CN-121998175-A - Multi-disaster coupled modeling-based multi-scale toughness assessment method for power distribution network

Abstract

The invention discloses a multi-disaster coupled modeling-based multi-scale toughness assessment method for a power distribution network, relates to the field of power system optimization scheduling, and particularly relates to a deep reinforcement learning-based active power distribution network toughness improvement method. The method comprises the following steps of (1) establishing a typhoon wind field model of a region where the power distribution network is located, (2) establishing a typhoon path model of the region where the power distribution network is located, (3) simulating a fault scene of the power distribution network, and (4) carrying out load reduction based on the importance of loads. (5) The method comprises the steps of (1) constructing a toughness evaluation index system, (6) calculating weights of all levels of toughness indexes based on fuzzy entropy, and (7) self-adaptive combination weighting fusion.

Inventors

• LI YAO
• HU TAIYUAN
• JIE YIMING
• KONG LINGYUAN
• WANG CHONG
• WANG QUN
• HAN SONG
• WANG LONG

Assignees

• 国网山东省电力公司枣庄供电公司

Dates

Publication Date

20260508

Application Date

20260108

Claims (10)

1. 1. A multi-scale toughness assessment method for a power distribution network based on multi-disaster coupling modeling is characterized in that a multi-disaster coupling modeling mechanism is introduced into the toughness assessment of the power distribution network, multiple influence factors such as typhoons, storm, terrains, equipment health states and the like are comprehensively considered, and space-time continuous expression of disaster intensity is realized by improving the linkage modeling of a wind field model, a lognormal rainfall model and a space coupling propagation matrix. The fuzzy entropy objective weight is organically fused with the Bayesian subjective weight, and the self-adjustment of the weight along with the disaster evolution is realized through a time change adjusting factor. The method overcomes the limitations of static state and rigidification of the traditional combined weighting method, can automatically adjust index weight according to disaster stages, and realizes dynamic evolution and intelligent adaptation of a toughness evaluation system.
2. 2. A multi-disaster coupled modeling-based multi-scale toughness assessment method for a power distribution network is characterized by comprising the following steps of: S1, establishing a typhoon wind field model of an area where a power distribution network is located, reasonably assuming that typhoon intensity is kept unchanged in the time range, and simulating wind speeds of all points in the influence range of the typhoon wind ring by utilizing superposition of gradient wind speeds in the wind ring and typhoon moving speeds based on the batts wind field model; s2, establishing a typhoon path model of a region where the power distribution network is located, fully considering the geographical position where the power distribution network is located and historical typhoon path information, and constructing a typhoon disaster path model by adopting a method of combining typhoon historical path data probability distribution fitting and Monte-Carlo random sampling; S3, simulating a fault scene of the power distribution network based on multi-disaster coupling modeling, synthesizing coupling relations of typhoons, rainfall, terrains and equipment health states, randomly simulating the fault scene of the power distribution network under extreme events through model Carlo sampling, and solving selected fault characteristics of the power distribution network by utilizing mathematical statistics; And S4, carrying out load reduction based on the importance degree of the load, and classifying the loads of each node into primary, secondary and tertiary loads through importance degree analysis of the loads of the power distribution network and different requirements of the loads of different grades on the power supply reliability. Determining load shedding schemes for different levels of loads respectively; S5, selecting toughness evaluation indexes, namely layering the evaluation indexes by using a hierarchical analysis method according to the complex decision-making problems of multiple criteria, wherein the change of the internal power supply capacity of the power distribution network and the regional characteristics of external disasters are considered when disasters occur, and establishing a hierarchical toughness evaluation index system; S6, calculating objective weights based on fuzzy entropy, and quantifying index information by adopting a fuzzy entropy method and calculating the objective weights in order to comprehensively reflect the objective difference and uncertainty of the toughness evaluation index, wherein the fuzzy entropy method can more accurately describe the information contribution degree of the index when the sample data has fuzzy, uncertainty or nonlinear characteristics; and S7, self-adaptive combined weighting fusion, wherein the importance of each index is dynamically adjusted along with the change of time and disaster stage under the condition of complex disasters, and the self-adjustment of the weight along with the stage is realized by combining real-time data in the disaster evolution process.
