0049] Furthermore, the improved genetic algorithm includes the following:

[0050] 6-1) Coding: Chromosomes use fault numbers to indicate the order in which repairs are performed;

[0051] 6-2) Decoding: Based on the chromosome sequence, power supply dependency constraints, resource constraints, and the corrected fault repair time. A scheduling simulator is used to determine the repair start time and repair completion time for each fault point, and the total weighted load loss is calculated as the fitness value.

[0052] 6-3) Genetic operations: New populations are generated using operations such as roulette wheel selection, sequential crossover, and crossover mutation;

[0053] 6-4) Feasibility handling: After crossover and mutation, sequences that violate dependencies may be generated. Use heuristic rules to adjust the chromosome to satisfy dependencies as much as possible.

[0054] 6-5) Output the optimal fault repair sequence obtained by the genetic algorithm. The system calculates the start and end times of repair for each fault point, as well as the estimated total weighted load loss. The dispatch center directs the repair team to work according to the optimal sequence for fault repair and promptly performs remote power restoration operations after the fault is repaired.

[0055] The beneficial effects of this invention are:

[0056] 1) This invention is a power supply restoration method for a distribution network that takes into account the impact mechanism of the deployment of de-icing devices and emergency repair. This invention can realize priority restoration of important nodes, and at the same time considers the influence between de-icing, emergency repair and power supply restoration, thus improving the efficiency of power supply restoration after disasters.

[0057] 2) This invention ensures the priority recovery of critical users and important nodes by quantifying the importance of nodes, thereby reducing core losses;

[0058] 3) Accurate identification of repair dependencies avoids ineffective emergency repair scheduling and improves emergency repair efficiency;

[0059] 4) This invention integrates the deployment of ice-melting devices, the fault repair process, and the power grid restoration operation into a unified model, achieving coordinated optimization of physical repair and operational control;

[0060] 5) The improved genetic algorithm used in this invention can effectively solve complex NP-hard problems, providing schedulers with directly executable decision-making solutions;

[0061] 6) This invention can dynamically adjust the recovery strategy according to environmental changes, adapting to complex and ever-changing application scenarios after disasters. Attached Figure Description

[0062] Appendix Figure 1 This is a schematic diagram of the method flow of the present invention.

[0063] Appendix Figure 2 This is the pseudocode of the reverse breadth-first search algorithm in a specific embodiment of the present invention.

[0064] Appendix Figure 3 This is the pseudocode of the improved genetic algorithm in a specific embodiment of the present invention.

[0065] Appendix Figure 4 This is the pseudocode for calculating the total weighted unload in a specific embodiment of the present invention. Detailed Implementation

[0066] The technical solution of the present invention will be clearly and completely described below with reference to the accompanying drawings and specific embodiments.

[0067] A distribution network power restoration method that considers node importance and repair dependency includes the following steps:

[0068] Step 1) Acquire distribution network fault information in real time to obtain detailed information on each fault point after the disaster and the deployment information of the pre-disaster de-icing device; the real-time acquisition of distribution network fault information includes multi-source data such as distribution SCADA (data acquisition and monitoring control system), UAV inspection, and fault detection and monitoring equipment, to acquire distribution network fault information in real time, and input the post-disaster distribution network topology, fault set, and pre-disaster de-icing device deployment information; the distribution network fault information includes the location, type, and occurrence time of each fault point;

[0069] Step 2) Select three characteristics as evaluation indicators: load level, number of key users, and network centrality. Use a comprehensive weighting method combining the analytic hierarchy process and entropy weighting to comprehensively evaluate the importance weight of each node.

[0070] Step 3) Use the backward breadth-first search (BFS) algorithm to identify the repair dependencies between each fault point and determine the set of upstream faults that need to be repaired first when repairing any fault point;

[0071] Step 4) Based on the coverage of the pre-disaster deployed de-icing devices, quantify the reduction in fault repair time caused by the pre-disaster deployed de-icing devices, and correct the fault repair time for each fault point in the fault set.

