Multi-objective optimization configuration method for energy storage system in distribution network

Abstract

The present invention discloses a multi-objective optimization configuration method for an energy storage system of a distribution network, and relates to the technical field of multi-objective optimization configuration. The method comprises the following steps: randomly generating a group of initial energy storage configuration schemes based on historical data; obtaining convergence rate data and optimization target satisfaction data in the analysis process of the energy storage configuration schemes; constructing an optimization configuration evaluation model based on a long-term and short-term neural network based on the convergence rate data and the optimization target satisfaction data; obtaining a number of different crossover rates and mutation rates, combining them two by two into a number of different optimization groups, and constructing a three-dimensional display model for the output of the optimization configuration evaluation model corresponding to the number of different optimization combinations, and performing curve analysis to obtain the analyzed crossover rate and mutation rate. The present application aims to solve the problem of imbalance between exploring new solutions and retaining excellent solutions for the multi-objective optimization configuration of an energy storage system of a distribution network.

Description

Technical Field

[0001] The present invention relates to the technical field of multi-objective optimization configuration, and more specifically, to a multi-objective optimization configuration method for an energy storage system of a distribution network. Background Art

[0002] The distribution network is an important part of the power system, responsible for transmitting electrical energy from the high-voltage transmission system to the end user. Its stability and reliability are crucial to power supply. In modern distribution networks, renewable energy and distributed generation facilities are increasingly being introduced, which brings new challenges and opportunities for the optimal configuration of the system; energy storage systems can store electrical energy when electricity demand is low and release it when demand is high, balancing the load and improving the flexibility and reliability of the grid. In the distribution network, energy storage systems can effectively alleviate load fluctuations, optimize energy use, and promote the integration of renewable energy.

[0003] Multi-objective optimization is a method for solving problems involving multiple conflicting objectives. In the configuration of energy storage systems in distribution networks, multiple factors such as cost, efficiency, environmental impact, and system reliability may need to be considered simultaneously. Through multi-objective optimization, the best balance can be sought between different objectives to achieve overall performance improvement; configuration methods refer to specific plans and strategies formulated to achieve optimal system operation. In the configuration of energy storage systems, this may involve selecting appropriate energy storage technologies, determining the optimal energy storage capacity, optimizing scheduling strategies, etc.

[0004] Genetic algorithms are widely used in the multi-objective optimization configuration of distribution network energy storage systems to solve the balance problem between multiple objectives such as cost, efficiency, environmental impact and system reliability. By simulating the process of natural selection, genetic algorithms can quickly search for configuration solutions close to the Pareto optimal solution in a large-scale solution space. Specifically, the algorithm generates new solutions through operations such as selection, crossover and mutation, and evaluates their advantages and disadvantages through fitness functions, and then continuously iterates and optimizes. In distribution network energy storage systems, genetic algorithms not only improve the flexibility of energy storage equipment selection and capacity configuration, but also effectively cope with the volatility and uncertainty of renewable energy, and promote the economy and reliability of the overall power system.

[0005] For example, the invention patent with announcement number: CN111435788A discloses a method and device for improving the ability of a distribution network to accept distributed power sources, including determining the optimization result of the distribution network based on a first characterization parameter of the ability of the distribution network to accept distributed power sources using a multi-objective optimization algorithm; judging whether the ability of the distribution network to accept distributed power sources meets the improvement requirements based on the optimization results, and if the improvement requirements are not met, repeating the above steps until the ability of the distribution network to accept distributed power sources meets the improvement requirements. The present invention avoids the limitation of maximizing the access capacity of distributed power sources from only the single perspective of optimization, fundamentally solves the problem of improving the ability of the distribution network to accept distributed power sources, and has a high degree of improvement in the access capacity or allowed penetration rate of distributed power sources, greatly improving the maximum access capacity of distributed power sources, and ensuring that large-scale distributed power sources can be safely and reliably connected to the distribution network.

[0006] For example, the invention patent with announcement number: CN117578436A announces a distribution network energy storage system planning method based on a decomposition multi-objective optimization algorithm, including obtaining real-time basic data of the distribution network; determining the optimization target of the distribution network energy storage system planning according to the real-time basic data of the distribution network; decomposing the optimization target of the distribution network energy storage system planning according to the decomposition multi-objective optimization algorithm to obtain multiple sub-optimization targets; setting a corresponding weight vector for each sub-optimization target, and aggregating the sub-optimization targets according to the weight vector to obtain an aggregated objective function; solving the aggregated objective function, and using the obtained optimal solution as the distribution network optimal energy storage system planning method. In this way, the distribution network energy storage system can be effectively planned while satisfying various constraints and achieving a balance between multiple targets such as optimizing investment costs, maximizing reliability, and minimizing energy loss.

[0007] The above disclosed technical solution has at least the following technical problems: in the application of genetic algorithm in the multi-objective optimization configuration of distribution network energy storage system, the setting of crossover rate and mutation rate has a huge impact on the configuration effect.

