NSGA-II-based power grid emergency scheduling method and system under extreme disaster

Abstract

The invention discloses an NSGA-II-based power grid emergency scheduling method and system under extreme disasters, and relates to the technical field of power grid emergency management and optimization, and the method comprises the steps: collecting power grid data, building power grid side and user side optimization targets, and constructing a multi-target scheduling model; generating an optimal solution set of the multi-target scheduling model through population initialization, genetic crossover mutation operation and non-dominated sorting; and calculating the satisfaction degree of a multi-target scheme by using a fuzzy membership function, and comprehensively evaluating and selecting an optimal scheme. According to the method, the multi-objective optimization model is constructed, the chromosome coding mode is improved, the fuzzy membership function is introduced, and the number of first-aid repair teams, task allocation and path optimization are subjected to comprehensive scheduling, so that the problems of complex power grid first-aid repair scheduling tasks, low response efficiency and insufficient utilization of first-aid repair resources under the flood disaster condition are solved; and the effects of improving the response speed of the first-aid repair task, optimizing the resource utilization rate and realizing optimal balance in time, resources and task completion degree are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of power grid emergency management and optimization, and in particular to a power grid emergency dispatching method and system under extreme disasters based on NSGA-II. Background Art

[0002] In recent years, with the frequent occurrence of natural disasters and the rapid development of new energy technologies, extreme disasters have posed severe challenges to the safe operation and recovery capabilities of power grids. Extreme disasters are sudden and wide-area, often resulting in damage to power system equipment and large-scale power outages. Existing methods achieve distributed load recovery after disasters by quickly estimating the operating status of the power grid. However, when facing large-scale line damage caused by extreme disasters, this method fails to fully consider the optimization of repair paths and time, and cannot quickly formulate a reasonable repair plan, thus delaying the recovery process. A post-disaster recovery priority evaluation method can also be proposed based on a multi-objective optimization algorithm, considering the degree of damage to power equipment and its importance. However, this method lacks consideration of the actual repair resource allocation and dynamic scheduling issues in extreme disasters, especially in scheduling, it fails to take into account the path optimization of the repair team and the coordination of multi-objective tasks. In addition, this method is not optimized in terms of algorithm complexity, which easily leads to excessive occupation of computing resources and response delays when dealing with large-scale power grid restoration tasks. In addition, there are some existing methods based on neural network learning that have made preliminary explorations on post-disaster power grid scheduling, but they are highly dependent on model data and are prone to overfitting or non-convergence risks in complex task environments. Especially for extreme disaster scenarios, these methods fail to effectively combine the special characteristics of disasters and lack the model design for multi-objective scheduling of emergency repair tasks, resulting in limited response capabilities and insufficient optimization of scheduling schemes in practical applications. Therefore, most existing technologies have explored single dimensions or local problems, lacking comprehensive solutions for power grid emergency dispatch tasks under extreme disaster conditions, especially the design of multi-objective scheduling models and efficient algorithms for emergency repair task allocation and path optimization, and failing to provide a systematic, robust and efficient comprehensive solution. Summary of the invention

[0003] In view of the above-mentioned problems, the present invention is proposed.

[0004] Therefore, the technical problem solved by the present invention is: how to quickly and efficiently formulate a power grid emergency repair scheduling plan under extreme disaster conditions to minimize the emergency repair time and resource costs and maximize the rationality of task allocation and user satisfaction.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions: a power grid emergency dispatching method under extreme disasters based on NSGA-II, which comprises the following steps:

[0006] Collect power grid data to establish optimization goals for the grid and user sides, and build a multi-objective dispatch model;

[0007] The improved NSGA-II algorithm is used to generate the optimal solution set of the multi-objective scheduling model through population initialization, genetic crossover mutation operation and non-dominated sorting;

[0008] The fuzzy membership function is used to calculate the satisfaction of multi-objective solutions, and the best solution that meets the balance of time, resources and tasks is comprehensively evaluated and selected.

[0009] As a preferred solution of the NSGA-II-based power grid emergency dispatch method under extreme disasters described in the present invention, the collection of power grid data includes collecting the number of power outage users in the regional grid after the disaster and the importance of the power outage area.

[0010] As a preferred solution of the power grid emergency dispatch method under extreme disasters based on NSGA-II described in the present invention, wherein: the establishment of the optimization objectives of the power grid side and the user side and the construction of the multi-objective dispatch model include:

[0011] Calculate the basic quantity comprehensive repair time, the expression is:

[0012] T i sy =P i T i s

[0013] Where, T i sy is the comprehensive repair time required to repair node i, P i is the number of repairmen required to repair node i, T i s is the repair time required to repair node i;

[0014] The optimization goal of the power grid side is to minimize the total repair distance of the repair team and construct an objective function to minimize the repair cost, which is expressed as:

[0015]

[0016] Among them, H1 is the optimization target of the power grid side, which represents the total repair distance of the repair team; m xyr is the decision variable, indicating that the power repair team r moves from point x to point y, d xy represents the distance from point x to point y; R is the set of repair teams, and D is the set including power supply companies and damaged points;

[0017] The user-side optimization goal is to minimize the maximum regional repair time, expressed as:

[0018]

[0019] Where, s represents the speed of the vehicle; T i s is the time required for emergency repair of damaged point i; D includes the power supply company and the damaged points in a certain area, D′ is the damaged points in a certain area only, y ir is a decision variable, indicating that damaged point i is served by power repair team r, and u represents the upper limit time of the repair team.