3. 3. The multi-scale toughness assessment method for a power distribution network based on multi-disaster coupling modeling according to claim 1, wherein in step S1, wind field distribution of a power distribution network area is firstly constructed by utilizing Batts wind field models, as a research object is an area-level power distribution network, the duration of the influence of typhoons is relatively limited, so that the typhoons strength can be reasonably assumed to be kept constant in the period, the wind field models are used for describing the wind speed distribution of each position in the range of the typhoons by superposing the gradient wind speed in the wind ring and the moving speed of the typhoons, and friction effects are not considered in the process, wherein the calculation formula of the gradient wind speed is as follows: Wherein: K is an empirical coefficient, and the value is 6.72; Is the center differential pressure (hpa); maximum wind speed radius (km) for typhoons; Taking a geographic latitude of 30 DEG for the earth rotation Coriolis force coefficient, and based on the geographic latitude, obtaining a gradient wind speed And typhoon movement speed The average maximum wind speed in typhoon wind circle within 10 minutes at the position with the sea surface height of 20m can be obtained by superposition The calculation formula is as follows: further, when any point on the sea surface is at a distance r from the typhoon center, the typhoon wind speed at the position is: Wherein: an average wind speed (km/h) of a typhoon center distance r from any point in a typhoon wind field on the sea surface; The average maximum wind speed in the typhoon wind ring is the average maximum wind speed; 7r is the distance (km) from the research point to the typhoon center, and when the typhoon logs in, the reduction effect of the terrain roughness (such as mountainous regions, basins or building groups) on the wind speed needs to be considered, so that the terrain correction coefficient is introduced into the model to correct the wind speed, and the calculation formula of the average wind speed at the position of 10m on land and 10 minutes is obtained: Wherein: average wind speed (km/h) for any point in the land typhoon farm; the method is characterized in that the method is used for obtaining the average wind speed (km/h) of any point in a typhoon wind field on the sea surface, p is an obstacle factor, 0.85 is taken, z is the ground roughness length, and as urban power distribution networks are researched, 1 is taken according to the roughness classification standard z.
4. 4. The multi-scale toughness assessment method for the power distribution network based on multi-disaster coupling modeling according to claim 1 is characterized in that in step S2, a typhoon path model of a research area needs to be established, and the model is used for reflecting the motion characteristics of historical typhoons and considering the geographic condition of the position of a target power distribution network. The specific process is as follows: (1) Setting simulation area, namely defining radius by taking research point as circle center Is used to determine if typhoons will have an impact on the study site. When the minimum distance between the research point and the typhoon path Smaller than the radius of the simulated circle If the typhoon is not considered to affect the research point, the typhoon is not considered to affect the research point; (2) Historical data screening and preprocessing, namely, only keeping typhoons, strong typhoons and super typhoons with the wind power level exceeding 12 levels during login, and ensuring that parameter values are in a reasonable range in order to ensure the rationality of probability distribution fitting; (3) Carrying out statistical modeling and distribution fitting on key parameters such as the moving speed, wind direction, minimum distance between a research point and a typhoon path and the like of typhoons; (4) Typhoon path generation, namely extracting a group of parameters (moving speed, wind direction and minimum distance) from the fitted probability distribution by random Monte-Carlo sampling to construct a virtual typhoon path; (5) Multipath expansion, namely if a plurality of virtual typhoon paths are required to be generated, repeating sampling and construction processes, so as to form a typhoon path set of multiple scenes; the typhoon path model established through the steps can introduce uncertainty modeling while taking historical statistical characteristics into consideration, and lays a foundation for subsequent disaster scene simulation.