[0008] When the crossover rate is too high, the solution with better configuration will disappear in the next generation, affecting the overall optimization quality. And because the information of good solutions is frequently disturbed, the number of generations required for convergence increases, reducing the efficiency of the algorithm.

[0009] When the mutation rate is too high, the features of an existing excellent solution may be randomly changed. Such frequent mutations may cause good configurations to be covered by random noise, affecting the stability of the optimization and the reliability of the final result.

[0010] However, there is a balance between the crossover rate and the mutation rate, which respectively affect the exploration of new solutions and the retention of excellent solutions in the multi-objective optimization configuration of the energy storage system of the power supply and distribution network. The existing technology does not have a sufficient balance between the crossover rate and the mutation rate in energy scheduling, resulting in the loss of excellent solution information and slow convergence when the crossover rate is too high, while the excessively high mutation rate leads to the destruction of the solution and the reduction of the search efficiency, thus affecting the overall optimization effect. In view of the above problems, the present invention proposes a solution. Summary of the invention

[0011] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a multi-objective optimization configuration method for an energy storage system of a distribution network, which solves the problem of imbalance between exploring new solutions and retaining excellent solutions in the multi-objective optimization configuration of an energy storage system of a distribution network by analyzing the settings of the crossover rate and the mutation rate.

[0012] To achieve the above object, the present invention provides the following technical solutions:

[0013] The multi-objective optimization configuration method of the energy storage system of the distribution network includes the following steps: randomly generating a group of initial energy storage configuration schemes based on historical data; performing crossover operations and mutation operations on the initial energy storage configuration schemes based on genetic algorithms according to preset crossover rates and mutation rates to obtain a final energy storage configuration scheme; obtaining convergence rate data and optimization target satisfaction data in the analysis process of the energy storage configuration schemes; constructing an optimization configuration evaluation model based on a long-term and short-term neural network based on the convergence rate data and the optimization target satisfaction data; obtaining a number of different crossover rates and mutation rates, combining them into a number of different optimization groups in pairs, and constructing a three-dimensional display model with the output of the optimization configuration evaluation model corresponding to the different optimization combinations, and performing curve analysis to obtain the analyzed crossover rate and mutation rate.

[0014] In a preferred embodiment, the method of randomly generating a set of initial energy storage configuration schemes based on historical data is specifically as follows: according to historical data analysis, the capacity range, quantity limit and other necessary parameters of various types of energy storage equipment are set to obtain the parameter range; a random number generator is used to generate the number and capacity of energy storage equipment within the set parameter range to obtain several energy storage configuration schemes; and the energy storage configuration schemes are screened based on physical and economic feasibility to obtain a set of initial energy storage configuration schemes.

[0015] In a preferred embodiment, the initial energy storage configuration scheme is subjected to crossover and mutation operations based on a genetic algorithm according to a preset crossover rate and mutation rate to obtain a final energy storage configuration scheme, specifically: according to a preset crossover rate, a part of excellent individuals are selected from the initial energy storage configuration scheme as parents; through a crossover operation, the characteristics of the parent individuals are combined to generate a new offspring configuration scheme; according to a preset mutation rate, the newly generated offspring configuration scheme is mutated; the mutation operation includes randomly adjusting the capacity and quantity of any energy storage device and changing the working mode of the device; the new configuration scheme after the crossover and mutation operations is merged with the original scheme to form an optimized energy storage configuration scheme.

[0016] In a preferred embodiment, the convergence rate data includes a convergence rate coefficient; the specific method for obtaining the convergence rate coefficient is as follows: obtaining the convergence rate at the same time interval during the analysis of the energy storage configuration scheme; the convergence standard deviation and the convergence average value of the convergence rate at the same time interval during the analysis of the energy storage configuration scheme; calculating the convergence rate variation coefficient based on the convergence standard deviation and the convergence average value; calculating the convergence rate coefficient based on the convergence rate variation coefficient based on a preset convergence rate coefficient calculation formula.

[0017] In a preferred embodiment, the optimization target satisfaction data includes a model fitness value matching coefficient and an optimization target satisfaction coefficient; the specific method for obtaining the model fitness value matching coefficient is as follows: obtaining the fitness value corresponding to the iterative output data of several model iterations; calculating the cosine similarity between each model iteration and the fitness value corresponding to other model iterations; obtaining the average cosine similarity, maximum cosine similarity and minimum cosine similarity of the fitness value corresponding to each model iteration; calculating the model fitness value matching coefficient based on the average cosine similarity, maximum cosine similarity and minimum cosine similarity based on a preset model fitness value matching coefficient calculation formula.