[0020] As a preferred solution of the power grid emergency dispatch method under extreme disasters based on NSGA-II described in the present invention, wherein: the establishment of the optimization objectives of the power grid side and the user side and the construction of the multi-objective dispatch model also include the constraints of the limited nature of the repair team and the continuity of the repair team;

[0021] The constraint condition of the limited nature of the repair team is expressed as:

[0022]

[0023] The continuity constraint of the repair team is expressed as:

[0024]

[0025] Where a is the number of damaged points.

[0026] As a preferred solution of the NSGA-II-based power grid emergency dispatch method under extreme disasters described in the present invention, wherein: the improved NSGA-II algorithm is used to generate the optimal solution set of the multi-objective dispatch model by initializing the population, genetic crossover mutation operation and non-dominated sorting, including:

[0027] Aiming at the problem of optimizing the number and paths of emergency repair teams under extreme disasters, the chromosome encoding method is improved. The entire population is regarded as a solution to the problem. Each individual in the population represents a emergency repair team, and the chromosome in the individual corresponds to the path of the vehicle.

[0028] Generate the initial path segment. The path segment is generated based on the spatial distribution of the repair points and the task priority. The complete path is randomly divided, and the task points are divided into multiple path segments. The generated path segments are distributed to the population to form multiple initial individuals. Each individual corresponds to a repair team. The fitness value of each individual is calculated. The expression for the initial population is:

[0029] P0={X1,X2,...,X N}

[0030] Among them, X i represents the path of vehicle No. i;

[0031] The fitness value is calculated based on the optimization objective using the fitness function.

[0032] As a preferred solution of the power grid emergency dispatch method under extreme disasters based on NSGA-II described in the present invention, wherein: the use of the improved NSGA-II algorithm to generate the optimal solution set of the multi-objective dispatch model by initializing the population, genetic crossover mutation operation and non-dominated sorting also includes:

[0033] Iterate the population to generate new offspring populations, generate new offspring by exchanging some path segments of parent chromosomes, optimize task allocation by reallocating path segments, and dynamically adjust the number of teams and task allocation;

[0034] The path optimization population of parent chromosomes is exchanged, and after randomly selecting a crossover point from the parent chromosomes, the path segment after the crossover point is exchanged;

[0035] Sort the merged parent and child populations to divide different non-dominated frontiers;

[0036] Each individual in the population is compared based on its performance on all objectives. If individual X1 is not inferior to individual X2 in all objectives and is better than X2 in at least one objective, then X1 is said to dominate X2. 2， The population is divided into multiple non-dominated levels. The optimal solution set belongs to the first frontier F1, and the suboptimal solution set belongs to the second frontier F2. The next generation population P is selected from the sorted population through the crowding comparison operator. t+1 , the expression is:

[0037]

[0038] Among them, m represents the fitness function on the target component, m represents the target component, They represent the maximum and minimum values ​​of the target component m respectively.

[0039] As a preferred solution of the NSGA-II-based power grid emergency dispatch method under extreme disasters described in the present invention, wherein: the fuzzy membership function is used to calculate the satisfaction of the multi-objective scheme, and the best scheme that satisfies the time, resource and task balance is comprehensively evaluated and selected, including:

[0040] Based on the fuzzy membership function, the membership of each objective function value is calculated, and each objective function value is mapped to the membership function in the interval [0,1] to obtain the satisfaction degree of each plan on the target. The single target satisfaction degree is aggregated to form the comprehensive satisfaction degree of the candidate plans. The candidate plans are ranked according to the size of the comprehensive satisfaction degree, and the plan with the highest comprehensive satisfaction degree is selected as the optimal solution.

[0041] Another object of the present invention is to provide an emergency dispatch system for power grids under extreme disasters based on NSGA-II, which can solve the problems of insufficient optimization of post-disaster repair paths, low resource scheduling efficiency and poor multi-objective coordination in existing methods by constructing a multi-objective dispatch model, an improved NSGA-II optimization algorithm and a comprehensive evaluation of fuzzy membership function.

