A multi-period fast power supply recovery method for distribution network based on flexibility resources

Abstract

The application discloses a kind of based on flexibility resource's power distribution network multi-period fast power supply recovery method, including steps as follows: collection meteorological information and historical failure data, combine power distribution network topological characteristics, establish failure prediction model, based on failure prediction result, unified modeling is carried out to distributed power supply, energy storage and adjustable load, considering capacity, response speed and constraint condition, dynamically assess the flexibility resource that can participate in recovery, in combination with resource capacity evaluation result, construct three-stage optimization model of key load priority-regional expansion-full recovery, formulate time-sharing power supply recovery strategy.The application not only realizes multi-period stage power supply recovery after natural disaster and sudden accident, guarantees the priority of key users and important load, but also significantly improves the scheduling efficiency and utilization of flexibility resource in disaster recovery, reduces user outage loss and power grid operation risk.

Description

Technical Field

[0001] This invention relates to the field of power system operation and dispatching technology, and in particular to a method for rapid power supply restoration in distribution networks across multiple time periods based on flexible resources. Background Technology

[0002] With the rapid development of distributed energy, energy storage devices, and demand-side response technologies, distribution networks are gradually transforming from the traditional unidirectional power supply structure to active networks with bidirectional power flow and flexible dispatch capabilities. However, in the event of natural disasters such as typhoons, snowstorms, floods, or earthquakes, or sudden accidents such as equipment failures or line trips, distribution networks still face the risk of widespread power outages. In such situations, traditional power restoration methods mainly rely on manual inspections and fixed restoration strategies, which often suffer from slow response times and low resource dispatch efficiency, making it difficult to meet the demands for speed and reliability under disaster conditions.

[0003] Existing power restoration methods mostly involve emergency repairs and dispatching only after a disaster occurs, lacking proactive modeling of disaster prediction and fault development. This results in reactive restoration efforts, lacking initiative and preventative measures. Meanwhile, although distributed generation, energy storage systems, and adjustable loads have been applied to a certain extent in distribution networks, these flexible resources are not fully utilized in existing restoration strategies. Most methods still rely primarily on main grid power or backup power, failing to fully leverage the rapid response and local support capabilities of flexible resources, leading to decreased restoration efficiency.

[0004] Furthermore, existing restoration strategies often aim to restore power to the entire grid as quickly as possible, lacking a phased and time-specific approach. In the initial stages of a disaster, critical loads such as hospitals and emergency command centers often cannot be prioritized, and the order of power restoration between residential and commercial areas lacks scientific rigor, leading to uneven resource allocation and difficulty in balancing the social and economic benefits of power supply. A more prominent problem is that during actual restoration, the grid's operational status and user demand constantly change over time. For example, the state of charge of energy storage systems dynamically adjusts with charging and discharging, some loads may experience temporary new demands, and some damaged lines may be gradually repaired. However, existing restoration plans are usually pre-defined, static strategies lacking real-time adjustments. This makes timely correction and optimization during implementation difficult, resulting in a disconnect between power restoration and actual demand, reducing the reliability and flexibility of grid restoration.

[0005] Existing technologies for power distribution network disaster recovery suffer from problems such as slow response, insufficient flexibility, and rigid recovery processes, making it difficult to meet the requirements of new power systems for speed, flexibility, and robustness. There is an urgent need for an innovative approach that integrates disaster prediction, flexible resource scheduling, multi-stage recovery, and dynamic correction mechanisms to achieve efficient and adaptive power supply recovery and enhance the power grid's emergency response capabilities under disaster conditions. Summary of the Invention

[0006] The purpose of this invention is to provide a method for rapid power restoration of distribution networks in multiple time periods based on flexible resources. This method not only enables phased power restoration in multiple time periods after natural disasters and emergencies, ensuring priority power restoration for critical users and important loads, but also significantly improves the scheduling efficiency and utilization rate of flexible resources in disaster recovery, reducing power outage losses for users and risks to grid operation.

[0007] The technical solution of the present invention to solve the above-mentioned technical problems is as follows:

[0008] This invention provides a method for rapid power restoration in a distribution network across multiple time periods based on flexible resources, comprising the following steps:

[0009] S1: After a natural disaster occurs, collect meteorological information and historical fault data, and combine them with the characteristics of the power distribution network topology to establish a fault prediction model;

[0010] S2: Based on the prediction results of the fault prediction model, a unified model is performed on distributed power sources, energy storage and adjustable loads, taking into account capacity, response speed and constraints, and dynamically evaluating the flexible resources that can participate in the recovery.