[0018] In a preferred embodiment, a specific method for obtaining the satisfaction degree coefficient of the optimization target is as follows: based on the demand types of historical optimization targets, a demand set of the optimization target is obtained, wherein the demand set of the optimization target includes the demand quantities and target proportions of various targets to be optimized and their optimization targets; a preset demand score is assigned to the demand types of various optimization targets within the demand set of the optimization target; a regression model is constructed based on the demand set of the optimization target and the corresponding demand scores to generate the satisfaction degree coefficient of the optimization target.

[0019] In a preferred embodiment, the method comprises obtaining several different crossover rates and mutation rates, combining them two by two into several different optimization groups, and constructing a three-dimensional display model with the outputs of the optimization configuration evaluation models corresponding to the several different optimization combinations, specifically: obtaining several different crossover rates and mutation rates, combining them two by two into several different optimization groups, and constructing a three-dimensional display model with the outputs of the optimization configuration evaluation models corresponding to the several different optimization combinations, specifically: constructing vector groups with several different crossover rates and mutation rates, and constructing a plane rectangular coordinate system with the crossover rate and mutation rate as the vertical axis and the horizontal axis respectively; using the output of the optimization configuration evaluation model corresponding to each combination of crossover rate and mutation rate as the third axis of the plane rectangular coordinate system to construct a three-dimensional display model.

[0020] In a preferred embodiment, the curve analysis is performed to obtain the crossover rate and mutation rate after analysis, specifically: the maximum values ​​of several three-dimensional display models are obtained, and the derivative analysis of the slope of the change curve of each maximum value is performed to obtain the change rate, and the crossover rate and mutation rate corresponding to the maximum value of the three-dimensional display model with the smallest change rate are used as the crossover rate and mutation rate after analysis, and are applied to the multi-objective optimization configuration process of the energy storage system of the distribution network.

[0021] The technical effects and advantages of the multi-objective optimization configuration method of the energy storage system of the distribution network of the present invention are as follows:

[0022] 1. The present invention randomly generates an initial energy storage configuration scheme through historical data analysis. This step ensures that the type, capacity and scheduling strategy of the energy storage equipment are reasonable and feasible. The generated scheme is screened for physical and economic feasibility so that the subsequent optimization process is based on an effective foundation. Secondly, the genetic algorithm operation is performed on the initial scheme using the set crossover rate and mutation rate to generate a new configuration scheme. This process not only improves the diversity of solutions, but also effectively prevents the algorithm from falling into local optimality, ensuring that the final scheme is more advantageous. Finally, the convergence rate data and the optimization target satisfaction data are analyzed to quantify the optimization effect and improve the stability of the algorithm. The convergence speed coefficient and the model fitness value matching coefficient provide dynamic feedback on the algorithm performance, helping researchers and decision makers to make scientific adjustments.

[0023] 2. The present invention generates an optimization configuration evaluation coefficient by integrating convergence rate data, model fitness value matching coefficient and optimization target satisfaction coefficient to quantify the performance of various configuration schemes. By performing three-dimensional display and curve analysis on different crossover rate and mutation rate combinations, the optimal parameter combination can be intuitively identified, thereby optimizing the energy storage configuration scheme. The advantages of this technology include: first, the powerful processing ability of the LSTM model for time series data makes the optimization evaluation more accurate and dynamic; second, through the construction of a three-dimensional display model, the impact of different parameter combinations can be clearly visualized, which is convenient for decision makers to quickly find the best configuration; finally, the systematic evaluation and feedback mechanism can effectively balance the relationship between exploring new solutions and retaining excellent solutions, and improve the stability and efficiency of the optimization process. This method optimizes the configuration of the energy storage system while ensuring multiple goals such as economy, efficiency and environmental impact, and promotes the development of smart grids. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 It is a structural schematic diagram of the multi-objective optimization configuration method of the energy storage system of the distribution network of the present invention. DETAILED DESCRIPTION

[0025] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0026] Embodiment 1, Figure 1 A multi-objective optimization configuration method for an energy storage system of a distribution network of the present invention is provided, comprising the following steps:

[0027] S1, randomly generates a set of initial energy storage configuration schemes based on historical data.

[0028] In this embodiment, the energy storage configuration scheme includes energy storage device type, capacity and scheduling strategy;

[0029] The method of randomly generating a set of initial energy storage configuration schemes based on historical data is specifically as follows:

[0030] According to historical data analysis, the capacity range, quantity limit and other necessary parameters of various energy storage devices are set to obtain the parameter range;

[0031] Use a random number generator to generate the number and capacity of energy storage devices within a set parameter range to obtain several energy storage configuration schemes;

[0032] The energy storage configuration schemes are screened based on physical and economic feasibility to obtain a set of initial energy storage configuration schemes.

[0033] It should be noted that through in-depth analysis of historical data, the capacity range and quantity limit of various energy storage devices can be determined. For example, based on previous load demand and equipment performance, the capacity range of lithium batteries is set to 50-200kWh, and the quantity is limited to 2-10. At the same time, other necessary parameters such as charging and discharging efficiency, equipment life and investment cost are considered to form a reasonable parameter range, which provides a basis for the subsequent generation of energy storage configuration plans.