[0042] In order to solve the above technical problems, the present invention provides the following technical solutions: a power grid emergency dispatch system under extreme disasters based on NSGA-II, comprising: a power grid emergency dispatch module under extreme disasters, a dispatch solution module based on NSGA-II, and an optimal solution selection module based on membership function;

[0043] The command and dispatch model of the repair team under the power grid emergency dispatch module under extreme disasters includes optimization objectives and constraints. The optimization objectives are divided into grid-side optimization objectives and user-side optimization objectives. The constraints are divided into limited repair team constraints, independent repair team constraints, and continuous repair team constraints. The established model objectives and initial conditions are sent to the dispatch solution module of NSGA-II for optimization calculation.

[0044] The scheduling solution module of NSGA-II includes the initialization population module, the genetic crossover mutation module, the fast non-dominated sorting module and the screening analysis module. In the initialization population module, the traditional chromosome encoding method is improved to solve the number and path optimization problem of emergency repair teams under extreme disasters. The entire population is regarded as a solution to the problem, and each individual in the population represents a repair team. The chromosome in the individual corresponds to the path of the vehicle. The genetic crossover mutation module is used for path optimization and team task allocation optimization in the emergency repair team scheduling model under extreme disasters. The population is iteratively updated through genetic operations to generate new offspring populations. The fast non-dominated sorting module sorts the merged parent and offspring populations and divides different non-dominated frontiers. The screening analysis module selects the next generation of populations from the sorted population through the crowding comparison operator.

[0045] The best solution selection model based on the membership function includes a single-objective satisfaction calculation module, a comprehensive satisfaction calculation module and an optimal solution selection module;

[0046] The single-objective satisfaction calculation module is based on the fuzzy membership function and calculates the membership of each objective function value; the comprehensive satisfaction calculation module aggregates the single-objective satisfaction to form the overall satisfaction of the candidate solutions; the optimal solution selection module sorts the candidate solutions according to the size of the comprehensive satisfaction and selects the solution with the highest comprehensive satisfaction as the optimal solution.

[0047] A computer device includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the above-mentioned method for emergency dispatching of power grids under extreme disasters based on NSGA-II are implemented.

[0048] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the above-mentioned method for emergency dispatching of power grids under extreme disasters based on NSGA-II.

[0049] Beneficial effects of the present invention: The present invention constructs a multi-objective optimization model based on the NSGA-II algorithm, improves the chromosome encoding method and introduces a fuzzy membership function, and comprehensively dispatches the number of emergency repair teams, task allocation and path optimization, thereby solving the problems of complex emergency repair dispatching tasks, low response efficiency and insufficient utilization of emergency repair resources under flood disaster conditions, and achieves the effects of improving the response speed of emergency repair tasks, optimizing resource utilization and achieving the best balance in time, resources and task completion. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them:

[0051] Figure 1 A structural diagram of a power grid emergency dispatch method under extreme disasters based on NSGA-II in a power grid emergency dispatch method under extreme disasters based on NSGA-II provided in a first embodiment of the present invention;

[0052] Figure 2 A structural diagram of a power grid emergency dispatch module under extreme disasters in a power grid emergency dispatch method under extreme disasters based on NSGA-II provided in a second embodiment of the present invention;

[0053] Figure 3 A structural diagram of a dispatch solution module based on NSGA-II of a power grid emergency dispatch method under extreme disasters based on NSGA-II provided in a second embodiment of the present invention;

[0054] Figure 4 A structural diagram of the best solution selection module based on the membership function in the NSGA-II-based extreme disaster power grid emergency dispatch method provided in the second embodiment of the present invention;

[0055] Figure 5A circuit diagram of a repair route of a power grid emergency dispatch method under extreme disasters based on NSGA-II provided in a third embodiment of the present invention;

[0056] Figure 6 The number of vehicles and the grid-side optimization target graph of the grid emergency dispatch method under extreme disasters based on NSGA-II provided in the third embodiment of the present invention;

[0057] Figure 7 The number of vehicles and user-side optimization target graph of the power grid emergency dispatch method under extreme disasters based on NSGA-II provided in the third embodiment of the present invention. DETAILED DESCRIPTION

[0058] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0059] Example 1, reference Figure 1 to Figure 4 As an embodiment of the present invention, a method for emergency dispatching of power grids under extreme disasters based on NSGA-II is provided, comprising:

[0060] Figure 1 It is a structural diagram of the power grid emergency dispatch method under extreme disasters based on NSGA-II. The power grid emergency dispatch method under extreme disasters based on NSGA-II includes: power grid emergency dispatch module 1 under extreme disasters, dispatch solution module 2 based on NSGA-II, and optimal solution selection module 3 based on membership function.

[0061] Collecting power grid data includes collecting the number of users without power under the regional grid after the disaster and the importance of the power outage area.

[0062] After research, it was found that for the city-level prediction area, a 5km×5km grid can reduce the difficulty of sample collection while ensuring the prediction accuracy as much as possible.

