Distributed new energy power distribution network dispatching multi-objective optimization method and system

Abstract

The invention relates to a distributed new energy power distribution network scheduling multi-objective optimization method and system, and the method comprises the specific steps: determining a multi-objective task which comprises the maximization of the new energy utilization rate, the minimization of the power grid loss, and the maximization of the power grid stability; related attribute data influencing the target task are collected based on the target task, the related attribute data are preprocessed, and preprocessing comprises data cleaning and normalization; constructing a multi-target task prediction model based on multiple optimization algorithms, and training the multi-target task prediction model by taking the collected related attribute data influencing the multi-target tasks as input and taking the real-time measurement and calculation value of each target task as output; the multiple optimization algorithms comprise a non-dominated sorting genetic algorithm, multi-target particle swarm optimization, multi-target Bayesian optimization, Pareto optimization, a weighted sum method and a multi-target genetic algorithm.

Description

Technical Field

[0001] The present invention relates to the technical field of distribution network dispatching, and in particular to a multi-objective optimization method and system for distributed new energy distribution network dispatching. Background Art

[0002] Distributed new energy distribution network dispatching refers to the use of distributed energy resources (such as solar energy, wind energy, etc.) to produce electricity, and effectively manage and dispatch through intelligent systems to meet the electricity needs of different regions and users. With the widespread application of distributed new energy, traditional distribution networks face many challenges, such as supply and demand balance, grid stability and environmental protection requirements.

[0003] Existing dispatching strategies often find it difficult to achieve the optimal balance between economy, environmental protection and grid security. Multi-objective optimization of distributed renewable energy distribution network dispatch can not only effectively integrate clean energy resources and reduce dependence on traditional energy, but also promote the development of power systems towards intelligence and high efficiency, which has important economic and environmental significance. Summary of the invention

[0004] The purpose of the present invention is to provide a multi-objective optimization method for distributed renewable energy distribution network scheduling to solve the problems raised in the above background technology that the current traditional distribution network is facing many challenges, such as supply and demand balance, grid stability and environmental protection requirements, and the scheduling strategy is often difficult to achieve an optimal balance between economy, environmental protection and grid security.

[0005] The technical solution of the present invention is as follows:

[0006] On the one hand, the present invention proposes a multi-objective optimization method for distributed renewable energy distribution network scheduling, and the specific steps include:

[0007] Determining a multi-objective task, wherein the multi-objective task includes maximizing the utilization rate of new energy, minimizing grid losses, and maximizing grid stability;

[0008] Collect relevant attribute data affecting the target task based on the target task, and preprocess the relevant attribute data, wherein the preprocessing includes data cleaning and normalization;

[0009] Construct a multi-objective task prediction model based on multiple optimization algorithms, use the collected relevant attribute data affecting the multi-objective tasks as input, and use the real-time measured values ​​of each objective task as output to train the multi-objective task prediction model; the multiple optimization algorithms include non-dominated sorting genetic algorithm, multi-objective particle swarm optimization, multi-objective Bayesian optimization, Pareto optimization, weighted sum method and multi-objective genetic algorithm;

[0010] The Pareto optimal method is used to evaluate the prediction value of the pre-trained multi-objective task prediction model and determine the set of non-inferior solutions.

[0011] Then, the final distribution network dispatching plan and specific dispatching strategy are determined based on the set of non-inferior solutions.

[0012] As a preferred implementation, the relevant attributes affecting the target task include the generation of new energy, load demand, grid structure and parameters, equipment operating status, line parameters, transformer capacity and real-time meteorological data.

[0013] As a preferred implementation, the multiple optimization algorithms are used to optimize the objective functions of each task of the multi-objective task prediction model. Before each optimization algorithm is optimized, the grid constraints need to be considered. The grid constraints include physical limitations of the grid, equipment capacity limitations, and set policy and regulatory constraints, such as renewable energy quotas and emission standards.