5. 5. The multi-disaster coupled modeling-based multi-scale toughness assessment method for the power distribution network, as set forth in claim 1, is characterized in that in step S3, a multi-disaster coupled model is established considering typhoons, storm, topography and equipment health status. The method comprises the steps of constructing a probability model for describing disaster influence, describing that an active power distribution network is subjected to multidimensional influence in typhoon and storm composite disasters, wherein the typhoon generates high wind speed to cause mechanical damage to overhead lines and pole tower structures, the storm causes the rise of soil humidity to cause the change of ground resistance and the insulation degradation of equipment, the topography and topography characteristics influence the distribution of a wind and rain field and the disaster propagation path, the equipment aging and health state determine disaster susceptibility difference and the like, the parameters represent the running state and the failure probability of power distribution network elements, and a multi-disaster coupling intensity function is introduced Comprehensively characterizing the node at the moment Disaster impact strength: Wherein, the Is a node At the wind speed of the wind, the wind speed is controlled, Is a node The intensity of rainfall at the position is calculated, Is a node The exposure coefficient of the terrain, For the device health index ([ 0,1], smaller and more fragile), , , , Is disaster factor weight coefficient, meets the following conditions ; In the modeling process of disasters, the rain of old donkey can cause equipment to soak and the insulation performance to deteriorate, and a corrected lognormal rainfall model is adopted to describe rainfall intensity distribution of different areas: Wherein, the Is the maximum rainfall intensity; the distance from the node to the storm center is set; As a time-dependent disturbance factor for reflecting the change of rainfall intensity with time, the degradation rate of insulation caused by rainfall is expressed as: Wherein the method comprises the steps of For the equipment burn-in rate under normal conditions, Is the rainfall sensitivity coefficient.
6. 6. The method for evaluating the multi-scale toughness of the power distribution network based on the multi-disaster coupling modeling according to claim 1, wherein the method for calculating the node failure rate and the equipment accumulated damage model is improved to achieve the health degree of the equipment Reflecting the residual performance of the device under the action of disasters, and defining a degradation differential equation according to the historical operation and maintenance data of the device and the stress level of the disasters: Wherein the method comprises the steps of A coefficient of weakness that is device type dependent. By integration it is possible to obtain: When (when) When the equipment is regarded as failure, the failure rate of the node or the line Integrated intensity with disaster The logic relationship is as follows: Wherein the method comprises the steps of As a spatial coupling term, used to describe the neighbor node fault conduction effect: Wherein the method comprises the steps of As a factor of the spread of the disaster, And (5) the delay influence time of the fault of the node j on the node i is obtained.
7. 7. The multi-scale toughness assessment method for the power distribution network based on multi-disaster coupling modeling according to claim 1, wherein disaster scenes are generated based on disaster space-time coupling characteristics, coupling relations of typhoons, rainfall, terrains and equipment health states are synthesized, a Monte Carlo sampling method is adopted to generate a multi-disaster scene set, each scene comprises wind speed, rainfall intensity, failure rate and equipment states under different time sequences, and comprehensive disaster intensity distribution of each scene can be expressed as: And generating random numbers through a computer, judging whether the elements are faulty according to probability distribution, so as to obtain the running state of the power distribution network at a certain moment, and if the total number of the elements is N, the running state can be expressed as a vector: operational status of each element The judgment can be carried out by uniformly distributing the random number r and the fault probability p: And repeating the process for M times by sampling for multiple times to obtain M power distribution network system state samples, namely, multiple fault scenes possibly occurring under extreme disasters: And judging convergence and precision, namely gradually approaching a simulation result to a true value along with the increase of sampling times according to a law of large numbers. In combination with the central limit theorem, a variance coefficient can be used as a convergence criterion: And (3) topology analysis, namely when the sampling result shows that the element fails to work, the topology structure of the power distribution network needs to be recalculated, and whether connectivity and power flow distribution change or not is judged. The distribution network consists of nodes and switch circuits, topology analysis can be performed by using a graph theory method, and node accessibility and system partition conditions are verified. When the sampling obtains that the state of a system element is a fault, the element is out of operation, and at the moment, the connectivity of the original power distribution network can be changed, so that the power flow distribution of the power distribution network system is affected, therefore, the topology analysis of the system after the fault is needed, and the node number of the current system and the connection condition of each node are judged. The power distribution network consists of nodes and switch circuits, the topology analysis can be carried out on the power distribution network by using the knowledge of graph theory, the searching of the graph is often needed in the problem of solving the graph theory, namely, from a certain vertex in the graph, the rest vertices of the graph are sequentially visited, each vertex is visited at most once, the basic idea is that one vertex is selected and marked, then the adjacent point is searched and marked, and then the adjacent point of the adjacent point is continuously searched until all the vertices in the graph are marked.