[0063] Table 1 Importance level classification

[0064] Number of users experiencing power outage Importance level meaning {0，200} 1 Low risk grid {200，400} 2 More dangerous grid {400，800} 3 Dangerous Grid 800 and above 4 High risk grid

[0065] The command and dispatch model of the repair team under the power grid emergency dispatch module 1 under extreme disasters includes two categories: optimization target 11 and constraint condition 12. The optimization target 11 can be divided into the power grid side optimization target 110 and the user side optimization target 111. The constraint condition 12 can be divided into the repair team finiteness restriction 121, the repair team independence restriction 122, and the repair team continuity restriction 123. Finally, the established model target and initial condition 10 are sent to module 2 for optimization calculation.

[0066] In the case of extreme disasters, in order to effectively carry out emergency dispatch of the power grid, the emergency dispatch module 1 of the power grid under extreme disasters is constructed. Since the power failures caused by extreme disasters are diverse, the emergency repair methods for these failures are also diverse. Different from the conventional vehicle path optimization problem, power emergency repair emphasizes more on the comprehensive emergency repair time, allocating different manpower and material resources to areas of different importance levels, giving priority to better serving the people on the power grid side, and ensuring that sufficient manpower and material resources are allocated to each disaster-stricken point. The basic quantity comprehensive emergency repair time calculation formula of this module is shown in formula (1):

[0067] T i sy =P i T i s (1)

[0068] Where: T i sy is the comprehensive repair time required to repair node i, P i is the number of repairmen required to repair node i, T i s is the repair time required to repair node i.

[0069] In addition, the model mainly includes two major elements: optimization objectives 10 and constraints 11. Specifically, the optimization objective 10 is divided into a grid-side optimization objective 110 and a user-side optimization objective 111. The grid-side optimization objective 110 aims to minimize the total repair distance of the repair team, thereby saving resources such as gasoline and reducing repair costs, as shown in formula (2):

[0070]

[0071] Where: H1 is the optimization target of the power grid side, indicating the total repair distance of the repair team; m xyr is the decision variable, indicating that the power repair team r moves from point x to point y, d xy represents the distance from point x to point y; R is the set of repair teams; D is the set including power supply companies and damaged points.

[0072] The user-side optimization goal 111 focuses on minimizing the maximum regional repair time and ensuring that power supply is restored to each disaster-stricken area as soon as possible, as shown in formula (3):

[0073]

[0074] Where: s represents the vehicle's speed; T i s is the time required for emergency repair of damaged point i; D includes the power supply company and damaged points in a certain area, and D′ includes only damaged points in a certain area.

[0075] At the same time, the module also sets several constraints12. First, the repair team is limited121 to ensure that the manpower and material resources of each team are within a reasonable range, as shown in formula (4).

[0076]

[0077] Where: y ir is a decision variable, indicating that damaged point i is served by power repair team r, and u represents the upper limit time of the repair team.

[0078] Secondly, the independence constraint 122 of the emergency repair team ensures that each damaged point has only one emergency repair team responsible for repair, avoiding waste of resources and duplication of work. In order to ensure that the teams do not affect each other, the independence of the emergency repair team needs to be guaranteed as shown in formula (5).

[0079]

[0080] Finally, the continuity limit of the repair team is 123, which ensures the continuity and efficiency of the repair work, as shown in formula (6).

[0081]

[0082] After setting these goals and constraints, we send the model's goals and initial quantities into module 2 for optimization calculation, and obtain the non-inferior solution set of the Pareto frontier through algorithm solution.

[0083] The scheduling solution module 2 of NSGA-II is a multi-objective optimization model based on genetic algorithm, which aims to solve the optimization problem of emergency repair team scheduling under extreme disaster conditions. Figure 2 The structure of the module is clearly described, including the population initialization module 21, the genetic crossover mutation module 22, the fast non-dominated sorting module 23 and the screening analysis module 24, which together constitute a complete optimization process for solving the scheduling problem.

[0084] In the initialization population module 21, the traditional chromosome encoding method is improved to optimize the number and path of emergency repair teams under extreme disasters. The entire population is regarded as a solution to the problem, and each individual in the population represents a repair team. The chromosome in the individual corresponds to the path of the vehicle. Different from the traditional fixed population size and chromosome encoding method, this improvement can flexibly adapt to dynamic optimization needs, such as adjusting the number of repair teams and replanning the path. At the same time, empty chromosome individuals are introduced to optimize resource allocation and form a diverse initial population.

[0085] In the initialization phase, the initial path segments need to be generated first. The path segments are generated based on the spatial distribution of the repair points and the task priorities. Assume that the task points that need to be repaired are arranged in a certain order, for example: 0 (starting point) → 1 → 2 → 3 → 4 → 5 → 6 → 7 → 8 → 0 (end point). These task points contain all the repair tasks that need to be completed. Subsequently, the complete path is randomly divided into multiple path segments. For example, the above path is randomly divided into two segments: 0 → 1 → 3 → 4 → 5 → 0 and 0 → 2 → 6 → 8 → 7 → 0. The divided path segments maintain integrity while ensuring that each path segment can constitute a valid vehicle driving route.