[0014] As a preferred implementation, the final distribution network dispatching plan is continuously tracked, data is collected in real time, the dispatching effect is evaluated, and the feedback data is used for parameter correction and iteration of the optimization algorithm in the multi-objective task prediction model.

[0015] On the other hand, the present invention proposes a distributed renewable energy distribution network dispatching multi-objective optimization system, comprising:

[0016] A multi-objective task determination module determines a multi-objective task, wherein the multi-objective task includes maximizing the utilization rate of new energy, minimizing the power grid loss and maximizing the power grid stability;

[0017] A data collection module collects relevant attribute data affecting the target task based on the target task, and preprocesses the relevant attribute data, wherein the preprocessing includes data cleaning and normalization;

[0018] A prediction model building module is used to build a multi-objective task prediction model based on a variety of optimization algorithms, using the collected relevant attribute data affecting the multi-objective tasks as input and the real-time measured values ​​of each target task as output to train the multi-objective task prediction model; the multiple optimization algorithms include non-dominated sorting genetic algorithm, multi-objective particle swarm optimization, multi-objective Bayesian optimization, Pareto optimization, weighted sum method and multi-objective genetic algorithm;

[0019] The module for determining the optimal solution of the target task uses the Pareto optimal method to evaluate the predicted value of the pre-trained multi-target task prediction model and determine the set of non-inferior solutions;

[0020] The distribution network dispatching plan generation module then determines the final distribution network dispatching plan and specific dispatching strategy based on the set of non-inferior solutions.

[0021] As a preferred implementation, the relevant attributes affecting the target task include the generation of new energy, load demand, grid structure and parameters, equipment operating status, line parameters, transformer capacity and real-time meteorological data.

[0022] As a preferred implementation, the multiple optimization algorithms are used to optimize the objective functions of each task of the multi-objective task prediction model. Before each optimization algorithm is optimized, the grid constraints need to be considered. The grid constraints include physical limitations of the grid, equipment capacity limitations, and set policy and regulatory constraints, such as renewable energy quotas and emission standards.

[0023] As a preferred implementation, the final distribution network dispatching plan is continuously tracked, data is collected in real time, the dispatching effect is evaluated, and the feedback data is used for parameter correction and iteration of the optimization algorithm in the multi-objective task prediction model.

[0024] On the other hand, the present invention proposes an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, a multi-objective optimization method for dispatching a distributed new energy distribution network as described in any embodiment of the present invention is implemented.

[0025] On the other hand, the present invention proposes a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a multi-objective optimization method for scheduling a distributed new energy distribution network as described in any embodiment of the present invention.

[0026] The present invention has the following beneficial effects:

[0027] 1. Through multi-objective optimization, the present invention can effectively improve the operating efficiency of distributed new energy distribution networks, reduce operating costs, and effectively balance the requirements of economy, environmental impact, and system reliability.

[0028] 2. When optimizing the objective function using the optimization algorithm, the present invention takes into account the grid constraints; at the same time, when extracting relevant attribute data of the target task, it takes into account various influencing factors, thereby improving the overall operating efficiency and sustainability of the distributed new energy distribution network, and contributing to the reliable power supply and environmental protection of the power system.

[0029] 3. The present invention continuously monitors the distribution network, continuously adjusts the prediction model, and continuously optimizes the dispatching strategy, thereby effectively improving the operating efficiency and reliability of the distributed new energy distribution network, and has broad application prospects and market value. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 It is a schematic diagram of the steps of the present invention. DETAILED DESCRIPTION

[0031] 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.

[0032] It should be understood that the step numbers used in this document are only for convenience of description and are not intended to limit the order in which the steps are executed.

[0033] It should be understood that the terms used in the present specification are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the present specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0034] The terms “include” and “comprising” indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or combinations thereof.

[0035] The term "and / or" means and includes any and all possible combinations of one or more of the associated listed items.

