Microgrid Coordination Optimization Method and System to Meet the Requirements of Friendly Interaction with Large Power Grids

Abstract

The present invention provides a microgrid coordination optimization method and system that meet the requirements of friendly interaction of large power grids, which relates to the field of power supply systems. The method includes: determining the operation correlation characteristics among multiple microgrids based on the historical operation information of the multiple microgrids included in the microgrid group; determining multiple typical large power grid interaction scenarios based on the historical interaction information between the multiple microgrids included in the microgrid group and the large power grid, and combining the operation correlation characteristics among the multiple microgrids to perform configuration optimization on the multiple microgrids included in the microgrid group; determining the real-time interaction requirements between the large power grid and the microgrid based on the real-time operation information of the large power grid and the real-time operation information of the multiple microgrids included in the microgrid group; determining the microgrid coordination strategy based on the real-time interaction requirements between the large power grid and the microgrid and the operation correlation characteristics among the multiple microgrids; and scheduling the microgrid group based on the microgrid coordination strategy, which has the advantage of improving the quality of microgrid coordination.

Description

Technical Field

[0001] The present invention relates to the field of power supply systems, and particularly to a microgrid coordination optimization method and system that meet the requirements of friendly interaction with large power grids. Background Art

[0002] With the continuous maturity of microgrid technology and the advancement of power market reform, microgrids have gradually become important participants in the power market. Traditionally, microgrids operate in the mode of "self-generation for self-use and surplus power feeding into the grid", and their main benefits rely on energy supply. In the main energy market, microgrids can guide the operation of internal distributed energy through electricity prices by participating in power trading. In the ancillary service market, microgrids can flexibly dispatch power supply and demand because they integrate various distributed energy resources, such as photovoltaic, wind energy, and energy storage. The local management and flexible operation mode of microgrids enable them to quickly respond to market changes, adjust power generation and consumption, and thus optimize economic benefits. Therefore, microgrids can also provide ancillary services such as peak shaving and frequency modulation, with great potential.

[0003] In the prior art, some studies on microgrid scheduling optimization have established linear scheduling models, but these models are too rough and not conducive to in-depth research. Some studies on microgrid scheduling optimization have established non-linear optimization models and solved them using intelligent optimization algorithms such as particle swarm optimization, ant colony optimization, and genetic algorithm. Generally, these models are only applicable to a few microgrid cases, and the applicability of the models needs to be improved.

[0004] Therefore, there is a need to provide a microgrid coordination optimization method and system that meet the requirements of friendly interaction with large power grids to improve the quality of microgrid coordination. Summary of the Invention

[0005] The present invention provides a microgrid coordination optimization method that meets the requirements of friendly interaction with large power grids, including: obtaining the historical operation information of multiple microgrids included in a microgrid group; determining the operation correlation characteristics between multiple microgrids based on the historical operation information of multiple microgrids included in the microgrid group; obtaining the historical interaction information between multiple microgrids included in the microgrid group and the large power grid; determining multiple typical large power grid interaction scenarios based on the historical interaction information between multiple microgrids included in the microgrid group and the large power grid; optimizing the configuration of multiple microgrids included in the microgrid group based on the operation correlation characteristics between multiple microgrids and multiple typical large power grid interaction scenarios; obtaining the real-time operation information of the large power grid and the real-time operation information of multiple microgrids included in the microgrid group; determining the real-time interaction requirements between the large power grid and the microgrids based on the real-time operation information of the large power grid and the real-time operation information of multiple microgrids included in the microgrid group; determining the microgrid coordination strategy based on the real-time interaction requirements between the large power grid and the microgrids and the operation correlation characteristics between multiple microgrids; and scheduling the microgrid group based on the microgrid coordination strategy.

