Micro-grid planning optimization method based on robust evolutionary computation

Abstract

The invention discloses a micro-grid planning optimization method based on robust evolutionary computation, and the method comprises the following steps: collecting real-time and historical data to generate a plurality of typical scenes, the data including load, power generation, energy storage state and equipment fault data; constructing a multi-objective optimization model, wherein objective functions of the multi-objective optimization model comprise an economic objective, an environment objective and a robustness objective; solving the multi-objective optimization model by adopting a self-adaptive multi-objective genetic algorithm; obtaining a Pareto optimal solution set, and generating a recommended planning scheme; and in the process of solving the multi-objective optimization model, predicting a low-efficiency solution region in an evolution path by using a large model and dynamically adjusting a search strategy. According to the method, robust evolutionary computation is combined with a large model technology, so that the problem of complex, multi-target and high-uncertainty micro-grid planning design is solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of power grid planning, and particularly relates to a method for optimizing microgrid planning based on robust evolutionary computation. Background Art

[0002] With the transformation of the global energy structure and the increasing demand for sustainable development, the microgrid, as a distributed power system, has become an important part of the modern power network. The microgrid can not only effectively integrate renewable energy resources (such as wind energy, solar energy, etc.), but also provide power balance and stability through energy storage technology to meet the power demand of specific regions or communities. At the same time, the microgrid can operate independently when disconnected from the main grid, improving the reliability and resilience of the energy system, so it has a broad application prospect in the power system.

[0003] However, despite the significant advantages of the microgrid, its planning and design still face a series of challenges, including uncertainty factors and multi-objective optimization problems:

[0004] The power generation of renewable energy in the microgrid has a high degree of uncertainty, and the load demand fluctuates greatly. In addition, there is a risk of performance degradation and failure of microgrid equipment, and these factors will greatly affect the operation stability and economy of the microgrid. How to effectively cope with these uncertainties and ensure the stable operation of the microgrid under extreme conditions has become an important problem in the design.

[0005] The design of the microgrid usually involves multiple conflicting objectives, including economy, environment, and robustness. How to balance these objectives, ensure the economic benefits of the system, achieve environmental protection goals, and at the same time ensure high reliability in the face of uncertainties is the core issue in microgrid planning. Summary of the Invention

[0006] In view of this, the objective of the present invention is to propose a method for optimizing microgrid planning based on robust evolutionary computation, and the microgrid planning optimization method includes the following steps:

[0007] Step 1, collecting real-time and historical data to generate multiple typical scenarios, and the data includes load, power generation, energy storage status, and equipment failure data;

[0008] Step 2, constructing a multi-objective optimization model, and the objective function of the multi-objective optimization model includes economic objectives, environmental objectives, and robustness objectives;

[0009] Step 3, using an adaptive multi-objective genetic algorithm to solve the multi-objective optimization model;

[0010] Step 4, obtaining the Pareto optimal solution set and generating a recommended planning scheme;

[0011] In the process of solving the multi-objective optimization model, a large model is used to predict the inefficient solution region in the evolutionary path and dynamically adjust the search strategy.

[0012] Specifically, the use of the large model to predict the inefficient solution region in the evolutionary path includes the following steps:

[0013] Input the fitness distribution of the population, the change of the objective function value, and the distribution of the solution space into the large model. Let \(P\) be the current population, which contains \(N\) individuals \(p_1, p_2, \cdots, p_N\). N , \(F(P)\) is the fitness value of each individual in the population, and \(\Delta F(P)\) is the change of the current objective function value, that is, the difference and fluctuation of the objective function value. \(\Omega\) is the current solution space.

[0014] By calculating the variance of the fitness of individuals in the population, judge the convergence degree of the population. If the variance of the fitness is less than the preset threshold, it means that the population tends to converge. The fitness convergence formula is: where \(F(p_i)\) is the fitness of individual \(p_i\). i \(i\) i and \(\overline{F}\) is the mean value of the fitness of the current population.

[0015] Calculate the change trend of the objective function value between different generations. If the improvement amplitude of the objective function value is less than the preset threshold, it means that the exploration of the current solution space has tended to be saturated.