[0036] Embodiment 1:

[0037] See also Figure 1 , a distributed renewable energy distribution network dispatching multi-objective optimization method, characterized in that the specific steps include:

[0038] Determining a multi-objective task, wherein the multi-objective task includes maximizing the utilization rate of new energy, minimizing grid losses, and maximizing grid stability;

[0039] In this embodiment, goal setting includes determining multiple goals to be optimized, such as maximizing the utilization rate of new energy, minimizing grid losses, maximizing grid stability, etc., to ensure that the specific definition and optimization direction of each goal are clear. This step can provide clear guidance for the optimization algorithm by setting different goals. Different optimization algorithms may focus on different goals, and the set goals can help select a suitable optimization algorithm or adjust algorithm parameters to more effectively achieve the expected optimization results. In multi-objective optimization, trade-offs and traction effects between different goals are usually involved. The goal setting stage can help identify these potential relationships and balance the mutual influence between different goals by setting appropriate optimization weights or constraints to obtain better solutions. At the same time, setting clear goals helps support the decision-making process. The goal set of the optimization results can provide decision makers with multiple options, enabling them to choose the most appropriate solution or decision path based on the trade-offs of specific goals.

[0040] Collect relevant attribute data affecting the target task based on the target task, and preprocess the relevant attribute data, wherein the preprocessing includes data cleaning and normalization;

[0041] In this embodiment, data collection includes the generation of new energy, load demand, grid structure and parameters, equipment operation status, etc., records the physical characteristics of the distribution network, such as line parameters, transformer capacity, etc., integrates real-time meteorological data and other relevant external information, and data preprocessing includes data cleaning, normalization and other operations. This step can ensure the data basis and quality of the optimization process through data collection and preprocessing, and provide accurate and complete data input required by the optimization model and algorithm, thereby supporting effective optimization decisions and real-time operation management. The primary task in the data collection stage is to obtain necessary real-time and historical data, such as wind power, photovoltaic power generation, grid load, battery energy storage status, etc. In addition, it is necessary to clean the data, handle outliers and missing data, and ensure the quality and integrity of the data. The purpose of cleaning and preparing data is to provide high-quality input data to ensure that subsequent optimization models and algorithms can be analyzed and decided based on accurate data, thereby improving the reliability and practicality of the optimization results. Feature extraction and data conversion can be performed through the data preprocessing stage, such as smoothing of time series data, construction of feature engineering (such as statistical features such as hourly and daily averages), and normalization or standardization of data. The purpose of feature extraction and transformation is to convert raw data into a form suitable for optimization model and algorithm processing, while retaining the key information and statistical characteristics of the data, thereby improving the efficiency and accuracy of the optimization process. By collecting and preprocessing data, a complete and reliable data set can be established. These data sets can contain historical data and real-time data, covering various key variables and parameters, such as power generation, load demand, battery status, etc., to provide input for subsequent optimization algorithms for model training and optimization decisions. The establishment of a data set needs to ensure that it covers a sufficiently long time period and sufficient data samples to support the robustness and generalization ability of the model. Preprocessed data can provide support for real-time decision-making and adjustment. These data can not only be used as model input for optimization algorithms, but also provide a basis for feedback and adjustment in real-time operation.

[0042] Construct a multi-objective task prediction model based on multiple optimization algorithms, use the collected relevant attribute data affecting the multi-objective tasks as input, and use the real-time measured values ​​of each objective task as output to train the multi-objective task prediction model; the multiple optimization algorithms include non-dominated sorting genetic algorithm, multi-objective particle swarm optimization, multi-objective Bayesian optimization, Pareto optimization, weighted sum method and multi-objective genetic algorithm;