[0011] S3: Based on the resource capacity assessment results, construct a three-stage optimization model of critical load priority—regional expansion—full restoration, and formulate time-based power restoration strategies.

[0012] Optionally, in step S1, the collected meteorological information includes wind speed. Rainfall Snow thickness Ambient temperature Historical fault data includes the frequency of line faults. Average repair time Equipment reliability indicators The distribution network topology is represented by the node set N and the branch set E, where the node load is... The branch road has a carrying capacity of .

[0013] Optionally, in step S1, establishing the fault prediction model includes:

[0014] Establish a node failure probability model:

[0015] ;

[0016] in, Let be the weighting coefficient, satisfying ;

[0017] The probability of line damage is determined by a combination of meteorological parameters and historical fault frequency.

[0018] ;

[0019] in, branch road Historical failure frequency This is the adjustment coefficient;

[0020] When predicting the duration of a power outage, the repair time and the severity of the disaster should be considered:

[0021] ;

[0022] in, As a disaster intensity indicator, This is the disaster amplification factor;

[0023] The final risk assessment function is formed as follows:

[0024] ;

[0025] in, This represents the importance weight of the nodes.

[0026] Alternatively, in step S2,

[0027] The distributed power sources include distributed photovoltaic, wind power, and small gas turbine units, whose available power output Due to its maximum installed capacity With resource availability coefficient Joint decision:

[0028] ;

[0029] in, Depends on weather conditions;

[0030] The modeling of the energy storage system focuses on the change process of the state of charge (SOC). The SOC evolution model is as follows:

[0031] ;

[0032] in, and These represent charging power and discharging power, respectively. and For charging and discharging efficiency, Let the total energy storage capacity be denoted by ; constraints are introduced as follows:

[0033] ;

[0034] The formula for adjustable load modeling is:

[0035] ;

[0036] in, This indicates the load reduction ratio, with a value range of [value range missing]. .

[0037] Alternatively, in step S2, the availability vector of the unified flexibility resource is:

[0038] ;

[0039] in, This represents the adjustable capacity of energy storage at time t.

[0040] Alternatively, in step S3, the construction of the three-stage optimization model specifically involves:

[0041] The objective function of the critical load priority restoration phase is to minimize the power outage losses of the critical loads:

[0042] ;

[0043] in, Representing the importance weights of different loads, Indicates load The duration of the power outage;

[0044] The constraints are:

[0045] ;

[0046] The objective function of the area expansion and recovery phase is to maximize power supply coverage.

[0047] ;

[0048] in, For nodes The power supply status is indicated by a value of 1, which means power has been restored, and a value of 0 means power is still out. The total number of nodes in residential and industrial / commercial areas;

[0049] The constraints are:

[0050] ;

[0051] The objective function of the full recovery and operation optimization phase is to minimize operating costs while taking into account network losses.

[0052] ;

[0053] in, These represent the cost coefficients for distributed power generation, energy storage, and network losses, respectively.

[0054] The constraints are:

[0055] ;

[0056] Maintain tidal balance;

[0057] .

[0058] Alternatively, in step S3, the time-segmented power restoration strategy incorporates rolling time-domain optimization, adjusting the power supply within each time window. Within this framework, the objective function is uniformly expressed as:

[0059] ;

[0060] in, The weighting coefficients are for different time periods, initially... Larger capacity to ensure critical loads, medium term Increase to improve coverage, later The dominant approach is to pursue economic efficiency and stability.

[0061] Optionally, step S4 is also included, which involves real-time monitoring of power grid operation and user demand during the recovery process, iteratively revising and hierarchically scheduling the established strategies, dynamically adapting to the evolution of the disaster, and achieving a rapid power restoration closed loop from initial emergency response to final stability.

[0062] Alternatively, step S4 may specifically be as follows:

[0063] In the real-time monitoring phase, node voltage is collected. Line flow Energy storage state of charge Actual output of distributed power sources User load demand The data is compared with the predicted values, and the deviation is calculated.