[0016] Analyze the coverage of the current population in the solution space, calculate the solution space coverage. The solution space coverage formula is: where \(\delta(p_i, \Omega)\) is the indicator function, which takes the value of 1 when the individual \(p_i\) falls into different regions of the solution space, and 0 otherwise. i

[0017] Based on the analysis results of fitness convergence, the change trend of the objective function value, and the solution space coverage, the large model predicts and identifies the inefficient solution region.

[0018] The dynamic adjustment of the search strategy includes the following steps:

[0019] The large model transfers the search direction from the inefficient solution region to the region that has not been fully explored to increase the breadth and depth of the search space.

[0020] If the diversity of the current population is low, increase the crossover rate to promote the exploration of the solution space. Conversely, if the population diversity is high, reduce the crossover rate and increase the mutation rate to avoid over-exploration.

[0021] ​​If the current population converges to a local optimal solution, increase the mutation rate to prompt the algorithm to jump out of the local optimum and explore new solution space regions.

[0022] Further, generating multiple typical scenarios includes the following steps:

[0023] Use Monte Carlo simulation to generate multiple scenarios Ω, Ω = {ω1, ω2,..., ω M}, where M is the number of generated scenarios, and each scenario ω m is composed of a set including load, generation, energy storage state, and equipment failure variables, and is represented by the following formula:

[0024]

[0025] where, respectively represent the load, power generation, energy storage system state, and equipment failure condition at time t in scenario ω m ;

[0026] Generate an extreme event scenario set ω extreme , including scenarios of extremely high load and extremely low generation;

[0027] Calculate the critical load guarantee probability for each scenario

[0028]

[0029] where, is an indicator function indicating whether the load is satisfied at time t, is the total load demand at time t in scenario ω, is the total power generation at time t in scenario ω, is the energy storage state of the energy storage system at time t in scenario ω;

[0030] Calculate the cost-benefit evaluation for each scenario: where, are respectively the costs related to generation, energy storage, and faults;

[0031] Retain the typical scenario data where the critical load guarantee probability and cost-benefit evaluation meet the requirements.

[0032] Further, the objective function of the multi-objective optimization model is expressed as: min(αC total +βC CO2 -γP load ), where C total is the economic objective, C CO2 is the environmental objective, and P loadIt is the robustness objective, and α, β, and γ are the weights of the economic objective, environmental objective, and robustness objective respectively;

[0033] The economic objective is expressed as: C total = C generation + C storage + C operation ;

[0034] C generation represents the power generation cost, and its calculation formula is expressed as: where C g is the unit power generation cost of the g-th power generation resource, P g is the power generation of this power generation resource, and C fuel is the fuel cost of traditional power generation resources;

[0035] C storage represents the energy storage cost, and its calculation formula is expressed as: where C install is the initial installation cost of the energy storage system, C operation,t is the energy storage operation and maintenance cost at time t; C operation is the cost related to other operations;

[0036] The environmental objective is modeled as the total carbon emissions, and its calculation formula is expressed as: where γ g is the carbon emission coefficient per unit power generation of the g-th power generation method, and P g is the power generation of this power generation method;

[0037] The robustness objective is modeled as the load guarantee probability. Assuming that in the m-th scenario, the condition for the microgrid to meet the critical load is where is the load demand in scenario ω m , and the calculation formula for the load guarantee probability is: where is the indicator function, indicating whether the load is satisfied at time t, is the load guarantee probability in scenario ω m ;

[0038] According to the priorities of different scenarios and objectives, dynamically adjust the weights α, β, and γ.