[0034] It should be noted that the random number generator is used to randomly generate multiple sets of energy storage configuration schemes within the preset energy storage device capacity range and quantity limit. For example, by randomly selecting the number of devices with a capacity of 50 to 200 kWh, multiple different combinations of energy storage configurations are generated to ensure that each set of schemes has diversity and feasibility within the parameter range, thereby obtaining several effective energy storage configuration schemes for subsequent optimization.

[0035] It should be noted that the generated energy storage configuration schemes are screened based on physical and economic feasibility to ensure that each set of schemes is feasible in practical applications. For example, by checking whether the capacity of each scheme meets the load demand of the distribution network and evaluating the balance between its investment cost and expected benefits, qualified configuration schemes are screened out, thus ultimately obtaining a set of effective initial energy storage configuration schemes, which provide a basis for subsequent optimization.

[0036] S2, performing crossover operation and mutation operation on the initial energy storage configuration scheme based on the genetic algorithm according to the preset crossover rate and mutation rate, to obtain the final energy storage configuration scheme.

[0037] The crossover operation and mutation operation are performed on the initial energy storage configuration scheme based on the genetic algorithm according to the preset crossover rate and mutation rate to obtain the final energy storage configuration scheme, which is specifically:

[0038] According to the preset crossover rate, a portion of excellent individuals are selected from the initial energy storage configuration scheme as parents;

[0039] Through the crossover operation, the characteristics of the parent individuals are combined to generate a new offspring configuration scheme;

[0040] According to the preset mutation rate, the newly generated offspring configuration scheme is mutated;

[0041] The mutation operation includes randomly adjusting the capacity and quantity of any energy storage device and changing the working mode of the device;

[0042] The new configuration scheme after crossover and mutation operations is merged with the original scheme to form an optimized energy storage configuration scheme.

[0043] It should be noted that the crossover operation is, for example, to randomly select two parent individuals, exchange the capacity or quantity of some of their energy storage devices, create a new solution, and thus explore a better configuration; the mutation operation process aims to introduce new features and increase the diversity of the population to avoid the algorithm falling into a local optimum.

[0044] S3, obtaining the convergence rate data and optimization target satisfaction data during the analysis of the energy storage configuration scheme.

[0045] The convergence rate data includes a convergence speed coefficient, and the optimization target satisfaction data includes a model fitness value matching coefficient and an optimization target satisfaction degree coefficient;

[0046] The convergence speed coefficient is an indicator used to evaluate the number of iterations required by the genetic algorithm when finding the optimal configuration solution for the distribution network energy storage system. Its analysis method usually involves recording the changes in the fitness value in each iteration. By calculating the rate of change of the fitness value, the speed at which the algorithm reaches the optimal solution can be determined. Specifically, the convergence speed coefficient can be used to fit the curve of the change of the fitness value with the number of iterations to calculate the number of iterations required to reach the set fitness threshold, thereby reflecting the rationality of the parameter setting. Under different combinations of mutation rate and crossover rate, this coefficient helps to evaluate the effect of the multi-objective optimization configuration of the distribution network energy storage system, thereby balancing the relationship between exploring new solutions and retaining excellent solutions.

[0047] Analyzing the convergence speed coefficient is useful for analyzing the multi-objective optimization configuration effect of the energy storage system of the distribution network under different combinations of mutation rate and crossover rate, and solving the problem of imbalance between exploring new solutions and retaining excellent solutions for the multi-objective optimization configuration of the energy storage system of the distribution network. It has the following advantages:

[0048] Effectively evaluate optimization efficiency: The convergence speed coefficient can quantitatively evaluate the speed at which the algorithm reaches the optimal solution, providing a basis for direct comparison of different parameter combinations, thereby identifying the most effective mutation rate and crossover rate.

[0049] Balance exploration and development: By monitoring the convergence speed, it is possible to effectively determine whether the current parameter settings cause the algorithm to fall into a local optimum, thereby adjusting the mutation rate and crossover rate to enhance the diversity of solutions and avoid the loss of excellent solutions.

[0050] Guiding parameter adjustment: Changes in convergence speed can provide feedback for the optimization process, helping researchers dynamically adjust the mutation rate and crossover rate to achieve a better balance between exploring new solutions and retaining excellent solutions.

[0051] Optimized decision support: The analysis of the convergence speed coefficient provides a scientific basis for decision makers, helping to formulate more reasonable energy storage configuration plans and improve the overall operating efficiency and economy of the system.

[0052] Improve algorithm stability: By analyzing the convergence speed under different parameter combinations, it is possible to identify parameter settings that perform well under various conditions, thereby improving the stability and reliability of the optimization process.

[0053] In summary, the convergence speed coefficient plays a key role in analyzing and optimizing the energy storage system in the distribution network, and helps to achieve efficient and balanced multi-objective optimization configuration.