8. 8. The multi-scale toughness assessment method for the power distribution network based on multi-disaster coupling modeling according to claim 1 is characterized in that in step S4, a load importance concept is introduced to design reduction strategies, the load of different nodes has different dependence on the power supply reliability of the system, therefore, the power distribution network needs to be graded according to the importance, and a reasonable load reduction scheme is formulated based on grading results. And then, according to the grading result, corresponding reduction strategies are formulated for loads of different grades: Dividing node loads into three stages according to requirements on power supply continuity: The primary load requirement must in any case ensure a continuous supply. Typical application environments include: 1) Once power failure can directly cause life safety accidents, such as secondary and higher hospitals; 2) Industries in which a power outage may cause serious politics or economic losses, such as product processing based on important raw materials, steel smelting, fire-fighting power supply for rocket launching bases and high-rise buildings, and the like; 3) Important electric sites with great influence on political economy, such as important transportation hubs, star hotels, large conference centers, important public facilities and the like; And a power supply scheme aiming at the primary load, such as dual power supply, special line power supply and the like. Meanwhile, the application of the distributed power supply provides another convenient means for guaranteeing the continuous power supply of the loads, when the power distribution network fails, the distributed power supply can rapidly supply power for the first-level load within the capacity range, and although the scheme has a little shortage in economy, the investment of power distribution network equipment can be remarkably saved, and the comprehensive benefit is still better after the scheme is optimized; the secondary load requires that the power supply be maintained as much as possible, but in extreme cases some interruption may be tolerated. Common scenarios include: 1) Enterprises, such as continuous production enterprises, that can have a significant impact on politics or economics if a power outage occurs; 2) Sites where blackouts can interfere with normal operation of important public units, such as transportation hubs, large malls, large stadiums, theatres, and the like; For this type of load, a dual-loop power supply and a dual-transformer power supply are employed. Meanwhile, the distributed power supply can also be used as a supplementary measure. When the power distribution network fails, the distributed power supply can provide power support for part or even all of the secondary loads in time under the condition of capacity permission; The three-stage load has no strict requirement on power supply continuity. Typical applications include rural domestic electricity and electricity for most rural enterprises, and common loads which are not primary and secondary loads. For the load, the distributed power supply can be used as a random power supply to supply power so as to fully exert the economic, social and environmental benefits. It is emphasized that the load classification is relative, the local power conservation level is combined, and the political effect and the economic effect are considered for reasonable division; Because the toughness of the distribution network often shows the supporting and recovering capacities of key loads, different grades of loads are required to be endowed with different weights during evaluation, wherein the primary load is assigned with 6, the secondary load is 3, and the tertiary load is 1, and the influence of various loads in toughness evaluation can be reflected more intuitively through the differential weights; When the system fails and overload of the line is found through topology analysis, load shedding operation must be performed to restore safe and stable operation of the power distribution network, and in this process, the goal of shedding should be to reduce the load which is shed as much as possible on the premise of meeting the safety of the system, and to ensure that the critical load is not shed preferentially. The method comprises the following specific steps: 1) In the power distribution network topology after the fault, layering processing is carried out on each branch from a power supply node, and the branches are searched one by one according to a hierarchical order; 2) When one line is found to be overloaded, the overload power of the line is determined firstly; 3) And selecting proper load shedding combinations according to the load grades, and ensuring that the total load shedding amount is not smaller than the overload power. When the reduction is executed, the principle of 'from low to high' is followed, namely, firstly cutting off the three-level load, and then sequentially reducing the two-level load if the three-level load is still insufficient to eliminate the overload, and finally considering the first-level load; 4) And (3) continuing searching the system according to the topology level, and if overload exists in other lines, repeating the step (2) and the step (3) until all lines of the system are restored to a safe operation state.