[0064] ;

[0065] in, Represents various operating status parameters; when deviation Exceeding the set threshold When this happens, the strategy correction process is automatically triggered;

[0066] In the iterative correction phase, a rolling time-domain optimization method is adopted, assuming that the recovery process is divided into multiple time windows. At the end of each window, the objective function is recalculated:

[0067] ;

[0068] in, The strategy will be dynamically adjusted based on the stage of disaster development; if the difference between the new calculation results and the original strategy exceeds a threshold... If so, the scheduling instructions will be updated immediately;

[0069] In the hierarchical scheduling process, the upper-level scheduling center is responsible for global resource coordination, including power allocation of distributed power sources and cross-regional energy storage scheduling; the lower-level control unit is responsible for specific execution, and its constraints are as follows:

[0070] .

[0071] The present invention has the following beneficial effects:

[0072] This invention proposes a multi-period rapid power supply restoration method for distribution networks based on flexible resources. Through disaster scenario identification and fault evolution modeling, it achieves forward-looking prediction of high-risk areas and critical loads in the early stages of a disaster, providing a scientific basis for resource scheduling. By unifying the modeling and dynamic capability assessment of distributed power sources, energy storage, and adjustable loads, it fully leverages the rapid response and adjustment potential of flexible resources, effectively improving resource utilization and scheduling efficiency. Through a phased optimization model of "critical load priority—regional expansion—full restoration," it ensures the power supply security of important users and gradually expands the power supply range, ultimately achieving stable operation of the entire network. Simultaneously, the introduction of a real-time monitoring and iterative correction mechanism enables the restoration strategy to be dynamically adjusted according to the evolution of the disaster and changes in the grid's operating status, ensuring the speed, adaptability, and robustness of power supply restoration. This significantly reduces user power outage losses and grid operation risks, demonstrating promising engineering application prospects and widespread application value. Attached Figure Description

[0073] Figure 1 A flowchart of the multi-period rapid power supply restoration method for distribution networks based on flexible resources provided by the present invention; Detailed Implementation

[0074] The principles and features of the present invention are described below with reference to the accompanying drawings. The examples given are only for explaining the present invention and are not intended to limit the scope of the present invention.

[0075] Example

[0076] This invention provides a method for rapid power restoration in a distribution network across multiple time periods based on flexible resources, with reference to... Figure 1 As shown, the method includes:

[0077] S1: After a natural disaster, meteorological information and historical fault data are collected and combined with the characteristics of the distribution network topology to establish a fault prediction model, identifying potentially damaged areas and fault development trends as a prerequisite for subsequent resource scheduling. Disaster information includes not only meteorological factors such as wind speed, rainfall, and ice thickness, but also fault records and outage durations from the power grid's operational history, forming a multi-dimensional risk analysis. By comparing disaster characteristics with power grid vulnerabilities, potentially damaged areas and possible fault development trends are identified, thus providing a forward-looking basis for subsequent power restoration and resource scheduling.

[0078] The collected meteorological information includes wind speed. Rainfall Snow thickness Ambient temperature Parameters such as line fault frequency and historical fault data are included. Average repair time Equipment reliability indicators The distribution network topology is represented by the node set N and the branch set E, where the node load is... The branch road has a carrying capacity of .

[0079] To achieve quantitative analysis of disaster impact, a node failure probability model is established:

[0080] ;

[0081] in, Let be the weighting coefficient, satisfying .

[0082] The probability of line damage is determined by taking into account both meteorological parameters and historical fault frequency.

[0083] ;

[0084] in, branch road Historical failure frequency This is the adjustment coefficient.

[0085] When predicting the duration of a power outage, both the repair time and the severity of the disaster should be considered:

[0086] ;

[0087] in, For disaster intensity indicators (such as wind force level or snowfall). This is the disaster amplification factor.

[0088] The final risk assessment function is formed as follows:

[0089] ;

[0090] in, The importance weight of nodes is assigned, for example, hospitals and command centers have a higher weight than residential areas. According to... By sorting the data from largest to smallest, a "priority recovery list for key areas" can be generated based on the impact of the disaster.

[0091] Preferably, the prediction model can perform differentiated modeling for different disasters: when the disaster is a typhoon, the weight of the wind speed factor is increased. When the disaster is freezing, a constraint on the thickness of ice accretion on the power lines is introduced. When the disaster is flooding, increase the node's geographical altitude. The influencing factor is used to determine the probability of equipment damage in low-lying areas.