[0039] Furthermore, the adaptive multi-objective genetic algorithm is used to solve the multi-objective optimization model, including the following steps:

[0040] Initialize the population as P = {p1, p2,..., p N}, where p iThe design scheme of the microgrid for the i-th individual, N is the population size, and each design scheme p i consists of several genes;

[0041] Establish a multi-objective fitness evaluation:

[0042] F i = αC total + βC CO2 - γP load

[0043] where F i is the fitness of the individual p i ;

[0044] Through the crossover operation, the genes of the parent individuals are combined into new offspring individuals. The crossover formula:

[0045] p ′ = Crossover(p1, p2)

[0046] where p1 and p2 are the parent individuals, and p ′ is the offspring individual generated by crossover;

[0047] Through the mutation operation, the genes of the individual are slightly adjusted to increase the diversity of the population. The mutation formula:

[0048] p ″ = Mutation(p ′ )

[0049] where p ′ is the individual after the crossover operation, and p ″ is the individual after mutation;

[0050] Adaptively adjust the probabilities of crossover and mutation. The adaptive adjustment formula:

[0051] P cross (t) = f1(Diversity(t))

[0052] P mut (t) = f2(Diversity(t))

[0053] where P cross (t) is the crossover probability, P mut (t) is the mutation probability, Diversity(t) represents the diversity of the population in the t-th generation, and f1 and f2 are adjustment functions based on diversity;

[0054] According to the fitness of each individual, select the individuals with higher fitness from the current population as the parent individuals. The selection formula:

[0055]

[0056] Among them, F i is the fitness of individual p i , and is the sum of the fitnesses of all individuals;

[0057] After the algorithm executes several generations of evolution, the final design solution is selected by calculating the optimal solution in the population.

[0058] Specifically, the method for generating the recommended planning solution includes the following steps:

[0059] According to the results of the adaptive multi-objective genetic algorithm, through Pareto front analysis, the balance points of different objectives are obtained, where each balance point represents a solution that balances between different objectives;

[0060] According to the Pareto front results and multi-objective trade-off analysis, a microgrid planning solution is generated. The planning solution includes: the trade-off relationship between objectives, representing the trade-off between economic objectives, environmental objectives, and robustness objectives; candidate design solutions: showing multiple Pareto optimal solutions and their corresponding objective values; according to the set demand scenarios, the optimal microgrid design solution is recommended.

[0061] The beneficial effects of the present invention are as follows: This solution combines robust evolutionary computation with large model technology to solve the complex, multi-objective, and highly uncertain microgrid planning and design problems. Through the construction of a multi-objective model, various factors such as system cost, carbon emissions, and robustness can be comprehensively optimized. Through evolutionary computation to simulate the natural selection and genetic process, the design solution of the microgrid is continuously optimized. By introducing an intelligent large model to guide the optimization of the search space and a real-time feedback mechanism, the efficiency of evolutionary computation and the robustness of the results are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 shows the overall flowchart of the microgrid planning optimization method. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0063] In order to make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0064] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0065] As shown Figure 1 In this embodiment, a microgrid planning optimization method based on robust evolutionary computation is proposed, including the following steps:

[0066] Step 1: Collect real-time and historical data to generate multiple typical scenarios. The data includes load, power generation, energy storage status, and equipment failure data;

[0067] Step 2: Construct a multi-objective optimization model. The objective function of the multi-objective optimization model includes economic objectives, environmental objectives, and robustness objectives;

[0068] Step 3: Use an adaptive multi-objective genetic algorithm to solve the multi-objective optimization model;

[0069] Step 4: Obtain the Pareto optimal solution set and generate a recommended planning scheme;

[0070] During the process of solving the multi-objective optimization model, use a large model to predict the inefficient solution area in the evolutionary path and dynamically adjust the search strategy.

[0071] Specifically, the use of a large model to predict the inefficient solution area in the evolutionary path includes the following steps:

[0072] Input the fitness distribution of the population, the change in the objective function value, and the distribution of the solution space into the large model, Let P be the current population, containing N individuals p1, p2, …, p N , F(P) be the fitness value of each individual in the population, and ΔF(P) be the change in the current objective function value, that is, the difference and fluctuation of the objective function value, be the current solution space;

[0073] By calculating the variance of the individual fitness in the population, judge the convergence degree of the population. If the variance of the fitness is less than the preset threshold, it means that the population tends to converge. The fitness convergence formula is: where F(p i ) is the fitness of individual p i , is the mean of the current population fitness;

[0074] Calculate the change trend of the objective function value between different generations. If the improvement amplitude of the objective function value is less than the preset threshold, it means that the exploration of the current solution space has tended to saturation;