[0054] The specific method for obtaining the convergence speed coefficient is as follows:

[0055] Obtain the convergence rate at the same time interval during the analysis of the energy storage configuration scheme;

[0056] The convergence standard deviation and convergence mean of the convergence rate at the same time interval during the analysis of the energy storage configuration scheme;

[0057] Calculate the coefficient of variation of convergence speed based on the convergence standard deviation and convergence mean;

[0058] The convergence speed coefficient is calculated based on the convergence speed coefficient variation coefficient according to a preset convergence speed coefficient calculation formula.

[0059] The specific calculation formula of the convergence speed variation coefficient is as follows:

[0060]

[0061] The specific calculation formula of the convergence speed coefficient is as follows:

[0062]

[0063] In the formula, is the convergence speed variation coefficient, is the convergence rate at the jth moment, is the total number of moments collected during the analysis of the energy storage configuration scheme, and j is the moment number; is the convergence speed coefficient.

[0064] The model fitness value matching coefficient is an indicator used to measure the degree of match between the fitness value of each generation of energy storage configuration schemes and the preset target. The analysis method includes recording the fitness value of each configuration scheme in each generation, calculating the average and standard deviation of the fitness value, and observing the changes in the quality of the solution under different crossover rates and mutation rates. Specifically, the fitness value matching coefficient can be calculated by comparing the fitness value of the current generation with the historical optimal fitness value to evaluate the convergence and optimization ability of the algorithm under the current parameter combination. Through this coefficient, the impact of different parameter combinations on the multi-objective optimization configuration effect of the distribution network energy storage system can be systematically analyzed, helping to solve the imbalance between exploring new solutions and retaining excellent solutions, thereby improving optimization efficiency.

[0065] The matching coefficient of the fitness value of the analysis model is useful for analyzing the multi-objective optimization configuration effect of the energy storage system of the distribution network under different combinations of mutation rate and crossover rate, and is used to solve the problem of imbalance between exploring new solutions and retaining excellent solutions for the multi-objective optimization configuration of the energy storage system of the distribution network. It has the following advantages:

[0066] Quantitative optimization effect: Through the fitness value matching coefficient, the performance of each configuration scheme can be quantitatively evaluated, providing an objective comparison basis for different parameter combinations and helping to identify the optimal combination.

[0067] Dynamic feedback mechanism: Real-time recording and analysis of fitness values ​​can provide instant feedback to the algorithm, and timely adjust the mutation rate and crossover rate to better adapt to the needs of the current optimization stage and balance the relationship between exploring new solutions and retaining excellent solutions.

[0068] Improve convergence: By evaluating the degree of match between the fitness value and the preset target, it is possible to effectively determine whether the current algorithm has reached convergence and adjust the strategy in time, thereby improving the efficiency of the overall optimization.

[0069] Optimization decision support: The analysis of fitness value matching coefficients provides a scientific basis for decision makers, which can help formulate more targeted energy storage configuration plans and improve the economy and operation efficiency of the system.

[0070] Enhance the diversity of solutions: When analyzing fitness values, it is possible to identify under-explored solution spaces, prompting the algorithm to introduce more diversity during mutation and crossover, and prevent the loss of excellent solutions.

[0071] Systematic evaluation and adjustment: By analyzing the fitness value matching coefficient, the advantages and disadvantages of different parameter settings can be systematically evaluated, so as to make targeted adjustments to ensure the balance and effectiveness of the optimization process.

[0072] In summary, the model fitness value matching coefficient plays an important role in the multi-objective optimization configuration of distribution network energy storage systems, and helps to improve the optimization ability and stability of the algorithm.

[0073] The specific method for obtaining the model fitness value matching coefficient is as follows:

[0074] Obtain the fitness values ​​corresponding to the iterative output data of several model iterations;

[0075] Calculate the cosine similarity between each model iteration and the fitness values ​​corresponding to other model iterations;

[0076] Get the average cosine similarity, maximum cosine similarity, and minimum cosine similarity of the fitness values ​​corresponding to each model iteration;

[0077] The model fitness value matching coefficient is calculated based on the average cosine similarity, the maximum cosine similarity and the minimum cosine similarity based on a preset model fitness value matching coefficient calculation formula.

[0078] The specific calculation formula of the cosine similarity between the fitness values ​​is as follows:

[0079]

[0080] The specific calculation formula of the average cosine similarity is:

[0081]

[0082] The specific calculation formula of the model fitness value matching coefficient is:

[0083]

[0084] In the formula, is the cosine similarity between the fitness values ​​of the ith model iteration and the jth model iteration, is the kth vector value of the fitness value corresponding to the i-th model iteration, is the kth vector value of the fitness value corresponding to the jth model iteration, m is the dimension of the fitness value, is the average cosine similarity of the i-th model iteration, N is the total number of model iterations, is the fitness value of the i-th model iteration, is the fitness value of the jth model iteration, is the model fitness value matching coefficient, is the maximum cosine similarity of the i-th model iteration, is the minimum cosine similarity of the i-th model iteration.