9. 9. The multi-scale toughness assessment method for the power distribution network based on multi-disaster coupling modeling according to claim 1, wherein in step S5, a toughness assessment index system is constructed, and a set of toughness assessment index system is constructed. The system is used for reflecting dynamic change of internal power supply capacity of the power distribution network after disasters occur and reflecting regional characteristics of external disasters. In order to deal with such complex decision-making problems involving multiple criteria, the present invention employs Analytic Hierarchy Process (AHP) to build a multi-scale toughness index framework by layering the indices. The whole index system is divided into three levels: (1) Target layer, i.e. evaluating the overall toughness level of the distribution network (2) A first-level index layer, which is a framework for forming a toughness evaluation system and covers core indexes with different dimensions (3) A second index layer for refining to specific measurement parameters TABLE 1 toughness assessment index framework The specific content of the index system is as follows: (1) The power supply capacity of the power distribution network comprises three aspects, namely a system function curve missing area, disaster absorptivity and post-disaster recovery rate, wherein the system function curve missing area can be measured by drawing a dynamic function curve before disaster, during disaster and after disaster, and a load importance index is defined as a load importance weight at the current moment t And the amount of load loss The product of (2) is calculated as follows: The real-time function curve and system function curve missing area that can be considered for load importance can thus be expressed as: The disaster absorption rate index of the power distribution network represents the ratio of the load level which can be maintained by the system to the initial load level before disaster during disaster occurrence, and is used for representing the capability of the power distribution network for resisting external impact, and the calculation formula is as follows: Wherein: a load level that can maintain continuous power supply in a disaster; Is the initial load level of the system; the post-disaster recovery rate of the power distribution network measures the speed of the power distribution network to recover to an initial power supply level after disaster, and the calculation formula is as follows: Wherein: the loss value of the load level of the power distribution network in the disaster is obtained; From load loss to distribution network recovering the time length of power supply of all loads; (2) The disaster influence degree of the power distribution network is mainly described by the following three aspects, namely the occurrence frequency of typhoon disasters, the average typhoon speed and the typhoon influence duration of the power distribution network, wherein the calculation formula of the typhoon influence duration T is as follows: Wherein: Simulating a circle radius in the typhoon model; Is the minimum distance from the geographical center point of the distribution network to the typhoon path.
10. 10. The multi-scale toughness assessment method for the power distribution network based on multi-disaster coupling modeling according to claim 1, wherein in step S6, in order to comprehensively reflect objective differences and uncertainties of toughness assessment indexes, a fuzzy entropy method (Fuzzy Entropy Method) is adopted to quantify index information and calculate objective weights. Compared with the traditional entropy weight method, the fuzzy entropy method can more accurately describe the information contribution degree of the index when the sample data has ambiguity, uncertainty or nonlinear characteristics, and the specific method is as follows: 1) Constructing a data matrix Is provided with Individual evaluation index The raw data of each index can be represented by a matrix, and the rows of the matrix represent the number of the evaluated objects Columns represent the number of evaluation indicators : 2) Data normalization For data matrix Performing standardization processing to obtain a standardized data matrix The processing mode of each index can be selected according to the type of the index, and the specific method is as follows: Profitability index: cost index: 3) Membership function calculation, namely mapping each standardized index value into 'membership degree to ideal state' to reflect uncertainty of a sample, wherein the membership degree is defined as: at the same time, the influence of