[0075] Analyze the coverage of the current population in the solution space, calculate the solution space coverage, and the solution space coverage formula: where, is the indicator function, which takes the value of 1 when the individual p i falls into different regions of the solution space, and 0 otherwise;

[0076] Based on the analysis results of fitness convergence, the change trend of the objective function value, and the solution space coverage, the large model predicts and identifies inefficient solution regions;

[0077] The described dynamic adjustment search strategy includes the following steps:

[0078] The large model transfers the search direction from the inefficient solution region to the under-explored region to increase the breadth and depth of the search space;

[0079] If the current population diversity is low, increase the crossover rate to promote the exploration of the solution space. Conversely, if the population diversity is high, reduce the crossover rate and increase the mutation rate to avoid over-exploration.

[0080] If the current population converges to a local optimal solution, increase the mutation rate to prompt the algorithm to jump out of the local optimum and explore new solution space regions.

[0081] By analyzing the current evolutionary state and search space in real time, the large model provides the search direction, dynamically adjusts the parameters of the evolutionary algorithm, and generates knowledge prompts related to the problem. In this way, the optimization process can be carried out more efficiently and intelligently, avoiding the local search and inefficient exploration problems existing in traditional evolutionary algorithms and improving the overall optimization performance.

[0082] Furthermore, the generation of multiple typical scenarios includes the following steps:

[0083] Use Monte Carlo simulation to generate multiple scenarios Ω, Ω = {ω1, ω2,..., ω M}, where M is the number of generated scenarios, and each scenario ω m is composed of a set of variables including load, power generation, energy storage state, and equipment failure, and is represented by the following formula:

[0084]

[0085] where, respectively represent the load, power generation, energy storage system state, and equipment failure situation at time t in scenario ω m ;

[0086] Generate an extreme event scenario set ω extreme , including scenarios of extremely high load and extremely low power generation;

[0087] Calculate the critical load guarantee probability for each scenario

[0088]

[0089] where, It is an indicator function that represents whether the load is satisfied at time t. is the total load demand at time t under scenario ω. is the total power generation at time t under scenario ω. is the energy storage state of the energy storage system at time t under scenario ω;

[0090] Calculate the cost - benefit evaluation for each scenario: Among them, are the costs related to generation, energy storage, and faults respectively;

[0091] Retain the data of typical scenarios where the key load guarantee probability and cost - benefit evaluation meet the requirements.

[0092] By extracting features from multiple data sources, scenarios containing various typical and extreme events are generated. These scenarios provide rich inputs for the optimal design of the micro - grid, ensuring that the subsequent optimization algorithm has sufficient robustness when facing different operating environments.

[0093] Furthermore, the objective function of the multi - objective optimization model is expressed as: min(αC total +βC CO2 -γP load ), where C total is the economic objective, C CO2 is the environmental objective, P load is the robustness objective, and α, β, γ are the weights of the economic objective, environmental objective, and robustness objective respectively;

[0094] The economic objective is expressed as: C total =C generation +C storage +C operation ;

[0095] C generation represents the power generation cost, and the calculation formula is expressed as: Among them, C g is the unit power generation cost of the g - th power generation resource, P g is the power generation of this power generation resource, and C fuel is the fuel cost of traditional power generation resources;

[0096] C storage represents the energy storage cost, and the calculation formula is expressed as: Among them, C install is the initial installation cost of the energy storage system, C operation,t is the energy storage operation and maintenance cost at time t; C operation is the cost related to other operations;

[0097] The described environmental objective is modeled as total carbon emissions, and the calculation formula is expressed as: where γ g is the carbon emission coefficient per unit of power generation of the g-th power generation method, and P g is the power generation of this power generation method;

[0098] The described robustness objective is modeled as the load guarantee probability. Assuming that in the m-th scenario, the condition for the microgrid to meet the critical load is where is the load demand under scenario ω m , and the calculation formula for the load guarantee probability is: where is the indicator function, indicating whether the load is satisfied at time t, is the load guarantee probability under scenario ω m ;

[0099] Dynamically adjust the weights α, β, γ according to the priorities of different scenarios and objectives.