[0085] The degree of satisfaction coefficient of the optimization goal is used to evaluate the degree of realization of multiple optimization goals (such as cost, efficiency, environmental impact, etc.) to reflect the balance of different energy storage configuration schemes in meeting these goals. The analysis method includes comparing the actual value of each configuration scheme on each goal with the preset target value, and quantifying the degree of achievement of each goal by calculating the proportion or difference value of the degree of satisfaction. Under different combinations of mutation rate and crossover rate, this coefficient can help observe the balance relationship between the various goals, identify the trade-offs that may exist in the optimization process, and effectively solve the imbalance between exploring new solutions and retaining excellent solutions, thereby improving the multi-objective optimization configuration effect of the distribution network energy storage system.

[0086] Analyzing the satisfaction coefficient of the optimization target is useful for analyzing the multi-objective optimization configuration effect of the energy storage system of the distribution network under different combinations of mutation rate and crossover rate, and solving the problem of imbalance between exploring new solutions and retaining excellent solutions for the multi-objective optimization configuration of the energy storage system of the distribution network. It has the following advantages:

[0087] Comprehensively evaluate the achievement of multiple goals: By quantifying the degree of achievement of each goal, the satisfaction coefficient can provide a clear effect evaluation for different parameter combinations and help identify the best configuration solution under multiple optimization goals.

[0088] Promote the balance between goals: The satisfaction coefficient can reveal the mutual influence between different goals, ensuring that while pursuing the optimization of a certain goal, other goals can also reach a reasonable level of satisfaction, thereby effectively avoiding excessive bias towards one aspect.

[0089] Dynamically adjust the optimization strategy: By analyzing the satisfaction coefficient in real time, the mutation rate and crossover rate can be dynamically adjusted according to the performance of the current parameter settings to optimize the search process and balance the relationship between exploring new solutions and retaining excellent solutions.

[0090] Support decision-making: Based on the analysis of the satisfaction coefficient, decision makers can formulate energy storage configuration plans more scientifically to ensure that the selected plan meets the expected goals in terms of cost, efficiency, environmental impact and other aspects.

[0091] Enhance the adaptability of the algorithm: The analysis of this coefficient can help identify parameter combinations that perform well under specific conditions, enhance the algorithm's adaptability to different environments and requirements, and improve overall optimization efficiency.

[0092] Reducing imbalance risk: By evaluating the degree of satisfaction of the optimization objectives, the imbalance risk that may occur between exploring new solutions and retaining excellent solutions can be effectively identified and reduced, thereby improving the stability and reliability of the optimization process.

[0093] In summary, the satisfaction degree coefficient of the optimization objective plays an important role in the multi-objective optimization configuration of the distribution network energy storage system, which helps to improve the optimization effect and decision-making quality.

[0094] The specific method for obtaining the satisfaction degree coefficient of the optimization target is as follows:

[0095] Based on the demand types of historical optimization targets, a demand set of optimization targets is obtained, wherein the demand set of optimization targets includes demand quantities and target proportions of various targets to be optimized and their optimization targets;

[0096] Assigning preset demand scores to demand types of various optimization targets within the demand set of optimization targets;

[0097] A regression model is constructed based on the demand set of the optimization target and the corresponding demand score to generate the satisfaction degree coefficient of the optimization target.

[0098] The specific calculation formula of the satisfaction degree coefficient of the optimization target is as follows:

[0099]

[0100] In the formula, To optimize the satisfaction coefficient of the target, … The preset demand score corresponding to the demand of the optimization target, … is the proportion of the demand for the optimization target in the demand for all optimization targets, is the number of types of requirements for the optimization target, n is the number of requirements for the optimization target, and e is a natural number.

[0101] This embodiment randomly generates an initial energy storage configuration scheme through historical data analysis. This step ensures that the type, capacity and scheduling strategy of the energy storage equipment are reasonable and feasible. The generated scheme is screened for physical and economic feasibility so that the subsequent optimization process is based on an effective foundation. Secondly, the initial scheme is operated by genetic algorithm using the set crossover rate and mutation rate to generate a new configuration scheme. This process not only improves the diversity of solutions, but also effectively prevents the algorithm from falling into local optimality, ensuring that the final scheme is more advantageous. Finally, the convergence rate data and the optimization target satisfaction data are analyzed to quantify the optimization effect and improve the stability of the algorithm. The convergence speed coefficient and the model fitness value matching coefficient provide dynamic feedback on the algorithm performance, helping researchers and decision makers to make scientific adjustments.

[0102] Embodiment 2, S4, constructs an optimization configuration evaluation model based on a long-term and short-term neural network based on convergence rate data and optimization target satisfaction data.