[0086] Next, the generated path segments are distributed to the population to form multiple initial individuals, each of which corresponds to a repair team. During the distribution process, the balance of the repair tasks is taken into account, and the task points and total task volume of each path segment are relatively balanced. In addition, in order to introduce the ability of dynamic adjustment in subsequent optimization, an empty chromosome individual is allowed to exist in the population. For example, after allocating the path segments, the following initial repair plan can be obtained:

[0087] Vehicle 1 path: 0→1→3→4→5→0

[0088] Vehicle 2 path: 0→2→6→8→7→0

[0089] Vehicle Path 3: Empty

[0090] Finally, the fitness value of each individual is calculated. The fitness function is based on optimization objectives, such as minimizing the repair time, minimizing the total path length, and balancing the task load of each team. On this basis, the initial quality of each individual in the population is evaluated to lay the data foundation for subsequent selection, crossover and mutation operations. Described by formula, the initialization population 210 can be expressed as: population P0 = {X1, X2, ..., X N}, where X i represents the path of vehicle No. i.

[0091] The genetic crossover and mutation module 22 is a key module designed based on the core operation of the genetic algorithm, which is used for path optimization and team task allocation optimization in the repair team scheduling model under extreme disasters. The population 210 is iteratively updated through genetic operations to generate a new offspring population 220. In view of the particularity of the path chromosome, that is, the path points must be unique and all repair teams must cover all points, the traditional crossover and mutation methods are not applicable. Therefore, this module designs a set of improved genetic operations, including transposition, shift, crossover and mutation operations, aiming to ensure the rationality of the solution while improving the quality of the solution and the diversity of the population.

[0092] First, the transposition operation generates new offspring by exchanging some path segments of the parent chromosomes. Specifically, continuous path segments are selected from each parent chromosome for exchange, while the integrity of other path segments is preserved, thereby exploring new solution spaces. For example, as shown in formula (7):

[0093]

[0094] In the formula, P1 and P2 are parent chromosomes, and P'1 and P'2 are daughter chromosomes. This transposition method not only retains the integrity of the path, but also increases the diversity of understanding.

[0095] Secondly, the shift operation optimizes the task allocation by reallocating the path segments. This shifting method can dynamically adjust the number of teams and task allocation, making the solution more flexible. This operation is divided into two cases:

[0096] (1) When both parent chromosomes are not empty, several path segments are randomly selected from one chromosome and inserted into any position of the other chromosome, as shown in formula (8):

[0097]

[0098] (2) When a chromosome is empty, the path segment is directly copied to the empty chromosome to generate a new solution, as shown in formula (9).

[0099]

[0100] Again, the crossover operation further improves the diversity of the population by exchanging the path segments of the parent chromosomes. In the operation, after randomly selecting the crossover point from the parent chromosomes, the path segments after the crossover point are exchanged. This method plays an important role in the recombination of the solution, making the algorithm more likely to jump out of the local optimum. As shown in formula (10):

[0101]

[0102] Finally, based on the traditional method, the mutation operation is improved according to the uniqueness of the path points, and the mutation is achieved by inverting the path segments. The mutation operation can break the local convergence of the population, further enrich the diversity of solutions, and provide more optimization possibilities for the population. For example, as shown in formula (11):

[0103] P1:(1,3,4,5)→P1':(1,4,3,5) (11)

[0104] The fast non-dominated sorting module 23 is an important part of the NSGA-II algorithm. Its main goal is to sort the merged parent and child populations to divide different non-dominated frontiers. Specifically, each individual in the population will be compared based on its performance on all targets. If individual X1 is not inferior to individual X2 on all targets and is better than X2 on at least one target, then X1 is said to dominate X2. Through this dominance relationship, the population is divided into multiple non-dominated levels. The optimal solution set belongs to the first frontier F1, the suboptimal solution set belongs to the second frontier F2, and so on. Fast non-dominated sorting implements sorting operations through an efficient algorithm.

[0105] The screening analysis module 24 selects the next generation population Pt+1 from the sorted population through the crowding comparison operator. The crowding degree is used to measure the distribution density of individuals in the solution space to ensure the diversity of the population in the next generation. The crowding degree calculation formula is shown in formula (12):

[0106]

[0107] Where: f m represents the fitness function on the target component, m represents the target component, Represent the maximum and minimum values ​​of the target component m, respectively. In order to ensure the diversity of solutions, individuals with larger crowding are retained first. The screening process also combines non-dominated rank and crowding to sort to ensure a balance between the quality and distribution of solutions. The final screened population is used for the next generation of iterations until the termination condition is reached, such as the maximum number of iterations or the convergence of the Pareto solution set.