[0072] Step 5) Construct a repair sequence optimization model with the objective of minimizing the total weighted load loss. The repair sequence optimization model comprehensively considers the node importance weight, repair dependency, load size, and the reduction of power outage time by the pre-disaster deployed de-icing device.

[0073] Step 6) The dispatch control center uses an improved genetic algorithm to solve the repair sequence optimization model and generate the optimal sequence for fault repair. The improved genetic algorithm uses a scheduling simulator based on repair dependency and resource constraints to calculate the fitness value of chromosomes and output the optimal sequence for fault repair to guide emergency repair and power restoration operations.

[0074] Step 7) During the recovery process, after the scheduling center receives information about newly discovered faults or changes in existing faults, it re-executes steps 3) to 6) to dynamically adjust the optimal sequence of fault repair until no new faults occur and the existing fault information remains unchanged.

[0075] Step 1) Detailed information on each post-disaster failure point and pre-disaster de-icing device deployment information is required, specifically including the following steps:

[0076] Step 1-1: Input the post-disaster distribution network topology Fault set Pre-disaster ice-melting device deployment assembly ,in: Represents the set of nodes in the power grid. Represents the set of power grid lines, and ;

[0077] Step 1-2: Define the fault set For each fault point , The information is represented by a tuple: , These represent the fault points. Location coordinates, fault type, and fault start time; for each de-icing device , The information is represented by a tuple: ,in, Indicates the deployment location. Indicates the coverage area.

[0078] Step 2) To prioritize the recovery of critical nodes, the importance of distribution network nodes (load points) needs to be quantitatively assessed. This method selects three core indicators (load level, number of critical users, and network centrality) and uses a combination of the analytic hierarchy process (AHP) (subjective weighting) and entropy weighting (objective weighting) to comprehensively assess the importance of nodes. The comprehensive weighting method combining AHP and entropy weighting to comprehensively assess the importance weight of each node specifically includes the following steps:

[0079] Step 2-1) Construct the evaluation index matrix: Assume there are a total of nodes to be evaluated. There are 1,000 evaluation indicators. One, forming the original evaluation indicator matrix. , It is the first The distribution network node at the first The evaluation values ​​under each indicator; the three core indicators are: (1) load level: the load carried by the node, (2) number of key users: the number of key users such as hospitals, governments, and factories within the power supply range of the node, (3) network centrality: the degree centrality of the node, reflecting its connectivity importance in the network topology; these three indicators, from the three indispensable dimensions of economy, socio-political, and network structure, together constitute the standard for evaluating the importance of the node;

[0080] Step 2-2) Use the extreme value method to analyze the original evaluation index matrix. Standardization is performed to obtain the standardization matrix. : ,in They are the first The maximum and minimum values ​​of each indicator across all nodes. For the first The distribution network node at the first The standardized evaluation values ​​under each indicator are used to unify all indicators to the same scale for comparison and weighted calculation, thereby eliminating differences in the dimensions and directions of different indicators.

[0081] Steps 2-3) Calculate the combined weights. This invention uses a weighted sum of the subjective weights from the analytic hierarchy process (AHP) and the objective weights from the entropy weight method to construct the combined weights. This method takes into account both the decision-maker's experience and the objective laws of the data. The specific steps include:

[0082] Step 2-3-1) Constructing the subjective weights of the analytic hierarchy process A judgment matrix is ​​constructed through expert scoring, and the subjective weight vector is obtained after calculating the eigenvectors and passing a consistency check. ;

[0083] Step 2-3-2) Constructing the objective weights of the entropy weight method Includes the following steps:

[0084] (1) Calculate the first The node at the th The proportion under each indicator : ;

[0085] (2) Calculate the first Entropy value of the item index : It is a constant. , The goal is to standardize the entropy value to the [0, 1] interval;

[0086] (3) Calculate the first Information utility value of the indicator : , ;

[0087] (4) Calculate the objective weights The Each component : , ;

[0088] Step 2-3-3) Calculate the combined weights , By objective weight and subjective weight We get the weighted sum. The Each component , ,in, for The One portion, This is the preference coefficient, usually taken as 0.5;

[0089] Steps 2-4) Calculate nodes Importance score : , .