[0006] Further, the historical operation information of the microgrid at least includes the photovoltaic output, wind turbine output, and load of the microgrid in multiple historical time periods; based on the historical operation information of multiple microgrids included in the microgrid group, determine the operation correlation characteristics between the multiple microgrids, including: for any two microgrids, calculate the photovoltaic output correlation coefficient between the two microgrids based on the photovoltaic output of the two microgrids in multiple historical time periods, calculate the wind turbine output correlation coefficient between the two microgrids based on the wind turbine output of the two microgrids in multiple historical time periods, calculate the load correlation coefficient between the two microgrids based on the load of the two microgrids in multiple historical time periods, and calculate the comprehensive output correlation coefficient between the two microgrids based on the photovoltaic output correlation coefficient and the wind turbine output correlation coefficient of the two microgrids; based on the comprehensive output correlation coefficient and the load correlation coefficient between any two microgrids, group the multiple microgrids to determine multiple microgrid groups; for any two microgrid groups, determine the operation complementary coefficient between the two microgrid groups based on the comprehensive output correlation coefficient and the load correlation coefficient between any one microgrid included in one microgrid group and any one microgrid included in the other microgrid group; based on the operation complementary coefficient between any two microgrid groups, determine the complementary microgrid group for each microgrid group.

[0007] Further, based on the comprehensive output correlation coefficient and the load correlation coefficient between any two microgrids, group the multiple microgrids to determine multiple microgrid groups, including: determine the first k value according to the comprehensive output correlation coefficient between any two microgrids; according to the first k value, cluster the multiple microgrids based on the comprehensive output correlation coefficient between any two microgrids through a clustering algorithm to determine multiple microgrid clusters; for each microgrid cluster, determine the second k value corresponding to the microgrid cluster according to the load correlation coefficient between any two microgrids included in the microgrid cluster, and according to the second k value corresponding to the microgrid cluster, cluster the multiple microgrids included in the microgrid cluster based on the load correlation coefficient between any two microgrids included in the microgrid cluster through a clustering algorithm to determine multiple microgrid groups included in the microgrid cluster.

[0008] Further, determining a first k value according to the comprehensive output correlation coefficients of any two microgrids includes: obtaining multiple groups of first samples, where each first sample consists of the comprehensive output correlation coefficients of any two sample microgrids included in a sample microgrid group and the optimal first k value; for each first sample, sorting the comprehensive output correlation coefficients of any two sample microgrids included in the sample microgrid group from smallest to largest to generate a comprehensive output correlation coefficient sequence, performing variational mode decomposition on the comprehensive output correlation coefficient sequence to generate multiple sequence components, and determining the eigenvalues of multiple sequence component factors; based on the eigenvalues of multiple sequence component factors corresponding to each first sample and the optimal first k value, determining multiple target sequence component factors from multiple sequence component factors; based on multiple target sequence component factors and multiple groups of first samples, generating multiple groups of second samples, where each second sample includes the eigenvalues of multiple target sequence component factors and the optimal first k value; based on multiple target sequence component factors and multiple groups of second samples, establishing and training a first k value prediction model; sorting the comprehensive output correlation coefficients of any two microgrids from smallest to largest to generate a comprehensive output correlation coefficient sequence, performing variational mode decomposition on the comprehensive output correlation coefficient sequence to generate multiple sequence components, and determining the eigenvalues of multiple target sequence component factors; determining the first k value through the first k value prediction model based on the eigenvalues of multiple target sequence component factors.

[0009] Further, determining multiple typical large grid interaction scenarios based on the historical interaction information between multiple microgrids included in a microgrid group and the large grid includes: determining multiple large grid interaction scenarios based on the historical interaction information between multiple microgrids included in a microgrid group and the large grid; calculating the scenario similarity between any two large grid interaction scenarios; clustering multiple large grid interaction scenarios based on the scenario similarity between any two large grid interaction scenarios; and determining multiple typical large grid interaction scenarios according to the clustering results.

[0010] Further, optimizing the configuration of multiple microgrids included in the microgrid group based on the operation correlation characteristics between multiple microgrids and multiple typical large grid interaction scenarios includes: determining multiple optimization indicators and the index weights corresponding to each optimization indicator; generating multiple configuration optimization schemes; for each configuration optimization scheme, determining the scores of multiple optimization indicators of the configuration optimization scheme in each typical large grid interaction scenario based on the operation correlation characteristics between multiple microgrids; generating an optimal configuration optimization scheme through a genetic algorithm based on the scores of multiple optimization indicators of each configuration optimization scheme in each typical large grid interaction scenario; and optimizing the configuration of multiple microgrids included in the microgrid group based on the optimal configuration optimization scheme.