[0100] The cost objective mainly focuses on minimizing the cost, the environmental objective focuses on reducing carbon emissions, and the robustness objective ensures that the microgrid can operate stably and guarantee the supply of critical loads in the face of uncertainties. By establishing a multi-objective optimization model, multiple conflicting objectives can be integrated into an optimization problem, and the relationship between different objectives can be balanced by adjusting the weights.

[0101] Furthermore, the described adaptive multi-objective genetic algorithm is used to solve the multi-objective optimization model, including the following steps:

[0102] Initialize the population as P = {p1, p2, …, p N}, where p i is the microgrid design scheme of the i-th individual, N is the population size, and each design scheme p i is composed of several genes;

[0103] Establish a multi-objective fitness evaluation:

[0104] F i = αC total + βC CO2 - γP load

[0105] where F i is the fitness of individual p i ;

[0106] Through the crossover operation Crossover, combine the genes of the parent individuals into new offspring individuals, and the crossover formula:

[0107] p ′ = Crossover(p1, p2)

[0108] where p1 and p2 are parent individuals, and p ′ is the offspring individual generated by crossover;

[0109] The individual genes are slightly adjusted through the mutation operation Mutation to increase the diversity of the population. The mutation formula:

[0110] p ″ = Mutation(p ′ )

[0111] where p ′ is the individual after the crossover operation, and p ″ is the individual after mutation;

[0112] The probabilities of crossover and mutation are adaptively adjusted. The adaptive adjustment formula:

[0113] P cross (t) = f1(Diversity(t))

[0114] P mut (t) = f2(Diversity(t))

[0115] where P cross (t) is the crossover probability, P mut (t) is the mutation probability, Diversity(t) represents the diversity of the population in the t-th generation, and f1 and f2 are adjustment functions based on diversity;

[0116] According to the fitness of each individual, individuals with higher fitness are selected from the current population as parent individuals. The selection formula:

[0117]

[0118] where F i is the fitness of individual p i , is the sum of the fitnesses of all individuals;

[0119] After the algorithm performs several generations of evolution, the final design solution is selected by calculating the optimal solution in the population.

[0120] The evolutionary computation in this step combines uncertainty modeling and robust optimization to ensure that the microgrid can operate stably and meet the requirements of critical loads in the face of uncertainties (such as load fluctuations, uncertainties in renewable energy generation, equipment failures, etc.).

[0121] Specifically, the planning scheme for generating recommendations includes the following steps:

[0122] Based on the results of the adaptive multi-objective genetic algorithm, through Pareto front analysis, the balance points of different objectives are obtained, where each balance point represents a solution that balances between different objectives;

[0123] Based on the Pareto front results and multi-objective trade-off analysis, a microgrid planning scheme is generated. The planning scheme includes: the trade-off relationship between objectives, representing the trade-off between economic objectives, environmental objectives, and robustness objectives; candidate design schemes: showing multiple Pareto optimal solutions and their corresponding objective values; and according to the set demand scenarios, the optimal microgrid design scheme is recommended.

[0124] As used herein, the term "preferred" is intended to be used as an example, illustration, or exemplification. Any aspect or design described herein as "preferred" should not necessarily be construed as more advantageous than other aspects or designs. Instead, the use of the term "preferred" is intended to present concepts in a specific manner. As used in this application, the term "or" is intended to mean an inclusive "or" rather than an exclusive "or". That is, unless otherwise specified or clear from the context, "X uses A or B" is intended to naturally include any one of the permutations. That is, if X uses A; X uses B; or X uses both A and B, then "X uses A or B" is satisfied in any of the foregoing examples.

[0125] Moreover, although the present disclosure has been shown and described with respect to one or more implementations, those skilled in the art will envision equivalent variations and modifications based on reading and understanding this specification and the drawings. The present disclosure includes all such modifications and variations and is limited only by the scope of the appended claims. Specifically with respect to the various functions performed by the above-described components (e.g., elements, etc.), the terms used to describe such components are intended to correspond to any component that performs the specified function of the component (e.g., it is functionally equivalent), unless otherwise indicated, even if structurally different from the disclosed structure that performs the functions in the exemplary implementations of the present disclosure shown herein. Additionally, although a particular feature of the present disclosure has been disclosed with respect to only one of several implementations, such a feature may be combined with one or other features of other implementations as may be desired and advantageous for a given or particular application. Moreover, insofar as the terms "comprises," "has," "contains," or any variation thereof are used in a particular embodiment or claim, such terms are intended to be inclusive in a manner similar to the term "includes."