[0090] In step 3), a reverse breadth-first search algorithm is used to identify the repair dependencies between faults from the fault points, and to determine the set of upstream faults that need to be repaired first when repairing any fault point. Specifically, this includes: identifying each fault point based on the radial topology of the distribution network. The repair dependency relationship means that a fault can only be truly restored through a power restoration operation (such as closing the switch) after all its upstream faults have been repaired and the path to the power source is clear; definition Fault point Given a set of all upstream faults, use a reverse breadth-first search (BFS) algorithm to find the faults from the fault points. Traverse in reverse order, recording all reachable and failed parent nodes, and for all failed nodes... Perform a reverse breadth-first search algorithm to obtain the global dependency repair relationships: .

[0091] In this invention, the de-icing device can shorten the fault repair time by preventing or mitigating icing. It is necessary to quantify the reduction in fault repair time caused by the de-icing device. Step 4) quantifies the reduction in fault repair time caused by the pre-disaster deployed de-icing device, specifically including:

[0092] Step 4-1): Based on each fault point of Each ice-melting device of and Calculate each fault point Was it caused by a certain ice-melting device? Override; Define an indicator function:

[0093]

[0094] Step 4-2): Assume that the de-icing device can reduce the repair time of faults within the covered area by a certain percentage. (0< <1), then the corrected fault repair time is:

[0095]

[0096] in, The indicator function defined in step 4-1), Emergency repair of fault points when no de-icing device is available Time required Depending on the type of fault and the location of the fault, This is the corrected fault repair time.

[0097] The goal of step 5) is to determine an optimal sequence for fault repair. This minimizes the total weighted load loss during the entire recovery process. The total weighted load loss is defined as the sum of the integrals of the products of the load loss per unit time of all nodes, their importance weights, and the duration of the outage. Since the repair sequence optimization model is discrete, the total weighted load loss ( It can be approximated as: ,in, Represents nodes per unit of time The load loss, if there is a power outage, is Otherwise, it is 0; For time step, The node obtained in step 2) Importance score, the objective function is defined as: Minimize the total weighted unload ( ));

[0098] The decision variable for the repair sequence optimization model is each fault point. Repair start time and repair completion time The constraints for fixing the sequence optimization model include:

[0099] (1) Power restoration dependency constraints: fault It must fail in all its upstream areas. Power restoration can only begin after the repairs are completed. ;

[0100] (2) Resource constraints: The number of faults that can be repaired in parallel at the same time does not exceed the total number of emergency repair teams; Assuming there are R emergency repair teams, at most R faults can be repaired at the same time.

[0101] (3) Repair time constraints: Consider the repair time after the pre-disaster de-icing device deployment. Repair completion time for each fault point. Equal to repair start time Adding the corrected fault repair time , ;

[0102] (4) Power supply restoration constraints: After a fault is repaired, the dispatch center can immediately restore the power supply to the downstream non-faulty areas through remote operation.

[0103] The improved genetic algorithm in step 6) employs fitness evaluation based on scheduling simulation, encodes it as a fault repair sequence, and considers power restoration dependency constraints and resource constraints during decoding. The improved genetic algorithm includes the following:

[0104] 6-1) Coding: Chromosomes use fault numbers to indicate the order in which repairs are performed;

[0105] 6-2) Decoding: Based on the chromosome sequence, power supply dependency constraints, resource constraints, and the corrected fault repair time. A scheduling simulator is used to determine the repair start time and repair completion time for each fault point, and the total weighted load loss is calculated as the fitness value (the smaller the better).

[0106] 6-3) Genetic operations: New populations are generated using operations such as roulette wheel selection, sequential crossover, and crossover mutation;

[0107] 6-4) Feasibility handling: After crossover and mutation, sequences that violate dependencies may be generated. Use heuristic rules to adjust the chromosome to satisfy dependencies as much as possible.

[0108] 6-5) Output the optimal fault repair sequence obtained by the genetic algorithm. The system calculates the start and end times of repair for each fault point, as well as the estimated total weighted load loss. The dispatch center directs the repair team to work according to the optimal sequence for fault repair and promptly performs remote power restoration operations after the fault is repaired.