[0043] In this embodiment, the mathematical model is established based on optimization theory, power system modeling technology, control theory and other knowledge, and is used to describe the operating behavior and objective function of the distributed new energy distribution network. The definition of the objective function includes minimizing costs, maximizing renewable energy utilization, minimizing carbon emissions, etc., and the weights and priorities are determined according to the actual situation. This step can help abstract the actual complex distributed new energy distribution network system into mathematical symbols and formulas through mathematical models, and formally describe various variables, constraints and goals. Through the formalization of the problem, complex practical problems can be accurately analyzed and solved within the mathematical framework, improving the logical rigor of problem understanding and solution design. Establishing a mathematical model helps to define the relationship and trade-off between multiple optimization objectives. By clearly defining the mutual influence and priority between different objectives, it helps decision makers understand and select the optimal solution, thereby achieving balance and comprehensive consideration in multi-objective optimization. At the same time, the establishment of a mathematical model provides a basis and standard for selecting a suitable optimization algorithm. Through the structure and characteristics of the mathematical model, the applicability and effect of different optimization algorithms in solving specific problems can be evaluated, and the most suitable optimization method can be selected to improve the calculation efficiency and the quality of the solution.

[0044] The optimization algorithms include non-dominated sorting genetic algorithm (NSGA-II), multi-objective particle swarm optimization (MOPSO), multi-objective Bayesian optimization (MOBO), Pareto optimization, weighted sum method, multi-objective genetic algorithm (MOGA), etc. Under the premise of considering the constraints of the power grid, multiple objective functions are optimized and the objective functions and constraints are encoded into the established mathematical model. The constraints include physical limitations of the power grid, equipment capacity limitations, and set policy and regulatory constraints, such as renewable energy quotas and emission standards. This step can significantly improve the computational efficiency of solving the problem through the selection of the optimization algorithm, which helps to find a solution close to the optimal solution in a reasonable time and avoid the computational burden caused by an excessively large search space. Distribution network scheduling problems usually involve multiple constraints, such as energy supply and demand balance, equipment operation restrictions, etc. Selecting a suitable algorithm can ensure that these constraints are fully considered and effectively implemented in the calculation.

[0045] The non-dominated sorting genetic algorithm is a genetic algorithm for multi-objective optimization that evaluates the distribution of solutions in its frontier. Its formula is: [d_i=\sum_{k}\frac{f_{k+1}-f_{k-1}}{f_{\text{max}}-f_{\text{min}}}] where (f_{k+1}) and (f_{k-1}) are the values ​​of the neighboring solutions of the (i)th solution on the (k)th objective function, and (f_{\text{max}}) and (f_{\text{min}}) are the maximum and minimum values ​​of the objective function, respectively.

[0046] Simulated Binary Crossover (SBX):

[0047] [x_{1}=x_{1}+\beta(x_{2}-x_{1})]

[0048] [x_{2}=x_{2}+\beta(x_{1}-x_{2})]

[0049] where (\beta) is generated according to a probability distribution.

[0050] Polynomial mutation:

[0051] [x_{i}^{'}=x_{i}+\varepsilon\cdot(x_{i}^{u}-x_{i}^{l})]

[0052] Where (\varepsilon) is a variation value generated according to the probability distribution, (x_{i}^{u}) and (x_{i}^{l}) are the upper and lower limits of the variable respectively.

[0053] NSGA-II solves multi-objective optimization problems through these formulas and steps, and performs well in terms of solution diversity and approaching the true Pareto frontier.

[0054] When dealing with multi-objective optimization problems, multi-objective particle swarm optimization usually combines the characteristics of multi-objective optimization and particle swarm optimization algorithm. The formula is:

[0055] [v_{i,d}^t=w\cdot v_{i,d}^{t-1}+c_1\cdot r_1\cdot(p_{i,d}^*-x_{i,d}^t)+c_2\cdot r_2\cdot(g_d^*-x_{i,d}^t)]

[0056] where (v_{i,d}^t) is the velocity of the (i)th particle in the (d)th dimension, (x_{i,d}^t) is the position, (p_{i,d}^) is the individual optimal position of the particle, (g_d^) is the global optimal position of the group, (w) is the inertia weight, (c_1) and (c_2) are learning factors, and (r_1) and (r_2) are random numbers.