[0011] Further, obtain the real-time operation information of multiple microgrids, including: based on the photovoltaic output correlation coefficients and wind turbine output correlation coefficients of any two microgrids, predict the future photovoltaic output information and future wind turbine output information of each microgrid; based on the load correlation coefficients of any two microgrids, predict the future load information of each microgrid.

[0012] Further, based on the real-time operation information of the large power grid and the real-time operation information of multiple microgrids included in the microgrid cluster, determine the real-time interaction requirements between the large power grid and the microgrids, including: based on the real-time operation information of the large power grid, predict the real-time interaction requirements of the large power grid; for each microgrid group, based on the future photovoltaic output information, future wind turbine output information and future load information of each microgrid included in the microgrid group, predict the real-time interaction requirements of the microgrid group.

[0013] Further, based on the real-time interaction requirements between the large power grid and the microgrids and the operation correlation characteristics among multiple microgrids, determine the microgrid coordination strategy, including: establish a policy generation model based on the deep Q-learning network; through the policy generation model, based on the real-time interaction requirements between the large power grid and the microgrids and the operation correlation characteristics among multiple microgrids, determine the microgrid coordination strategy.

[0014] The present invention provides a microgrid coordination and optimization system adapted to the friendly interaction requirements of the large power grid, which is used to execute the above-mentioned microgrid coordination and optimization method adapted to the friendly interaction requirements of the large power grid, including: an operation analysis module, which is used to obtain the historical operation information of multiple microgrids included in the microgrid cluster, and based on the historical operation information of multiple microgrids included in the microgrid cluster, determine the operation correlation characteristics among multiple microgrids; an interaction analysis module, which is used to obtain the historical interaction information between multiple microgrids included in the microgrid cluster and the large power grid, and based on the historical interaction information between multiple microgrids included in the microgrid cluster and the large power grid, determine multiple typical large power grid interaction scenarios; a configuration optimization module, which is used to perform configuration optimization on multiple microgrids included in the microgrid cluster based on the operation correlation characteristics among multiple microgrids and multiple typical large power grid interaction scenarios; an information acquisition module, which is used to obtain the real-time operation information of the large power grid and the real-time operation information of multiple microgrids included in the microgrid cluster; a demand determination module, which is used to determine the real-time interaction requirements between the large power grid and the microgrids based on the real-time operation information of the large power grid and the real-time operation information of multiple microgrids included in the microgrid cluster; a policy formulation module, which is used to determine the microgrid coordination strategy based on the real-time interaction requirements between the large power grid and the microgrids and the operation correlation characteristics among multiple microgrids; a policy execution module, which is used to schedule the microgrid cluster based on the microgrid coordination strategy.

[0015] Compared with the prior art, the microgrid coordination and optimization method and system provided by the present invention adapted to the friendly interaction requirements of the large power grid have at least the following beneficial effects:

[0016] 1. By analyzing the operational correlation characteristics among multiple microgrids, the mutual influences and dependencies between microgrids can be identified, enabling more precise configuration optimization. Based on multiple typical large-grid interaction scenarios determined from historical interaction information, potential changes in power demand can be foreseen, and resource allocation can be carried out in advance to improve energy utilization efficiency. The microgrid coordination strategy formulated according to real-time interaction requirements and operational correlation characteristics can ensure the coordinated operation among microgrids, avoiding resource waste and duplicate construction. By optimizing the operation strategy of microgrids, the local consumption of new energy sources (such as photovoltaic and wind power) can be promoted, reducing the phenomena of curtailment of photovoltaic power and curtailment of wind power. By implementing peak shaving and valley filling and demand management strategies, the energy storage can be discharged or distributed power sources can be preferentially used during peak power demand periods, reducing the dependence on fossil energy and achieving energy conservation and emission reduction.