[0103] The optimization configuration evaluation model is constructed based on the long-term and short-term neural network based on the convergence rate data and the optimization target satisfaction data, specifically:

[0104] The obtained convergence speed coefficient, model fitness value matching coefficient and optimization target satisfaction coefficient are used to establish a traffic shaping hidden danger assessment model to generate an optimization configuration assessment coefficient;

[0105] The specific calculation formula of the optimization configuration evaluation coefficient is as follows:

[0106]

[0107] In the formula, To optimize the configuration evaluation coefficient, is the preset proportional coefficient of the convergence speed coefficient, is the preset proportional coefficient of the model fitness value matching coefficient, The preset proportional coefficient for the degree of satisfaction of the optimization target, is the convergence speed coefficient, is the preset proportional coefficient of the model fitness value matching coefficient, The preset proportional coefficient for the degree of satisfaction of the optimization target.

[0108] S5, obtain several different crossover rates and mutation rates, combine them into several different optimization groups in pairs, and construct a three-dimensional display model for the output of the optimization configuration evaluation model corresponding to the several different optimization combinations, and perform curve analysis to obtain the analyzed crossover rate and mutation rate.

[0109] Obtain several different crossover rates and mutation rates, combine them into several different optimization groups, and construct a three-dimensional display model with the output of the optimization configuration evaluation model corresponding to the different optimization combinations, specifically:

[0110] Construct vector groups with different crossover rates and mutation rates respectively, and construct a rectangular coordinate system with the crossover rate and mutation rate as the vertical axis and horizontal axis respectively;

[0111] The output of the optimization configuration evaluation model corresponding to each combination of crossover rate and mutation rate is used as the third axis of the plane rectangular coordinate system to construct a three-dimensional display model.

[0112] The curve analysis is performed to obtain the crossover rate and mutation rate after analysis, which are specifically:

[0113] The maximum values ​​of several three-dimensional display models are obtained, and the derivative analysis of the slope of the change curve of each maximum value is performed to obtain the change rate. The crossover rate and mutation rate corresponding to the maximum value of the three-dimensional display model with the smallest change rate are used as the crossover rate and mutation rate after analysis, and are applied to the multi-objective optimization configuration process of the energy storage system of the distribution network.

[0114] This embodiment generates an optimization configuration evaluation coefficient by integrating convergence rate data, model fitness value matching coefficient and optimization target satisfaction coefficient to quantify the performance of various configuration schemes. By performing three-dimensional display and curve analysis on different crossover rate and mutation rate combinations, the optimal parameter combination can be intuitively identified, thereby optimizing the energy storage configuration scheme. The advantages of this technology include: first, the powerful processing ability of the LSTM model for time series data makes the optimization evaluation more accurate and dynamic; second, through the construction of a three-dimensional display model, the impact of different parameter combinations can be clearly visualized, which facilitates decision makers to quickly find the best configuration; finally, the systematic evaluation and feedback mechanism can effectively balance the relationship between exploring new solutions and retaining excellent solutions, and improve the stability and efficiency of the optimization process. This method optimizes the configuration of the energy storage system while ensuring multiple goals such as economy, efficiency and environmental impact, and promotes the development of smart grids.

[0115] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0116] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product.

[0117] Those of ordinary skill in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0118] In addition, each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0119] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0120] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A multi-objective optimization configuration method for energy storage systems in a distribution network, characterized in that: The steps include: Randomly generate a set of initial energy storage configuration solutions based on historical data; According to the preset crossover rate and mutation rate, the initial energy storage configuration scheme is subjected to crossover operation and mutation operation based on the genetic algorithm to obtain the final energy storage configuration scheme; Obtaining convergence rate data and optimization target satisfaction data during the analysis of energy storage configuration solutions; The convergence rate data includes a convergence rate coefficient; the specific method for obtaining the convergence rate coefficient is as follows: Obtain the convergence rate at the same time interval during the analysis of the energy storage configuration scheme; The convergence standard deviation and convergence mean of the convergence rate at the same time interval during the analysis of the energy storage configuration scheme; Calculate the coefficient of variation of convergence speed based on the convergence standard deviation and convergence mean; Calculate the convergence speed coefficient based on the convergence speed coefficient calculation formula using the convergence speed variation coefficient; The optimization target satisfaction data includes a model fitness value matching coefficient and an optimization target satisfaction coefficient; the specific method for obtaining the model fitness value matching coefficient is as follows: Obtain the fitness values ​​corresponding to the iterative output data of several model iterations; Calculate the cosine similarity between each model iteration and the fitness values ​​corresponding to other model iterations; Get the average cosine similarity, maximum cosine similarity, and minimum cosine similarity of the fitness values ​​corresponding to each model iteration; Calculate the model fitness value matching coefficient based on the average cosine similarity, the maximum cosine similarity and the minimum cosine similarity based on a preset model fitness value matching coefficient calculation formula; The specific method for obtaining the satisfaction degree coefficient of the optimization target is as follows: Based on the demand types of historical optimization targets, a demand set of optimization targets is obtained, wherein the demand set of optimization targets includes demand quantities and target proportions of various targets to be optimized and their optimization targets; Assigning preset demand scores to demand types of various optimization targets within the demand set of optimization targets; A regression model is constructed based on the demand set of the optimization target and the corresponding demand score to generate the satisfaction degree coefficient of the optimization target; Based on the convergence rate data and optimization target satisfaction data, an optimization configuration evaluation model is constructed based on the long-term and short-term neural network; Several different crossover rates and mutation rates are obtained, and they are combined into several different optimization groups in pairs. The output of the optimization configuration evaluation model corresponding to the several different optimization combinations is used to construct a three-dimensional display model, and curve analysis is performed to obtain the analyzed crossover rate and mutation rate.