[0057] Multi-objective Bayesian optimization is a technique used to solve optimization problems with multiple objective functions. The core idea of ​​Bayesian optimization is to guide the search process by building a probability model to efficiently find the optimal solution. In the case of multiple objectives, Bayesian optimization needs to consider the comprehensive optimization of multiple objective functions.

[0058] The prior distribution of the Gaussian process is:

[0059] [f(x)\sim\mathcal{GP}(m(x),k(x,x'))]

[0060] Among them, (m(x)) is the mean function, and (k(x,x')) is the covariance function (kernel function).

[0061] The posterior distribution of the Gaussian process is:

[0062] Given the observed data (\mathcal{D}={(x_j,\mathbf{y}j)}{j=1}^N), where (\mathbf{y}_j=[f_1(x_j),f_2(x_j),\dots,f_k(x_j)]), the posterior distribution of the Gaussian process is:

[0063] [p(f(x)\mid\mathcal{D})=\mathcal{N}\left(\mu(x),\sigma^2(x)\right)]

[0064] Among them, (\mu(x)) and (\sigma^2(x)) are the predicted mean and predicted variance respectively.

[0065] Set parameters for the selected optimization algorithm and initialize the parameters and initial solutions required by the algorithm. The parameter setting will affect the convergence speed of the algorithm and the final optimization results.

[0066] The Pareto optimal method is used to evaluate the prediction value of the pre-trained multi-objective task prediction model and determine the set of non-inferior solutions.

[0067] In this embodiment, the selected multi-objective optimization algorithm is run to start the optimization process. The algorithm generates a series of solutions and evaluates the quality of these solutions according to the set objective function.

[0068] Evaluation of solutions and determination of Pareto frontiers The generated solutions are evaluated, and the pros and cons of each solution are calculated according to the set multiple objective functions. The Pareto frontier technology is used to determine the set of non-inferior solutions, that is, those solutions that cannot be improved on any single objective. Pareto frontier refers to the set of all possible optimal solutions in multi-objective optimization problems. These optimal solutions are solutions that cannot be optimized when there is a trade-off and sacrifice relationship between different objectives. Determine multiple objectives that need to be optimized in distribution network scheduling, such as minimizing cost, maximizing reliability, etc.; construct the objective function into a multi-objective optimization model, which may contain multiple decision variables and constraints; use appropriate multi-objective optimization algorithms, such as multi-objective genetic algorithms, multi-objective particle swarm algorithms, etc., to search and find all possible Pareto optimal solutions; all found Pareto optimal solutions constitute the Pareto frontier. These solutions are non-inferior in multiple objective dimensions, that is, it is impossible to improve other objectives at the same time by improving one of them; according to specific decision-making needs and conditions, the optimal solution is selected from the Pareto frontier. The Pareto frontier can help decision makers make reasonable trade-offs between different objectives (such as cost, reliability, environmental impact, etc.) and find the optimal operating strategy or configuration plan.

[0069] Then, the final distribution network dispatching plan and specific dispatching strategy are determined based on the set of non-inferior solutions.

[0070] In this embodiment, the result analysis and dispatch strategy generation generates specific dispatch strategies by analyzing the solution set in the Pareto front, such as the output plan of distributed energy, the charging and discharging plan of the energy storage system, demand-side management measures, etc., and selects the optimal solution or provides decision support to determine the final distribution network dispatch plan.

[0071] As a preferred implementation mode of this embodiment, the relevant attributes affecting the target task include the generation of new energy, load demand, grid structure and parameters, equipment operating status, line parameters, transformer capacity and real-time meteorological data.

[0072] As a preferred implementation mode of this embodiment, the multiple optimization algorithms are used to optimize the objective functions of each task of the multi-objective task prediction model. Before each optimization algorithm is optimized, the grid constraints need to be considered. The grid constraints include physical limitations of the grid, equipment capacity limitations, and set policy and regulatory constraints, such as renewable energy quotas and emission standards.