[0017] 2. By calculating the correlation coefficients of photovoltaic output, wind turbine output, and load between any two microgrids, the operational correlation characteristics and complementarity between microgrids can be accurately evaluated. Grouping microgrids based on these correlation coefficients can ensure that the microgrids within a group have similar operational characteristics, facilitating subsequent coordinated optimization. Further determining the complementary microgrid groups for each microgrid group helps to achieve resource complementarity and load balancing among different microgrids. Based on the grouping, a more flexible scheduling plan between microgrids can be formulated to ensure the reasonable allocation and efficient utilization of resources. Especially during periods when the output of clean energy is insufficient or during peak load hours, through the coordinated scheduling among microgrids, the contradiction between supply and demand can be balanced, improving the stability and reliability of the entire microgrid cluster. The coordinated operation among microgrids can form a more stable and reliable power supply network. When a certain microgrid fails or the output of clean energy is insufficient, other microgrids can quickly provide support, reducing the power outage time and losses. Through the coordinated optimization among microgrids, clean energy can be better utilized, reducing the dependence on traditional energy. This helps to promote the transformation and upgrading of the energy structure and achieve sustainable development.

[0018] 3. Based on the eigenvalues and the optimal first k values of multiple sequence component factors, multiple target sequence component factors can be determined, and then the first k value prediction model can be established and trained. Through the first k value prediction model, the optimal k value can be predicted according to the eigenvalues of the comprehensive output correlation coefficient sequences of any two microgrids. This ensures the scientificity and accuracy of the k value selection during the grouping process. After determining the first k value, the clustering algorithm can be used to cluster multiple microgrids based on the comprehensive output correlation coefficient to form multiple microgrid clusters. Since the selection of the k value is more accurate, the clustering process will be more efficient, and the clustering results will be more in line with the actual operation requirements. Traditional clustering algorithms often require multiple attempts and adjustments when selecting the k value, which increases the computational complexity and time cost. However, the first k value determined by this method can be directly applied to the clustering algorithm without multiple attempts and adjustments, thus reducing the computational complexity. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] This specification will be further described in the form of exemplary embodiments, which will be described in detail through the drawings. These embodiments are not restrictive. In these embodiments, the same numbers represent the same structures, where:

[0020] Figure 1 is a schematic flowchart of a microgrid coordination optimization method that meets the requirements of friendly interaction with large power grids according to some embodiments of this specification;

[0021] Figure 2 is a schematic flowchart of determining the first k value according to some embodiments of this specification;

[0022] Figure 3 is a schematic block diagram of a microgrid coordination optimization system that meets the requirements of friendly interaction with large power grids according to some embodiments of this specification. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0023] To more clearly illustrate the technical solutions of the embodiments of this specification, the drawings required for the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some examples or embodiments of this specification. For those of ordinary skill in the art, without creative efforts, this specification can also be applied to other similar scenarios based on these drawings. Unless obvious from the language context or otherwise stated, the same reference numerals in the figures represent the same structures or operations.

[0024] Figure 1 is a schematic flowchart of a microgrid coordination optimization method that meets the requirements of friendly interaction with large power grids according to some embodiments of this specification. As Figure 1 shown, the microgrid coordination optimization method that meets the requirements of friendly interaction with large power grids may include the following steps:

[0025] S101. Obtain the historical operation information of multiple microgrids included in the microgrid cluster.

[0026] Among them, the historical operation information of the microgrid includes at least the photovoltaic output, wind turbine output, and load of the microgrid in multiple historical time periods.

[0027] S102. Based on the historical operation information of multiple microgrids included in the microgrid cluster, determine the operation correlation characteristics among the multiple microgrids.

[0028] In some embodiments, S102 may specifically include:

[0029] For any two microgrids, calculate the photovoltaic output correlation coefficient of the two microgrids based on the photovoltaic output of the two microgrids in multiple historical time periods, calculate the wind turbine output correlation coefficient of the two microgrids based on the wind turbine output of the two microgrids in multiple historical time periods, calculate the load correlation coefficient of the two microgrids based on the load of the two microgrids in multiple historical time periods, and calculate the comprehensive output correlation coefficient of the two microgrids based on the photovoltaic output correlation coefficient and the wind turbine output correlation coefficient of the two microgrids;

[0030] Group the multiple microgrids based on the comprehensive output correlation coefficient and the load correlation coefficient of any two microgrids, and determine multiple microgrid groups;

[0031] For any two microgrid groups, determine the operation complementary coefficient of the two microgrid groups based on the comprehensive output correlation coefficient and the load correlation coefficient of any one microgrid included in one microgrid group and any one microgrid included in the other microgrid group;

[0032] Based on the operation complementary coefficient of any two microgrid groups, determine the complementary microgrid group of each microgrid group. Among them, when the operation complementary coefficient of the two microgrid groups is greater than the operation complementary coefficient threshold, the two microgrid groups are complementary microgrid groups to each other.