2. The multi-objective optimization configuration method for the energy storage system of the distribution network according to claim 1 is characterized in that: The method of randomly generating a set of initial energy storage configuration schemes based on historical data is specifically as follows: According to historical data analysis, the capacity range and quantity limit of various energy storage devices are set to obtain the parameter range; Use a random number generator to generate the number and capacity of energy storage devices within a set parameter range to obtain several energy storage configuration schemes; The energy storage configuration schemes are screened based on physical and economic feasibility to obtain a set of initial energy storage configuration schemes.

3. The multi-objective optimization configuration method for the energy storage system of the distribution network according to claim 2 is characterized in that: The crossover operation and mutation operation are performed on the initial energy storage configuration scheme based on the genetic algorithm according to the preset crossover rate and mutation rate to obtain the final energy storage configuration scheme, which is specifically: According to the preset crossover rate, a portion of excellent individuals are selected from the initial energy storage configuration scheme as parents; Through the crossover operation, the characteristics of the parent individuals are combined to generate a new offspring configuration scheme; According to the preset mutation rate, the newly generated offspring configuration scheme is mutated; The mutation operation includes randomly adjusting the capacity and quantity of any energy storage device and changing the working mode of the device; The new configuration scheme after crossover and mutation operations is merged with the original scheme to form an optimized energy storage configuration scheme.

4. The multi-objective optimization configuration method for the energy storage system of the distribution network according to claim 3 is characterized in that: The method of obtaining a plurality of different crossover rates and mutation rates, combining them into a plurality of different optimization groups in pairs, and constructing a three-dimensional display model with the output of the optimization configuration evaluation model corresponding to the plurality of different optimization combinations, is specifically as follows: Obtain several different crossover rates and mutation rates, combine them into several different optimization groups, and construct a three-dimensional display model with the output of the optimization configuration evaluation model corresponding to the different optimization combinations, specifically: Construct vector groups with different crossover rates and mutation rates respectively, and construct a rectangular coordinate system with the crossover rate and mutation rate as the vertical axis and horizontal axis respectively; The output of the optimization configuration evaluation model corresponding to each combination of crossover rate and mutation rate is used as the third axis of the plane rectangular coordinate system to construct a three-dimensional display model.

5. The multi-objective optimization configuration method for the energy storage system of the distribution network according to claim 4 is characterized in that: The curve analysis is performed to obtain the crossover rate and mutation rate after analysis, which are specifically: The maximum values ​​of several three-dimensional display models are obtained, and the derivative analysis of the slope of the change curve of each maximum value is performed to obtain the change rate. The crossover rate and mutation rate corresponding to the maximum value of the three-dimensional display model with the smallest change rate are used as the crossover rate and mutation rate after analysis, and are applied to the multi-objective optimization configuration process of the energy storage system of the distribution network.

6. The multi-objective optimization configuration method for the energy storage system of the distribution network according to claim 5, characterized in that: The specific calculation formula of the convergence speed variation coefficient is as follows: The specific calculation formula of the convergence speed coefficient is as follows: In the formula, is the convergence speed variation coefficient, is the convergence rate at the jth moment, is the total number of moments collected during the analysis of the energy storage configuration scheme, and j is the moment number; is the convergence speed coefficient.

7. The multi-objective optimization configuration method for the energy storage system of the distribution network according to claim 6, characterized in that: The specific calculation formula of the cosine similarity between the fitness values ​​is as follows: The specific calculation formula of the average cosine similarity is: The specific calculation formula of the model fitness value matching coefficient is: In the formula, is the cosine similarity between the fitness values ​​of the ith model iteration and the jth model iteration, is the kth vector value of the fitness value corresponding to the i-th model iteration, is the kth vector value of the fitness value corresponding to the jth model iteration, m is the dimension of the fitness value, is the average cosine similarity of the i-th model iteration, N is the total number of model iterations, is the fitness value of the i-th model iteration, is the fitness value of the jth model iteration, is the model fitness value matching coefficient.

Images