[0109] This invention is a power supply restoration method for distribution networks that considers the impact mechanism of ice-melting device deployment and fault repair. It can provide a decision-making method for rapid power supply restoration of distribution networks after extreme weather such as snow and ice disasters. This invention is a rapid power supply restoration algorithm that comprehensively considers the importance of node loads, fault repair dependencies, the effectiveness of pre-disaster ice-melting device deployment, and the coordinated scheduling of repair resources. This invention is applicable to post-disaster emergency response scenarios with known fault location and type.

Claims

1. A distribution network power supply restoration method considering node importance and repair dependency, characterized in that: include: Step 1) Obtain real-time distribution network fault information to obtain detailed information on each fault point after the disaster and the deployment information of the ice melting device before the disaster; Step 2) Select three characteristics—load level, number of key users, and network centrality—as evaluation indicators, and use a comprehensive weighting method combining the analytic hierarchy process (AHP) and entropy weighting to comprehensively evaluate the importance weight of each node. Step 3) Use the reverse breadth-first search algorithm to identify the repair dependencies between each fault point and determine the set of upstream faults that need to be repaired first when repairing any fault point; Step 4) Based on the coverage of the pre-disaster deployed de-icing devices, quantify the reduction in fault repair time caused by the pre-disaster deployed de-icing devices, and correct the fault repair time for each fault point in the fault set. Step 5) Construct a repair sequence optimization model with the objective of minimizing the total weighted load loss; Step 6) The scheduling and control center uses an improved genetic algorithm to solve the repair sequence optimization model and generate the optimal sequence for fault repair. Step 7) During the recovery process, after the scheduling center receives information about newly discovered faults or changes in existing faults, it re-executes steps 3) to 6) to dynamically adjust the optimal sequence of fault repair until no new faults occur and the existing fault information remains unchanged.

2. The power supply restoration method for a distribution network considering node importance and repair dependence according to claim 1, characterized in that: The detailed information obtained about each fault point after the disaster and the deployment information of the ice-melting device before the disaster specifically includes: Step 1-1): Input the post-disaster distribution network topology. Fault set Pre-disaster ice-melting device deployment assembly ,in: Represents the set of nodes in the power grid. Represents the set of power grid lines, and ; Steps 1-2) Define the fault set For each fault point , The information is represented by a tuple: , These represent the fault points. Location coordinates, fault type, and fault start time; for each de-icing device , The information is represented by a tuple: ,in, Indicates the deployment location. Indicates the coverage area.

3. The power supply restoration method for a distribution network considering node importance and repair dependence according to claim 1, characterized in that: The comprehensive weighting method, which combines the analytic hierarchy process (AHP) and entropy weighting, is used to comprehensively evaluate the importance weight of each node. Specifically, it includes: Step 2-1) Construct the evaluation index matrix: Assume there are a total of nodes to be evaluated. There are 1,000 evaluation indicators. One, forming the original evaluation indicator matrix. , It is the first The distribution network node at the first Evaluation values ​​under each indicator; Step 2-2): Use the extreme value method to analyze the original evaluation index matrix. Standardization is performed to obtain the standardization matrix. : ,in They are the first The maximum and minimum values ​​of each indicator across all nodes. For the first The distribution network node at the first The standardized evaluation value under each indicator; Steps 2-3): Calculate the combined weights; Steps 2-4) Calculate nodes Importance score : , .

4. The power supply restoration method for a distribution network considering node importance and repair dependence according to claim 1, characterized in that: The load level refers to the load carried by the node; the number of critical users refers to the number of critical users within the power supply range of the node; and the network centrality refers to the degree centrality of the node, reflecting its connectivity importance in the network topology.

5. A distribution network power supply restoration method considering node importance and repair dependence according to claim 3, characterized in that: The calculation of the combined weights specifically includes: Step 2-3-1) Constructing the subjective weights of the analytic hierarchy process A judgment matrix is ​​constructed through expert scoring, and the subjective weight vector is obtained after calculating the eigenvectors and passing a consistency check. ; Step 2-3-2) Constructing the objective weights of the entropy weight method ; Step 2-3-3) Calculate the combined weights , By objective weight and subjective weight We get the weighted sum. The Each component , ,in, for The One portion, This is the preference coefficient.