[0033] Specifically, the Pearson correlation coefficient of the photovoltaic output of the two microgrids can be calculated based on the photovoltaic output of the two microgrids in multiple historical time periods as the photovoltaic output correlation coefficient of the two microgrids. The calculation methods of the wind turbine output correlation coefficient and the load correlation coefficient are similar to the calculation method of the photovoltaic output correlation coefficient, and will not be elaborated here.

[0034] The photovoltaic output correlation coefficient and the wind turbine output correlation coefficient of the two microgrids can be weighted and summed as the comprehensive output correlation coefficient of the two microgrids.

[0035] In some embodiments, grouping the multiple microgrids based on the comprehensive output correlation coefficient and the load correlation coefficient of any two microgrids to determine multiple microgrid groups includes:

[0036] Determine a first k value according to the comprehensive output correlation coefficients of any two microgrids, where the first k value is a positive integer;

[0037] According to the first k value, cluster multiple microgrids based on the comprehensive output correlation coefficients of any two microgrids through a clustering algorithm (such as K-Means, etc.) to determine multiple microgrid clusters;

[0038] For each microgrid cluster, determine a second k value corresponding to the microgrid cluster according to the load correlation coefficients of any two microgrids included in the microgrid cluster. According to the second k value corresponding to the microgrid cluster, cluster the multiple microgrids included in the microgrid cluster based on the load correlation coefficients of any two microgrids included in the microgrid cluster through a clustering algorithm to determine multiple microgrid groups included in the microgrid cluster, where the second k value is a positive integer.

[0039] Figure 2 It is a schematic flow chart of determining the first k value shown in some embodiments of this specification. As Figure 2 shown, in some embodiments, determining the first k value according to the comprehensive output correlation coefficients of any two microgrids includes:

[0040] Obtain multiple groups of first samples, where each first sample consists of the comprehensive output correlation coefficients of any two sample microgrids included in the sample microgrid group and the optimal first k value;

[0041] For each first sample, sort the comprehensive output correlation coefficients of any two sample microgrids included in the sample microgrid group from smallest to largest to generate a comprehensive output correlation coefficient sequence, perform variational mode decomposition on the comprehensive output correlation coefficient sequence to generate multiple sequence components, and determine the eigenvalues of multiple sequence component factors (central frequency, bandwidth, mean, variance, extreme values, etc.);

[0042] Based on the eigenvalues of multiple sequence component factors corresponding to each first sample and the optimal first k value, determine multiple target sequence component factors from multiple sequence component factors;

[0043] Based on multiple target sequence component factors and multiple groups of first samples, generate multiple groups of second samples, where each second sample includes the eigenvalues of multiple target sequence component factors and the optimal first k value;

[0044] Based on multiple target sequence component factors and multiple groups of second samples, establish and train a first k value prediction model, where the first k value prediction model can be a random forest regression model;

[0045] Sort the comprehensive output correlation coefficients of any two microgrids in ascending order to generate a comprehensive output correlation coefficient sequence. Perform variational mode decomposition on the comprehensive output correlation coefficient sequence to generate multiple sequence components, and determine the eigenvalues of multiple target sequence component factors.

[0046] Based on the eigenvalues of multiple target sequence component factors, determine the first k value through the first k value prediction model.

[0047] Specifically, for each sequence component factor, the Pearson correlation coefficient between the sequence component factor and the optimal first k value can be calculated according to the eigenvalue of the sequence component factor in each first sample and the optimal first k value of each first sample. The sequence component factors with Pearson correlation coefficients greater than the Pearson correlation coefficient threshold are used as target sequence component factors.

[0048] The following process can be used to determine multiple microgrid clusters:

[0049] S1021. Randomly select the number of microgrids corresponding to the first k value as the initial central microgrids.

[0050] S1022. Traverse each microgrid in the dataset and assign each microgrid to the microgrid cluster where the central microgrid with the smallest comprehensive output correlation coefficient is located.

[0051] S1023. For each microgrid cluster, recalculate its central microgrid. Among them, the new central microgrid can be the microgrid with the smallest average comprehensive output correlation coefficient of other microgrids included in the microgrid cluster.