[0073] As a preferred implementation method of this embodiment, the final distribution network dispatching plan is continuously tracked, data is collected in real time, the dispatching effect is evaluated, and the feedback data is used for parameter correction and iteration of the optimization algorithm in the multi-objective task prediction model.

[0074] In this embodiment, the finalized optimization scheme is implemented in the actual distributed renewable energy distribution network, and its effect is continuously monitored, and the dispatching strategy is adjusted according to the real-time situation to deal with uncertain factors; the feedback and optimization step collects actual data after the dispatching execution, evaluates the dispatching effect, and uses the feedback data for model correction and optimization algorithm iteration to improve the accuracy and effect of future dispatching, so as to continuously improve the efficiency and reliability of the dispatching of the distributed renewable energy distribution network.

[0075] Embodiment 2:

[0076] A distributed renewable energy distribution network dispatching multi-objective optimization system, comprising:

[0077] A multi-objective task determination module determines a multi-objective task, wherein the multi-objective task includes maximizing the utilization rate of new energy, minimizing the power grid loss and maximizing the power grid stability;

[0078] A data collection module collects relevant attribute data affecting the target task based on the target task, and preprocesses the relevant attribute data, wherein the preprocessing includes data cleaning and normalization;

[0079] A prediction model building module is used to build a multi-objective task prediction model based on a variety of optimization algorithms, using the collected relevant attribute data affecting the multi-objective tasks as input and the real-time measured values ​​of each target task as output to train the multi-objective task prediction model; the multiple optimization algorithms include non-dominated sorting genetic algorithm, multi-objective particle swarm optimization, multi-objective Bayesian optimization, Pareto optimization, weighted sum method and multi-objective genetic algorithm;

[0080] The module for determining the optimal solution of the target task uses the Pareto optimal method to evaluate the predicted value of the pre-trained multi-target task prediction model and determine the set of non-inferior solutions;

[0081] The distribution network dispatching plan generation module then determines the final distribution network dispatching plan and specific dispatching strategy based on the set of non-inferior solutions.

[0082] As a preferred implementation mode of this embodiment, the relevant attributes affecting the target task include the generation of new energy, load demand, grid structure and parameters, equipment operating status, line parameters, transformer capacity and real-time meteorological data.

[0083] As a preferred implementation mode of this embodiment, the multiple optimization algorithms are used to optimize the objective functions of each task of the multi-objective task prediction model. Before each optimization algorithm is optimized, the grid constraints need to be considered. The grid constraints include physical limitations of the grid, equipment capacity limitations, and set policy and regulatory constraints, such as renewable energy quotas and emission standards.

[0084] As a preferred implementation method of this embodiment, the final distribution network dispatching plan is continuously tracked, data is collected in real time, the dispatching effect is evaluated, and the feedback data is used for parameter correction and iteration of the optimization algorithm in the multi-objective task prediction model.

[0085] Embodiment three:

[0086] This embodiment provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, a multi-objective optimization method for scheduling a distributed new energy distribution network as described in any embodiment of the present invention is implemented.

[0087] Embodiment 4:

[0088] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements a multi-objective optimization method for scheduling a distributed new energy distribution network as described in any embodiment of the present invention.

[0089] The above descriptions are merely embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A multi-objective optimization method for distributed renewable energy distribution network scheduling, characterized in that: The specific steps include: Determining a multi-objective task, wherein the multi-objective task includes maximizing the utilization rate of new energy, minimizing grid losses, and maximizing grid stability; Collect relevant attribute data affecting the target task based on the target task, and preprocess the relevant attribute data, wherein the preprocessing includes data cleaning and normalization; Construct a multi-objective task prediction model based on multiple optimization algorithms, use the collected relevant attribute data affecting the multi-objective tasks as input, and use the real-time measured values ​​of each objective task as output to train the multi-objective task prediction model; the multiple optimization algorithms include non-dominated sorting genetic algorithm, multi-objective particle swarm optimization, multi-objective Bayesian optimization, Pareto optimization, weighted sum method and multi-objective genetic algorithm; The Pareto optimal method is used to evaluate the prediction value of the pre-trained multi-objective task prediction model and determine the set of non-inferior solutions. Then, the final distribution network dispatching plan and specific dispatching strategy are determined based on the set of non-inferior solutions.