[0092] S2: Based on fault prediction results, a unified model is constructed for distributed power sources, energy storage, and adjustable loads, considering capacity, response speed, and constraints. This dynamically assesses the flexible resources available for recovery, providing input for multi-period optimization. Distributed power source modeling considers daytime and nighttime output differences, energy storage devices are analyzed in conjunction with state of charge and charging / discharging capabilities, and adjustable loads are categorized based on user-side load characteristics and interruption tolerance. This unified modeling approach clearly defines the available time periods, capacity potential, and response speeds of different types of flexible resources, thus providing complete input for recovery optimization.

[0093] The distributed power sources include distributed photovoltaic, wind power, and small gas turbine units, etc., and their available power output... Due to its maximum installed capacity With resource availability coefficient Joint decision:

[0094] ;

[0095] in, This depends on meteorological conditions, such as sunlight intensity or wind speed. The formula indicates that at any given time, the output of distributed power sources cannot exceed their rated capacity and is subject to adjustment by environmental conditions. This modeling approach allows disaster prediction information (such as insufficient sunlight due to cloudy or rainy weather) to be directly translated into constraints on resource availability, making the model more closely reflect actual operation.

[0096] The modeling of the energy storage system focuses on the change in state of charge (SOC). Energy storage typically plays a role in "rapid response" and "stable support" during disaster recovery; its availability depends not only on rated power but also on the current charge level. Its SOC evolution model is as follows:

[0097] ;

[0098] in, and These represent charging power and discharging power, respectively. and For charging and discharging efficiency, This represents the total energy storage capacity. To avoid overcharging or deep discharging, constraints also need to be introduced:

[0099] ;

[0100] This model can dynamically reflect the remaining regulatory capacity of energy storage after a disaster. For example, when the State of Charge (SOC) is low, the discharge capacity of energy storage will be significantly limited, so the charging and replenishment process needs to be prioritized in the recovery strategy to ensure the continuous support capacity in subsequent stages.

[0101] Adjustable load (demand response) modeling focuses on user-side flexibility. Its baseline load... Under normal circumstances, no adjustments are made. However, during disaster recovery, some non-critical electricity demand can be appropriately reduced based on user tolerance. The modeling formula is as follows:

[0102] ;

[0103] in, This indicates the load reduction ratio, with a value range of [value range missing]. At the same time, to ensure users' basic needs are met, the reduction in playtime will be subject to a maximum tolerance period. There are limitations. For example, residential lighting can be reduced temporarily, but the load reduction in hospitals is almost zero. This modeling allows for the flexible release of demand-side adjustment potential while ensuring users' basic rights.

[0104] After completing the modeling of distributed power sources, energy storage, and adjustable loads, this invention further proposes an availability vector that unifies flexible resources:

[0105] ;

[0106] in, This represents the adjustable capacity of energy storage at time t, encompassing both its instantaneous power and reflecting the sustainability constraints of its State of Charge (SOC). This unified modeling approach allows for a comprehensive comparison of the availability of different resources within a single framework, providing standardized input for subsequent recovery optimization.

[0107] Preferably, in this embodiment, the response speed of various resources is also constrained. For example, the output change rate of distributed power sources. The rated ramp rate should not be exceeded to avoid voltage fluctuations caused by sudden changes; the charge / discharge switching time of the energy storage should be less than [a certain value]. This ensures its rapid adaptability during the recovery process; the response time of the adjustable load should not exceed [a certain value]. This ensures that reduction or recovery can be completed within a reasonable timeframe. These constraints further enhance the engineering feasibility of the model.

[0108] S3: Based on the resource capacity assessment results, construct a three-stage optimization model of "critical load priority - regional expansion - full restoration", and formulate time-based power restoration strategies to ensure optimal timing of power supply paths and resource allocation. The first stage focuses on ensuring the power restoration needs of critical loads such as hospitals, emergency command centers, and communication hubs to ensure the continuity of public safety and emergency response. The second stage gradually expands the power supply range to restore power to residential areas and some industrial and commercial areas. The third stage aims to achieve stable power supply across the entire network, balancing power distribution in different areas with the safe operation of the power grid, thereby achieving the orderly progress of power restoration.