index ambiguity is considered, and the adjustment parameters are introduced The fuzzy recognition sensitivity of the index is represented, and the improvement membership degree is as follows: Wherein the method comprises the steps of And the mean membership degree of the j index. This process helps mitigate the distortion effects of the extreme samples on the overall weight distribution; 4) Fuzzy entropy value calculation The information uncertainty of the j index is reflected, and the calculation formula is as follows: When the index sample has large difference and high information content, Is widely distributed in the [0,1] interval, resulting in Smaller, when the sample difference is smaller and the information amount is low, The distribution is concentrated, Larger; 5) Objective weight calculation The fuzzy entropy weight is obtained by normalizing the effective information amount: the weight reflects objective contribution degree of each index to system toughness change, and can effectively avoid the problem of excessive sensitivity of a traditional entropy weight method to a polar sample. Calculated and obtained Objective benchmarks for subsequent adaptive combination weighting fusion; Under complex disaster conditions, the importance of each index can be dynamically adjusted along with the change of time and disaster stages, so that a self-adaptive combined weighting algorithm based on fuzzy entropy-Bayesian fusion is provided, the self-adjustment of weight along with the stages is realized by combining real-time data in the disaster evolution process, and a multi-scale toughness dynamic evaluation framework is constructed, and the specific flow is as follows: 1) Bayesian subjective weight dynamic update: Wherein: And in the disaster stage t, the interpretation probability of the j index on the system toughness change is determined. Subjective weight of the previous stage; the formula enables the index weight to be adaptively updated according to the real-time observation data of the disaster situation. For example, if the change in the recovery rate of the system function curve during the disaster phase has a significant effect on the toughness result, the index The number of the cells to be processed is increased, Automatically ascending to embody a dynamic subjective correction mechanism; 2) And (5) self-adaptive weight fusion: To comprehensively consider fuzzy entropy weight and subjective Bayesian weight of objective data, a time variation adjustment factor is defined : Wherein: The standard deviation of fluctuation of the index j in a time window is shown; the standard deviation of average fluctuation of all indexes, when the fluctuation of the indexes is strong (the information change is fast), Increasing, the model trend adopts objective entropy weight, when the index changes slowly or is dominant empirically, The subjective Bayesian weight is reduced and is more depended; 3) And (3) calculating a combination weight: The final adaptive combining weights are defined as: The beneficial effects of the invention are as follows: compared with the prior art, the multi-scale toughness assessment method for the power distribution network based on multi-disaster coupling modeling has the following advantages and positive effects: According to the invention, a multi-disaster coupling modeling mechanism is introduced in the toughness evaluation of the power distribution network, and a plurality of influence factors such as typhoons, storm, topography, equipment health states and the like are comprehensively considered, so that the space-time continuous expression of disaster intensity is realized by improving the linkage modeling of a wind field model, a lognormal rainfall model and a space coupling propagation matrix. The model can reflect the local destructive action of a single disaster and reveal the node failure probability evolution and regional fragile propagation characteristics under the multi-disaster composite impact. The index information quantity is calculated by adopting a fuzzy entropy method, and the membership function and the fuzzy recognition factor are introduced, so that the calculation stability can be kept when the data has ambiguity, uncertainty or nonlinear distribution. The method can more accurately describe the information contribution degree of different toughness indexes under the disaster condition, remarkably improves the robustness and the interpretability of weight calculation, and provides a solid objective foundation for index evaluation under complex disaster scenes. And meanwhile, the fuzzy entropy objective weight and the Bayesian subjective weight are organically fused, and the self-adjustment of the weight along with the disaster evolution is realized through a time change adjusting factor. The method overcomes the limitations of static state and rigidification of the traditional combined weighting method, can automatically adjust index weight according to disaster stages, and realizes dynamic evolution and intelligent adaptation of a toughness evaluation system; In conclusion, the invention can obviously improve the risk resistance and recovery level of the power distribution network under the extreme disaster condition while ensuring the technical feasibility, and has higher practical value and application prospect.