2. A multi-objective optimization method for distributed renewable energy distribution network scheduling according to claim 1, characterized in that: The relevant attributes affecting the target task include the generation of new energy, load demand, grid structure and parameters, equipment operating status, line parameters, transformer capacity and real-time meteorological data.

3. A distributed renewable energy distribution network dispatching multi-objective optimization method according to claim 1, characterized in that: The multiple optimization algorithms are used to optimize the objective functions of each task of the multi-objective task prediction model. Before each optimization algorithm is optimized, the grid constraints need to be considered. The grid constraints include physical limitations of the grid, equipment capacity limitations, and set policy and regulatory constraints, such as renewable energy quotas and emission standards.

4. A distributed renewable energy distribution network dispatching multi-objective optimization method according to claim 1, characterized in that: The final distribution network dispatching plan is continuously tracked, data is collected in real time, the dispatching effect is evaluated, and the feedback data is used to correct parameters in the multi-objective task prediction model and iterate the optimization algorithm.

5. A distributed renewable energy distribution network dispatching multi-objective optimization system, characterized in that: include: A multi-objective task determination module determines a multi-objective task, wherein the multi-objective task includes maximizing the utilization rate of new energy, minimizing the power grid loss and maximizing the power grid stability; A data collection module collects relevant attribute data affecting the target task based on the target task, and preprocesses the relevant attribute data, wherein the preprocessing includes data cleaning and normalization; A prediction model building module is used to build a multi-objective task prediction model based on a variety of optimization algorithms, using the collected relevant attribute data affecting the multi-objective tasks as input and the real-time measured values ​​of each target task as output to train the multi-objective task prediction model; the multiple optimization algorithms include non-dominated sorting genetic algorithm, multi-objective particle swarm optimization, multi-objective Bayesian optimization, Pareto optimization, weighted sum method and multi-objective genetic algorithm; The module for determining the optimal solution of the target task uses the Pareto optimal method to evaluate the predicted value of the pre-trained multi-target task prediction model and determine the set of non-inferior solutions; The distribution network dispatching plan generation module then determines the final distribution network dispatching plan and specific dispatching strategy based on the set of non-inferior solutions.

6. A distributed renewable energy distribution network dispatching multi-objective optimization system according to claim 5, characterized in that: The relevant attributes affecting the target task include the generation of new energy, load demand, grid structure and parameters, equipment operating status, line parameters, transformer capacity and real-time meteorological data.

7. A distributed renewable energy distribution network dispatching multi-objective optimization system according to claim 5, characterized in that: The multiple optimization algorithms are used to optimize the objective functions of each task of the multi-objective task prediction model. Before each optimization algorithm is optimized, the grid constraints need to be considered. The grid constraints include physical limitations of the grid, equipment capacity limitations, and set policy and regulatory constraints, such as renewable energy quotas and emission standards.

8. A distributed renewable energy distribution network dispatching multi-objective optimization system according to claim 5, characterized in that: The final distribution network dispatching plan is continuously tracked, data is collected in real time, the dispatching effect is evaluated, and the feedback data is used to correct parameters in the multi-objective task prediction model and iterate the optimization algorithm.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, a distributed new energy distribution network scheduling multi-objective optimization system as described in any one of claims 1 to 4 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, a distributed new energy distribution network scheduling multi-objective optimization system as described in any one of claims 1 to 4 is implemented.

Images