[0109] In this step, the fault prediction results obtained in S1 and the availability of flexible resources in S2 are used as inputs to further establish a three-stage optimization model. Each stage corresponds to different recovery objectives and constraints, satisfying both the urgency of power restoration and the long-term stability of grid operation.

[0110] Phase 1: Prioritize the restoration of critical loads

[0111] In the initial stages of a disaster, the primary objective is to ensure power supply to critical loads such as hospitals, emergency command centers, and communication hubs. The objective function at this stage is to minimize power outage losses to these critical loads.

[0112] ;

[0113] in, The importance weights represent different workloads, with hospitals and command centers having a higher weight than ordinary users; Indicates load The duration of the power outage. This objective function allows for the priority allocation of limited capacity from energy storage and distributed power sources during scheduling.

[0114] The constraints are:

[0115] ;

[0116] This constraint ensures that at any given time, the available output of distributed power sources and energy storage, after deducting demand-side response reductions, must be sufficient to meet the needs of all critical loads. This means that even if the entire power grid is severely damaged, critical nodes can still maintain power supply.

[0117] Phase Two: Regional Expansion Recovery

[0118] After critical loads are secured, the recovery objective shifts to expanding power supply coverage, gradually restoring power to residential and commercial areas. The objective function at this stage is to maximize power supply coverage.

[0119] ;

[0120] in, For nodes The power supply status is indicated by a value of 1, which means power has been restored, and a value of 0 means power is still out. This represents the total number of nodes in residential and industrial / commercial areas.

[0121] The constraints are:

[0122] ;

[0123] The first requirement ensures that node voltage remains stable within the acceptable range, while the second ensures that line power flow does not exceed safe capacity. This phase emphasizes "point-to-area" expansion, using the coordination of distributed power sources and energy storage to maximize power supply coverage. However, when expanding power supply, voltage and current constraints must be strictly adhered to; otherwise, secondary accidents may easily occur. Therefore, the optimization goal of this phase is to restore power to as many users as possible without exceeding the grid's safety boundaries.

[0124] Phase 3: Full Recovery and Operation Optimization

[0125] As the disaster progresses into its later stages and the main grid is gradually repaired, the recovery objective shifts to providing full power supply across the entire grid. The objective function at this stage is to minimize operating costs while also considering network losses.

[0126] ;

[0127] in, These represent the cost coefficients for distributed power generation, energy storage, and network losses, respectively. This objective function aims to reduce ineffective losses and improve economic efficiency while simultaneously restoring the entire network.

[0128] The constraints are:

[0129] ;

[0130] Maintain tidal balance;

[0131] ;

[0132] The goal is to ensure that all nodes are fully restored to power upon completion of the restoration process. This phase emphasizes "efficiency and cost," meaning that while achieving full grid power restoration, it prioritizes the economic efficiency of operation and the rationality of resource allocation. Energy storage is no longer solely for emergency use but enters a phase of peak shaving and valley filling, optimizing operation. Adjustable loads are also gradually restored, ultimately achieving comprehensive grid stability.

[0133] Time-based rolling optimization mechanism

[0134] To improve the adaptability of the method, this invention introduces rolling time-domain optimization. In each time window... Within this framework, the objective function is uniformly expressed as:

[0135] ;

[0136] in, The weighting coefficients are for different time periods, initially... Larger capacity to ensure critical loads, medium term Increase to improve coverage, later The primary focus is on achieving both economic efficiency and stability. This mechanism ensures that recovery strategies are no longer "one-off and fixed," but rather dynamically adjusted and optimized based on the development of the disaster and the progress of power grid repair. For example, in the initial stages of a disaster, restoring power to hospitals is the highest priority; during the transitional recovery period, the restoration rate of power to residential areas becomes the main indicator; and in the post-disaster stabilization period, reducing operating costs becomes the goal.