Description

Multi-disaster coupled modeling-based multi-scale toughness assessment method for power distribution network Technical Field The invention relates to the field of power system optimization scheduling, in particular to a multi-scale toughness assessment method for a power distribution network based on multi-disaster coupling modeling. Background With the frequent occurrence of extreme events such as natural disasters, terrorist attacks, and serious technical faults, the power system faces unprecedented challenges in terms of safety and stability. The traditional evaluation means often lack comprehensiveness and dynamics when dealing with disaster impact, and the fragile links of the power distribution network under extreme conditions are difficult to accurately reveal. For this problem, the toughness assessment strategy is particularly important. The method can provide a systematic reference frame for analyzing the influence of natural disasters on the operation of the power distribution network, ensure the evaluation of key elements in the operation of the power distribution network, and quantitatively characterize the key elements. Through scientific index screening and design, the accuracy of the result can be improved while the objectivity is ensured, so that more valuable technical support is provided for decision departments. In addition, a reasonable evaluation system not only can identify potential weak points of the power distribution network in disaster defense and recovery processes, but also can indicate directions for improving the overall toughness of the system and optimizing emergency management means. The method comprises the core ideas of establishing a power distribution network scene model under typhoon disasters, including a wind field and a path model, estimating the failure rate of system elements by utilizing Monte Carlo sampling on the basis, further simulating different fault scenes, simultaneously carrying out grading reduction by combining the importance of loads, constructing a multi-scale toughness index system, and finally carrying out comprehensive evaluation by combining a weighting method, thereby quantifying the adaptability and the recovery capability of the power distribution network under extreme disaster conditions. Disclosure of Invention The invention aims to provide a multi-scale toughness assessment method for a power distribution network based on multi-disaster coupling modeling. By introducing multidimensional evaluation indexes, the change characteristics of the power supply capacity of the power distribution network when disasters occur are considered, and a comprehensive evaluation model capable of reflecting the global toughness level of the system is established by combining regional differences of external environments. The method utilizes a combined weighting mechanism to effectively balance the correlation and objectivity among indexes, thereby improving the scientificity and reliability of the evaluation result. Compared with the prior art, the method and the device can more comprehensively describe the dynamic influence of disaster impact on the power distribution network, can provide a quantitative basis for toughness improvement and emergency decision, and finally enhance the defending and recovering capabilities of the power distribution network in the face of extreme disasters. The technical scheme is that the method comprises the following specific steps: s1, establishing a typhoon wind field model of the region where the power distribution network is located, reasonably assuming that the typhoon intensity is kept unchanged in the time range, and simulating the wind speed of each point in the influence range of the typhoon wind ring by utilizing superposition of the gradient wind speed in the wind ring and the typhoon moving speed based on the batts wind field model. And S2, establishing a typhoon path model of the region where the power distribution network is located, fully considering the geographical position where the power distribution network is located and historical typhoon path information, and constructing a typhoon disaster path model by adopting a method of combining typhoon historical path data probability distribution fitting and Monte-Carlo random sampling. And S3, simulating a fault scene of the power distribution network based on multi-disaster coupling modeling, synthesizing coupling relations of typhoons, rainfall, terrains and equipment health states, randomly simulating the fault scene of the power distribution network under extreme events through model Carlo sampling, and solving the selected fault characteristics of the power distribution network by utilizing mathematical statistics. And S4, load reduction is performed based on the importance of the load. The loads of all nodes are classified into primary, secondary and tertiary loads through importance analysis of the loads of the power distribution network and different req