[0108] In summary, the scheduling solution module 2 based on NSGA-II achieves efficient optimization of multi-objective scheduling problems through steps such as population initialization, genetic operations, non-dominated sorting, and screening analysis. Each step of the module strictly follows the basic framework of the genetic algorithm, and further improves the diversity and uniformity of the solution through non-dominated sorting and congestion comparison, providing multiple high-quality optimization solutions for the scheduling problem of extreme disaster repair teams.

[0109] Figure 4The best solution selection module based on the membership function is shown in the structure diagram. The best solution selection module based on the membership function 3 is composed of three core modules: a single-objective satisfaction calculation module 31, a comprehensive satisfaction calculation module 32 and an optimal solution selection module 33. This module combines the fuzzy membership function and finally selects the optimal solution that meets the conditions by analyzing each candidate solution in the multi-objective optimization problem.

[0110] First, the single-objective satisfaction calculation module 31 calculates the membership of each objective function value based on the fuzzy membership function. In multi-objective optimization problems, each candidate solution has different performance for different objectives. By mapping each objective function value to a membership function in the interval [0,1], the satisfaction of each solution on the objective is obtained. Specifically, for a certain objective function value fi(x), the membership function can be calculated by the following formula:

[0111]

[0112] Among them, f i (x k ) represents the solution x k The function value on the i-th target component, f i max and f i min Respectively represent the maximum and minimum function values ​​on the i-th objective component. This process can quantify the performance of each candidate solution on each objective, making it easier for subsequent comprehensive analysis.

[0113] Secondly, the comprehensive satisfaction degree calculation module 32 aggregates the single-objective satisfaction degrees to form the overall satisfaction degree of the candidate solutions. The comprehensive satisfaction degree can be achieved by a weighted fuzzy decision method, and the formula is:

[0114]

[0115] Among them, the solution x k Indicates the satisfaction degree of all optimization objectives u k , that is, x k The satisfaction of all objective components accounts for the proportion of the satisfaction of the objective components of all solutions in the entire Pareto frontier non-inferior solution set. This process fully considers the trade-off relationship between multiple objectives, reflects the importance of each objective through weight distribution, and integrates the performance of multiple objectives into a comparable indicator.

[0116] Finally, the optimal solution selection module 33 selects the optimal solution according to the comprehensive satisfaction level u k The candidate solutions are sorted according to their size, and the solution with the highest comprehensive satisfaction is selected as the optimal solution.

[0117] In summary, the optimal solution selection module 3 based on the membership function effectively realizes the scientific evaluation and optimization of multi-objective solutions in the extreme disaster emergency repair scheduling problem through the collaborative work of three sub-modules. The single-objective satisfaction calculation module 31 provides the basic membership value for comprehensive analysis; the comprehensive satisfaction calculation module 32 integrates the satisfaction of multiple objectives; and the optimal solution selection module 33 provides the final solution for the scheduling problem. This module design ensures the rationality, scientificity and practicality of the solution, and provides theoretical support and practical guidance for actual emergency repair scheduling.

[0118] Embodiment 2 is an embodiment of the present invention, which provides a system for a power grid emergency dispatch method under extreme disasters based on NSGA-II, comprising: a power grid emergency dispatch module 1 under extreme disasters, a dispatch solution module 2 based on NSGA-II, and an optimal solution selection module 3 based on a membership function;

[0119] The command and dispatch model of the repair team under the power grid emergency dispatch module 1 under extreme disasters includes an optimization target 11 and a constraint condition 12. The optimization target 11 is divided into a power grid side optimization target 110 and a user side optimization target. The constraint condition 12 is divided into a repair team finiteness restriction 121, a repair team independence restriction 122, and a repair team continuity restriction 123. The established model target and initial condition 10 are sent to the dispatch solution module 2 of NSGA-II for optimization calculation.

[0120] The scheduling solution module 2 of NSGA-II includes an initialization population module 21, a genetic crossover mutation module 22, a fast non-dominated sorting module 23 and a screening analysis module 24. In the initialization population module 21, the traditional chromosome encoding method is improved to solve the number and path optimization problem of emergency repair teams under extreme disasters. The entire population is regarded as a solution to the problem, and each individual in the population represents a repair team. The chromosome in the individual corresponds to the path of the vehicle. The genetic crossover mutation module 22 is used for path optimization and team task allocation optimization in the emergency repair team scheduling model under extreme disasters. The population 210 is iteratively updated through genetic operations to generate a new offspring population 220. The fast non-dominated sorting module 230 sorts the merged parent and offspring populations to divide different non-dominated frontiers. The screening analysis module 24 selects the next generation population from the sorted population through a crowding comparison operator.

[0121] The best solution selection module 3 based on the membership function includes a single-objective satisfaction calculation module 31, a comprehensive satisfaction calculation module 32 and an optimal solution selection module 33;

[0122] The single-objective satisfaction calculation module 31 calculates the membership of each objective function value based on the fuzzy membership function; the comprehensive satisfaction calculation module 32 aggregates the single-objective satisfaction to form the overall satisfaction of the candidate solutions; the optimal solution selection module 33 sorts the candidate solutions according to the size of the comprehensive satisfaction, and selects the solution with the highest comprehensive satisfaction as the optimal solution.