[0137] S4: During the actual recovery process, the system monitors grid operation and user demand in real time, performs hierarchical scheduling and iterative correction of the established strategies, dynamically adapts to the evolution of the disaster, and achieves a rapid closed-loop power restoration from initial emergency response to final stability. The monitoring system collects real-time data on grid operation status, flexibility resource status, and user load demand, and compares it with the established recovery strategies. If resource capacity decreases, user demand increases, or local line repairs are completed, the system will trigger an update to the recovery strategy. Through a hierarchical scheduling mechanism, the upper layer is responsible for overall resource coordination across regions, while the lower layer executes specific line commissioning and load switching, ensuring that the recovery strategy can continuously adapt to environmental changes throughout the entire disaster recovery process, achieving closed-loop optimization from initial emergency response to final stability.

[0138] The core of this step lies in establishing a closed-loop mechanism of real-time monitoring, feedback correction, and hierarchical scheduling.

[0139] First, in the real-time monitoring phase, the collected data includes: node voltage. Line flow Energy storage state of charge Actual output of distributed power sources User load demand Etc. Compare these data with the predicted values ​​and calculate the deviation:

[0140] ;

[0141] in, This represents various operating status parameters. When the deviation... Exceeding the set threshold When this happens, the system will automatically trigger the policy correction process.

[0142] Secondly, in the iterative correction stage, this invention employs a rolling time-domain optimization method. It is assumed that the recovery process is divided into multiple time windows. At the end of each window, the objective function is recalculated:

[0143] ;

[0144] in, The strategy will be dynamically adjusted based on the stage of disaster development. If the difference between the new calculation results and the original strategy exceeds a threshold... If so, the scheduling instructions will be updated immediately.

[0145] Secondly, in the hierarchical scheduling stage, the upper-level scheduling center is responsible for global resource coordination, including power allocation of distributed power sources and cross-regional energy storage scheduling; the lower-level control unit is responsible for specific execution, such as switching operations, load switching, and energy storage charging, discharging, starting, and stopping. Its constraints are:

[0146] ;

[0147] Ensure that the local power supply does not exceed the actual available resources.

[0148] Preferably, the present invention also introduces a dynamic adjustment mechanism for user demand. For example, if a temporary increase in the load demand of residential areas is detected during the monitoring process (such as heating or cooling demand under extreme weather conditions), this demand is used as a new input parameter to adjust the recovery priority for the next period, ensuring user satisfaction.

[0149] Through the above mechanism, this step achieves dynamic adaptability in the recovery process. That is, in the early stage of a disaster, the focus of dispatching is on the rigid guarantee of critical loads; in the middle stage of recovery, the dispatching objective gradually shifts to coverage and regional balance; and in the later stage of a disaster, dispatching optimizes the stability and economy of the entire network operation. Ultimately, a closed loop of "real-time monitoring - dynamic correction - hierarchical dispatching - strategy update" is formed, realizing the entire process of rapid power restoration from initial emergency response to final stability.

[0150] This invention also provides a storage medium storing a computer program. When executed by a processor, the computer program implements some or all of the steps in various embodiments of the multi-period rapid power supply restoration method for distribution networks based on flexible resources provided by this invention. The storage medium may be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0151] Those skilled in the art will clearly understand that the techniques in the embodiments of the present invention can be implemented using software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solutions in the embodiments of the present invention, or the parts that contribute to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in various embodiments or certain parts of the embodiments of the present invention.

[0152] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for rapid power restoration in a distribution network across multiple time periods based on flexible resources, characterized in that, Includes the following steps: S1: After a natural disaster occurs, collect meteorological information and historical fault data, and combine them with the characteristics of the power distribution network topology to establish a fault prediction model; S2: Based on the prediction results of the fault prediction model, a unified model is performed on distributed power sources, energy storage and adjustable loads, taking into account capacity, response speed and constraints, and dynamically evaluating the flexible resources that can participate in the recovery. S3: Based on the resource capacity assessment results, construct a three-stage optimization model of critical load priority—regional expansion—full restoration, and formulate time-based power restoration strategies.

2. The method for rapid power supply restoration in a distribution network based on flexible resources according to claim 1, characterized in that, In step S1, the collected meteorological information includes wind speed. Rainfall Snow thickness Ambient temperature Historical fault data includes the frequency of line faults. Average repair time Equipment reliability indicators The distribution network topology is represented by the node set N and the branch set E, where the node load is... The branch road has a carrying capacity of .