[0126] Each functional unit in the embodiments of the present invention may be integrated into a processing module, or each unit may exist physically alone, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. When the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium. The above-mentioned storage medium may be a read-only memory, a magnetic disk or an optical disc, etc. Each of the above-mentioned devices or systems may execute the storage method in the corresponding method embodiment.

[0127] In summary, the above embodiments are an implementation manner of the present invention, but the implementation manner of the present invention is not limited by the above embodiments. Any other changes, modifications, substitutions, combinations, and simplifications made without departing from the spirit and principle of the present invention shall be equivalent replacement manners and are all included in the protection scope of the present invention.

Claims

1. A microgrid planning and optimization method based on robust evolutionary computation, characterized in that It includes the following steps: Step 1, collect real-time and historical data to generate multiple typical scenarios, where the data includes load, power generation, energy storage status, and equipment failure data; Step 2, construct a multi-objective optimization model, and the objective function of the multi-objective optimization model includes economic objectives, environmental objectives, and robustness objectives; Step 3, use an adaptive multi-objective genetic algorithm to solve the multi-objective optimization model; Step 4, obtain the Pareto optimal solution set and generate a recommended planning scheme; During the process of solving the multi-objective optimization model, use a large model to predict the inefficient solution area in the evolutionary path and dynamically adjust the search strategy.

2. The microgrid planning optimization method based on robust evolutionary computation according to claim 1, characterized in that The use of the large model to predict the inefficient solution area in the evolutionary path includes the following steps: Input the fitness distribution of the population, the change of the objective function value, and the distribution of the solution space to the large model. Let \(P\) be the current population, which contains \(N\) individuals \(p_1, p_2, \cdots, p_N\). N Let \(F(P)\) be the fitness value of each individual in the population, and \(\Delta F(P)\) be the change of the current objective function value, that is, the difference and fluctuation of the objective function value. Let \(S\) be the current solution space. By calculating the variance of the fitness of individuals in the population, the convergence degree of the population is judged. If the variance of the fitness is less than the preset threshold, it means that the population tends to converge. The fitness convergence formula is as follows: where F(p i ) is the fitness of individual p i , and is the mean value of the fitness of the current population; Calculate the change trend of the objective function value between different generations. If the improvement amplitude of the objective function value is less than the preset threshold, it means that the exploration of the current solution space has tended to saturation; Analyze the coverage of the current population in the solution space, calculate the solution space coverage, and the formula for the solution space coverage is: where is the indicator function, which takes the value of 1 when the individual p i falls into different regions of the solution space, and 0 otherwise; Based on the analysis results of fitness convergence, the change trend of the objective function value, and the solution space coverage, the large model predicts and identifies the inefficient solution area; The dynamic adjustment of the search strategy includes the following steps: The large model transfers the search direction from the inefficient solution area to the area that has not been fully explored to increase the breadth and depth of the search space; If the current population diversity is low, increase the crossover rate to promote the exploration of the solution space. Conversely, if the population diversity is high, reduce the crossover rate and increase the mutation rate to avoid over-exploration. If the current population converges to a certain local optimal solution, increase the mutation rate to prompt the algorithm to jump out of the local optimum and explore new solution space areas.

3. The microgrid planning optimization method based on robust evolutionary computation according to claim 1, characterized in that The generation of multiple typical scenarios includes the following steps: Generate multiple scenarios Ω using Monte Carlo simulation, where Ω = {ω1, ω2,..., ω M}, where M is the number of generated scenarios, and each scenario ω m is composed of a set including variables of load, power generation, energy storage state, and equipment failure, and is represented by the following formula: Among them, and respectively represent the load, power generation, energy storage system status, and equipment failure conditions at time t in scenario ω m ; Generate a set of extreme event scenarios ω extreme , including scenarios of extremely high load and extremely low power generation; Calculate the key load guarantee probability for each scenario wherein, is an indicator function indicating whether the load is satisfied at time t, is the total load demand at time t under scenario ω, is the total power generation at time t under scenario ω, is the energy storage state of the energy storage system at time t under scenario ω; Calculate the cost-benefit assessment for each scenario: Among them, and are the costs related to generation, energy storage, and faults respectively; Retain the typical scenario data where the key load guarantee probability and cost-benefit evaluation meet the preset requirements.