[0123] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program code.

[0124] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0125] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0126] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, a plurality of steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0127] Embodiment 3: In this embodiment, in order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments. This embodiment experiments the existing traditional method and the method of this embodiment respectively.

[0128] Use Matlab to set the population size of the NSGA-II algorithm to 100, the number of genetic generations to 200, the genetic probability to 0.75 and randomly perform one of the genetic operations, the mutation probability to 0.1, and the vehicle speed to 50km / h.

[0129] Use Matlab to draw a repair route map Figure 5 As shown. Figure 5 It can be seen that this solution distributes the entire repair task by three repair vehicles. Compared with the optimal solution of 1745.54km of total repair route distance and 132.5524h of regional overall repair time on the grid side, this solution uses 8.12% deterioration of the grid side target to improve the user side target by 75.15%; compared with the optimal solution of 3564.644km of total repair route distance and 49.81747h of regional overall repair time on the user side, this solution uses 34.17% deterioration of the user side target to improve the grid side target by 87.63%. Therefore, this solution can better ensure the balance between the two optimization goals and better solve the practical problem.

[0130] The relationship between the number of vehicles and the optimization target of the grid side is as follows: Figure 6 As shown in the figure, the relationship between the number of vehicles and the user-side optimization goal is as follows: Figure 7As shown. The correlation between the number of vehicles and the optimization target on the grid side calculated by Matlab is 0.895, which is significant at the 0.01 level; the correlation between the number of vehicles and the optimization target on the user side is -0.785, which is significant at the 0.05 level. Since the increase in the number of vehicles will increase the distance from the power supply company to the damaged point and reduce the repair time in the area, the number of vehicles is proportional to the optimization target on the grid side and inversely proportional to the optimization target on the user side. Therefore, when the economic environment permits, the number of vehicles can be appropriately increased to better achieve the optimization target on the user side.

[0131] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. The emergency dispatch method of power grid under extreme disasters based on NSGA-II is characterized by: include: Collect power grid data to establish optimization goals for the grid and user sides, and build a multi-objective dispatch model; The improved NSGA-II algorithm is used to generate the optimal solution set of the multi-objective scheduling model through population initialization, genetic crossover mutation operation and non-dominated sorting; The fuzzy membership function is used to calculate the satisfaction of multi-objective solutions, and the best solution that meets the balance of time, resources and tasks is comprehensively evaluated and selected.

2. The method for emergency dispatching of power grids under extreme disasters based on NSGA-II as claimed in claim 1, characterized in that: The collecting of power grid data includes collecting the number of power outage users in the regional grid after the disaster and the importance of the power outage area.

3. The method for emergency dispatching of power grid under extreme disasters based on NSGA-II as claimed in claim 2, characterized in that: The establishment of the optimization objectives of the grid side and the user side and the construction of the multi-objective scheduling model include: Calculate the basic quantity comprehensive repair time, the expression is: In the formula, is the comprehensive repair time required to repair node i, P i is the number of repairmen required to repair node i, is the repair time required to repair node i; The optimization goal of the power grid side is to minimize the total repair distance of the repair team and construct an objective function to minimize the repair cost, which is expressed as: Among them, H1 is the optimization target of the power grid side, which represents the total repair distance of the repair team; m xyr is the decision variable, indicating that the power repair team r moves from point x to point y, d xy represents the distance from point x to point y; R is the set of repair teams, and D is the set including power supply companies and damaged points; The user-side optimization goal is to minimize the maximum regional repair time, expressed as: Where s represents the vehicle's speed; is the time required for emergency repair of damaged point i; D includes the power supply company and the damaged points in a certain area, D′ is the damaged points in a certain area only, y ir is a decision variable, indicating that damaged point i is served by power repair team r, and u represents the upper limit time of the repair team.

4. The method for emergency dispatching of power grid under extreme disasters based on NSGA-II as claimed in claim 3, characterized in that: Based on establishing the optimization objectives on the grid side and the user side, the multi-objective dispatch model is constructed, which also includes the constraints of the limited nature of the repair team and the continuity of the repair team; The constraint condition of the limited nature of the repair team is expressed as: The continuity constraint of the repair team is expressed as: Where a is the number of damaged points.

5. The method for emergency dispatching of power grid under extreme disasters based on NSGA-II as claimed in claim 4, characterized in that: The improved NSGA-II algorithm is used to generate the optimal solution set of the multi-objective scheduling model by initializing the population, genetic crossover mutation operation and non-dominated sorting, including: Aiming at the problem of optimizing the number and paths of emergency repair teams under extreme disasters, the chromosome encoding method is improved. The entire population is regarded as a solution to the problem. Each individual in the population represents a emergency repair team, and the chromosome in the individual corresponds to the path of the vehicle. Generate the initial path segment. The path segment is generated based on the spatial distribution of the repair points and the task priority. The complete path is randomly divided, and the task points are divided into multiple path segments. The generated path segments are distributed to the population to form multiple initial individuals. Each individual corresponds to a repair team. The fitness value of each individual is calculated. The expression for the initial population is: P0={X1,X2,...,X N } Among them, Xi represents the path of the i-th vehicle; The fitness value is calculated based on the optimization objective using the fitness function.