3. The method for rapid power supply restoration in a distribution network based on flexible resources according to claim 2, characterized in that, In step S1, establishing the fault prediction model includes: Establish a node failure probability model: ； in, Let be the weighting coefficient, satisfying ; The probability of line damage is determined by a combination of meteorological parameters and historical fault frequency. ； in, branch road Historical failure frequency This is the adjustment coefficient; When predicting the duration of a power outage, the repair time and the severity of the disaster should be considered: ； in, As a disaster intensity indicator, This is the disaster amplification factor; The final risk assessment function is formed as follows: ； in, This represents the importance weight of the nodes.

4. The method for rapid power supply restoration in a distribution network based on flexible resources according to claim 1, characterized in that, In step S2 The distributed power sources include distributed photovoltaic, wind power, and small gas turbine units, whose available power output Due to its maximum installed capacity With resource availability coefficient Joint decision: ； in, Depends on weather conditions; The modeling of the energy storage system focuses on the change process of the state of charge (SOC). The SOC evolution model is as follows: ； in, and These represent charging power and discharging power, respectively. and For charging and discharging efficiency, Let the total energy storage capacity be denoted by ; constraints are introduced as follows: ； The formula for adjustable load modeling is: ； in, This indicates the load reduction ratio, with a value range of [value range missing]. .

5. The method for rapid power supply restoration in a distribution network based on flexible resources according to claim 4, characterized in that, In step S2, the availability vector of the unified flexibility resources is: ； in, This represents the adjustable capacity of energy storage at time t.

6. The method for rapid power supply restoration in a distribution network based on flexible resources according to claim 1, characterized in that, In step S3, the construction of the three-stage optimization model specifically involves, The objective function of the critical load priority restoration phase is to minimize the power outage losses of the critical loads: ； in, Representing the importance weights of different loads, Indicates load The duration of the power outage; The constraints are: ； The objective function of the area expansion and recovery phase is to maximize power supply coverage. ； in, For nodes The power supply status is indicated by a value of 1, which means power has been restored, and a value of 0 means power is still out. The total number of nodes in residential and industrial / commercial areas; The constraints are: ； The objective function of the full recovery and operation optimization phase is to minimize operating costs while taking into account network losses. ； in, These represent the cost coefficients for distributed power generation, energy storage, and network losses, respectively. The constraints are: ； Maintain tidal balance; 。 7. A method for rapid power supply restoration in a distribution network based on flexible resources according to claim 6, characterized in that, In step S3, the time-segmented power restoration strategy introduces rolling time-domain optimization, in each time window Within this framework, the objective function is uniformly expressed as: ； in, The weighting coefficients are for different time periods, initially... Larger capacity to ensure critical loads, medium term Increase to improve coverage, later The dominant approach is to pursue economic efficiency and stability.

8. The method for rapid power supply restoration in a distribution network based on flexible resources according to claim 1, characterized in that, It also includes step S4, which involves real-time monitoring of power grid operation and user demand during the recovery process, iteratively revising and hierarchically scheduling the formulated strategies, dynamically adapting to the evolution of the disaster, and achieving a rapid power restoration closed loop from initial emergency response to final stability.

9. A method for rapid power supply restoration in a distribution network based on flexible resources according to claim 8, characterized in that, Specifically, step S4 is as follows: In the real-time monitoring phase, node voltage is collected. Line flow Energy storage state of charge Actual output of distributed power sources User load demand The data is compared with the predicted values, and the deviation is calculated. ； in, Represents various operating status parameters; when deviation Exceeding the set threshold When this happens, the strategy correction process is automatically triggered; In the iterative correction phase, a rolling time-domain optimization method is adopted, assuming that the recovery process is divided into multiple time windows. At the end of each window, the objective function is recalculated: ； in, The strategy will be dynamically adjusted based on the stage of disaster development; if the difference between the new calculation results and the original strategy exceeds a threshold... If so, the scheduling instructions will be updated immediately; In the hierarchical scheduling process, the upper-level scheduling center is responsible for global resource coordination, including power allocation of distributed power sources and cross-regional energy storage scheduling; the lower-level control unit is responsible for specific execution, and its constraints are as follows: 。 10. A method for rapid power supply restoration in a distribution network based on flexible resources according to claim 8, characterized in that, In step S4, a dynamic adjustment mechanism for user demand is also introduced. If a temporary increase in the load demand of the residential area is found during the monitoring process, the demand will be used as a new input parameter to adjust the recovery priority for the next period.

Images