6. A distribution network power supply restoration method considering node importance and repair dependence according to claim 5, characterized in that: The construction of objective weights using the entropy weight method Specifically, it includes: Step 2-3-2-1) Calculate the first... The node at the th The proportion under each indicator : ; Step 2-3-2-2), calculate the first... Entropy value of the item index : It is a constant. , The goal is to standardize the entropy value to the [0, 1] interval; Step 2-3-2-3), calculate the first... Information utility value of the indicator : , ; Steps 2-3-2-4) Calculate the objective weights The Each component : , .

7. A distribution network power supply restoration method considering node importance and repair dependence according to claim 1, characterized in that: The process involves using a reverse breadth-first search algorithm to identify the repair dependencies between faults from the fault points, and determining the set of upstream faults that must be repaired first when repairing any fault point. Specifically, this includes identifying each fault point based on the radial topology of the distribution network. The repair dependency relationship means that a fault can only be truly restored through a power restoration operation after all its upstream faults have been repaired and the path to the power source is clear; (Definition) Fault point The set of all upstream faults is used to start from the fault point using a reverse breadth-first search algorithm. Traverse in reverse order, recording all reachable parent nodes that have failed, and for all failed nodes... Perform a reverse breadth-first search algorithm to obtain the global dependency repair relationships: .

8. A distribution network power supply restoration method considering node importance and repair dependence according to claim 1, characterized in that: The reduction in fault repair time caused by the pre-disaster deployment of de-icing devices specifically includes: Step 4-1) Based on each fault point of Each ice-melting device of and Calculate each fault point Was it caused by a certain ice-melting device? Override; Define an indicator function: Step 4-2) Assume that the de-icing device can reduce the repair time of faults in the covered area by a certain percentage. ,0< If the value is less than 1, the corrected fault repair time is: in, For indicator functions, Emergency repair of fault points when no de-icing device is available Time required This is the corrected fault repair time.

9. A distribution network power supply restoration method considering node importance and repair dependence according to claim 1, characterized in that: The total weighted unload is: ,in, Represents nodes per unit of time The load loss, if there is a power outage, is Otherwise, it is 0; For time step, For nodes Importance score, the objective function is defined as: ; The decision variable of the repair sequence optimization model is each fault point. Repair start time and repair completion time The constraints for fixing the sequence optimization model include: (1) Power restoration dependency constraints: fault It must fail in all its upstream areas. Power restoration can only begin after the repairs are completed; (2) Resource constraints: The number of faults that can be repaired in parallel at the same time does not exceed the total number of emergency repair teams; Assuming there are R emergency repair teams, at most R faults can be repaired at the same time. (3) Repair time constraint: the repair completion time for each fault point Equal to repair start time Adding the corrected fault repair time , ; (4) Power supply restoration constraints: After a fault is repaired, the dispatch center can immediately restore the power supply to the downstream non-faulty areas through remote operation.

10. A distribution network power supply restoration method considering node importance and repair dependence according to claim 1, characterized in that: The improved genetic algorithm includes the following: 6-1) Coding: Chromosomes use fault numbers to indicate the order in which repairs are performed; 6-2) Decoding: Based on the chromosome sequence, power supply dependency constraints, resource constraints, and the corrected fault repair time. A scheduling simulator is used to determine the repair start time and repair completion time for each fault point, and the total weighted load loss is calculated as the fitness value. 6-3) Genetic operations: New populations are generated using operations such as roulette wheel selection, sequential crossover, and crossover mutation; 6-4) Feasibility handling: After crossover and mutation, sequences that violate dependencies may be generated. Use heuristic rules to adjust the chromosome to satisfy dependencies as much as possible. 6-5) Output the optimal fault repair sequence obtained by the genetic algorithm. The repair start time and repair completion time for each fault point, as well as the estimated total weighted load loss; The dispatch center directed the repair team to work according to the optimal sequence for repairing the fault, and promptly performed remote power restoration operations after the fault was repaired.

Images