4. A microgrid planning and optimization method based on robust evolutionary computation according to claim 3, characterized in that, The objective function of the multi-objective optimization model is expressed as min(αC total +βC CO2 -γP load ), where C total is the economic objective, C CO2 is the environmental objective, P load is the robustness objective, and α, β, and γ are the weights of the economic objective, environmental objective, and robustness objective, respectively; The economic target described is expressed as: C total = C generation + C storage + C operation ; C generation represents the power generation cost, and the calculation formula is expressed as: C generation = Σ g∈G (C g · P g ) + C fuel , where C g is the unit power generation cost of the g-th power generation resource, G is the set of all power generation resources, P g is the power generation of this power generation resource, and C fuel is the fuel cost of traditional power generation resources; C storage Indicates the energy storage cost, and the calculation formula is expressed as: Among them, C install is the initial installation cost of the energy storage system, and C operation,t is the energy storage operation and maintenance cost at time t; C operation is the cost related to other operations; The described environmental goal is modeled as total carbon emissions, and the calculation formula is expressed as: C CO2 = Σ g∈G γ g ·P g , where γ g is the carbon emission coefficient per unit of power generation of the g-th power generation method, and P g is the power generation of this power generation method; The robustness objective is modeled as the load guarantee probability. Assuming that under the m-th scenario, the condition for the microgrid to meet the critical load is where is the load demand under scenario ω m The calculation formula for the load guarantee probability is: where is the indicator function, indicating whether the load is satisfied at time t, is the load guarantee probability under scenario ω m ; Dynamically adjust the weights α, β, γ according to the priorities of different scenarios and objectives.

5. A microgrid planning optimization method based on robust evolutionary computation according to claim 4, characterized in that, The use of an adaptive multi-objective genetic algorithm to solve the multi-objective optimization model includes the following steps: Initialize the population as P = {p1, p2, …, p N}, where p i is the microgrid design scheme of the i-th individual, N is the population size, and each design scheme p i is composed of several genes; Establish a multi-objective fitness evaluation: F i = αC total + βC CO2 - γP load Among them, F i is the fitness of individual p i ; Through the crossover operation Crossover, combine the genes of the parent individuals into new offspring individuals. The crossover formula: p ′ = Crossover(p1, p2) Among them, p1 and p2 are parental individuals, and p ′ is the offspring individual generated by crossover; Make minor adjustments to the individual genes through the mutation operation Mutation to increase the diversity of the population. The mutation formula: p ″ = Mutation(p ′ ) Among them, p ′ is the individual after crossover operation, and p ″ is the individual after mutation; Adaptive adjustment of the probabilities of crossover and mutation; According to the fitness of each individual, select the individuals with higher fitness from the current population as the parent individuals. The selection formula: Among them, F i is the fitness of individual p i , and is the sum of the fitnesses of all individuals; After the algorithm executes several generations of evolution, select the final design scheme by calculating the optimal solution in the population.

6. The microgrid planning optimization method based on robust evolutionary computation according to claim 5, characterized in that The generation of the recommended planning scheme includes the following steps: According to the results of the adaptive multi-objective genetic algorithm, through Pareto front analysis, obtain the balance points of different objectives, where each balance point represents a solution that balances different objectives; According to the Pareto front results and multi-objective trade-off analysis, generate a microgrid planning scheme. The planning scheme includes: the trade-off relationship between various objectives, representing the trade-off between economic objectives, environmental objectives, and robustness objectives; candidate design schemes: display multiple Pareto optimal solutions and their corresponding objective values; according to the set demand scenarios, recommend the optimal microgrid design scheme.

Images