6. The method for emergency dispatching of power grids under extreme disasters based on NSGA-II as claimed in claim 5, characterized in that: The improved NSGA-II algorithm is used to generate the optimal solution set of the multi-objective scheduling model by initializing the population, genetic crossover mutation operation and non-dominated sorting, and also includes: Iterate the population to generate new offspring populations, generate new offspring by exchanging some path segments of parent chromosomes, optimize task allocation by reallocating path segments, and dynamically adjust the number of teams and task allocation; The path optimization population of the parent chromosomes is exchanged, and after the crossover point is randomly selected from the parent chromosomes, the path segment after the crossover point is exchanged; Sort the merged parent and child populations to divide different non-dominated frontiers; Each individual in the population is compared based on its performance on all objectives. If individual X1 is not inferior to individual X2 in all objectives and is better than X2 in at least one objective, then X1 is said to dominate X2. 2， The population is divided into multiple non-dominated levels. The optimal solution set belongs to the first frontier F1, and the suboptimal solution set belongs to the second frontier F2. The next generation population P is selected from the sorted population through the crowding comparison operator. t+1 , the expression is: Among them, m represents the fitness function on the target component, m represents the target component, They represent the maximum and minimum values ​​of the target component m respectively.

7. The method for emergency dispatching of power grids under extreme disasters based on NSGA-II as claimed in claim 6, characterized in that: The method of calculating the satisfaction of multi-objective solutions by using fuzzy membership function, comprehensively evaluating and selecting the best solution that satisfies the balance of time, resources and tasks includes: Based on the fuzzy membership function, the membership of each objective function value is calculated, and each objective function value is mapped to the membership function in the interval [0,1] to obtain the satisfaction degree of each plan on the target. The single target satisfaction degree is aggregated to form the comprehensive satisfaction degree of the candidate plans. The candidate plans are ranked according to the size of the comprehensive satisfaction degree, and the plan with the highest comprehensive satisfaction degree is selected as the optimal solution.

8. A system using the NSGA-II-based emergency dispatch method for power grids under extreme disasters as claimed in any one of claims 1 to 7, characterized in that: It includes a power grid emergency dispatch module under extreme disasters (1), a dispatch solution module based on NSGA-II (2), and an optimal solution selection module based on membership function (3); The command and dispatch model of the emergency repair team under the power grid emergency dispatch module (1) under extreme disasters includes an optimization target (11) and a constraint condition (12). The optimization target (11) is divided into a power grid side optimization target (110) and a user side optimization target (111). The constraint condition (12) is divided into a finiteness restriction of the emergency repair team (121), an independence restriction of the emergency repair team (122), and a continuity restriction of the emergency repair team (123). The established model target and initial condition (10) are sent to the dispatch solution module (2) of NSGA-II for optimization calculation. The scheduling solution module (2) of NSGA-II includes an initialization population module (21), a genetic crossover mutation module (22), a fast non-dominated sorting module (23) and a screening analysis module (24). In the initialization population module (21), the traditional chromosome encoding method is improved to solve the problem of optimizing the number and path of emergency repair teams under extreme disasters. The entire population is regarded as a solution to the problem, and each individual in the population represents a emergency repair team. The chromosome in the individual corresponds to the path of the vehicle. The genetic crossover mutation module (22) is used for path optimization and team task allocation optimization in the emergency repair team scheduling model under extreme disasters. The population (210) is iteratively updated through genetic operations to generate a new offspring population (220). The fast non-dominated sorting module (230) sorts the merged parent and offspring populations to divide different non-dominated frontiers. The screening analysis module (24) selects the next generation population from the sorted population through a crowding comparison operator. The optimal solution selection module (3) based on the membership function includes a single-objective satisfaction degree calculation module (31), a comprehensive satisfaction degree calculation module (32) and an optimal solution selection module (33); The single objective satisfaction calculation module (31) calculates the membership of each objective function value based on the fuzzy membership function; the comprehensive satisfaction calculation module (32) aggregates the single objective satisfaction to form the overall satisfaction of the candidate solutions; the optimal solution selection module (33) sorts the candidate solutions according to the size of the comprehensive satisfaction and selects the solution with the highest comprehensive satisfaction as the optimal solution.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method for emergency dispatching of power grid under extreme disasters based on NSGA-II described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for emergency dispatching of power grid under extreme disasters based on NSGA-II according to any one of claims 1 to 7 are implemented.

Images