[0052] S1024. Repeat steps S1021 - S1023 until a certain stop condition is met. The stop conditions include: the number of microgrids whose microgrid clusters change in two adjacent iterations is less than the first quantity threshold, or the central microgrids of the microgrid clusters no longer change in two adjacent iterations, or the set maximum number of iterations is reached.

[0053] The method for determining the second k value is similar to the method for determining the first k value, and will not be elaborated here.

[0054] Based on the load correlation coefficients of any two microgrids included in the microgrid cluster, perform clustering on the multiple microgrids included in the microgrid cluster. The method for determining the multiple microgrid groups included in the microgrid cluster is similar to the method for determining multiple microgrid clusters, and will not be elaborated here.

[0055] The operating complementary coefficient between two microgrid groups can be calculated according to the following formula:

[0056] ,

[0057] where is the operation complementary coefficient between the $i$-th microgrid group and the $j$-th microgrid group. is the comprehensive output correlation coefficient between the $m$-th microgrid included in the $i$-th microgrid group and the $n$-th microgrid included in the $j$-th microgrid group. is the load correlation coefficient between the $m$-th microgrid included in the $i$-th microgrid group and the $n$-th microgrid included in the $j$-th microgrid group. is the total number of microgrids included in the $i$-th microgrid group. is the total number of microgrids included in the $j$-th microgrid group.

[0058] S103. Obtain the historical interaction information between multiple microgrids included in the microgrid cluster and the large power grid.

[0059] Specifically, the historical interaction information between the microgrid and the large power grid may include the operation information of the large power grid (such as voltage, frequency, load, etc.) during the time period when the microgrid interacts with the large power grid, the operation information of the microgrid (such as operation mode (such as grid-connected model, island mode, etc.), PV output, wind turbine output, stored electricity, load, etc.), and the power exchange information (such as total exchanged power, exchange power, etc.).

[0060] S104. Based on the historical interaction information between multiple microgrids included in the microgrid cluster and the large power grid, determine multiple typical large power grid interaction scenarios.

[0061] In some embodiments, S104 specifically includes:

[0062] Based on the historical interaction information between multiple microgrids included in the microgrid cluster and the large power grid, determine multiple large power grid interaction scenarios;

[0063] Calculate the scenario similarity between any two large power grid interaction scenarios;

[0064] Based on the scenario similarity between any two large power grid interaction scenarios, cluster multiple large power grid interaction scenarios;

[0065] According to the clustering result, determine multiple typical large power grid interaction scenarios.

[0066] Specifically, first, based on the historical interaction information between multiple microgrids included in the microgrid cluster and the main grid, multiple possible main grid interaction scenarios are initially identified and determined. These main grid interaction scenarios can have characteristics such as time period, weather conditions, main grid load demand, main grid of each microgrid, distributed energy output of each microgrid, energy storage situation of each microgrid, and load demand of each microgrid. In order to find scenarios with similar characteristics and perform clustering, it is necessary to calculate the scenario similarity between any two main grid interaction scenarios. The scenario similarity can be evaluated by comparing the similarity degrees of key parameters such as main grid load demand, main grid of each microgrid, distributed energy output of each microgrid, energy storage situation of each microgrid, and load demand of each microgrid under different scenarios. The calculation methods include Euclidean distance, cosine similarity, Pearson correlation coefficient, etc.

[0067] After calculating the scenario similarity, clustering algorithms (such as K-means clustering, hierarchical clustering, etc.) can be used to cluster multiple main grid interaction scenarios. The purpose of clustering is to group scenarios with similar characteristics into one category, thereby simplifying the subsequent analysis and decision-making processes. The clustering result will form multiple clustering clusters, and each cluster represents a typical main grid interaction scenario.

[0068] Finally, according to the clustering result, one or more representative scenarios are selected from each clustering cluster as typical main grid interaction scenarios. These typical scenarios should be able to comprehensively and accurately reflect the interaction characteristics between the microgrid cluster and the main grid under different conditions. When selecting representative scenarios, factors such as the frequency of scenario occurrence, scenario diversity, and the degree of impact of the scenario on grid operation can be considered.

[0069] S105. Optimize the configuration of multiple microgrids included in the microgrid cluster based on the operating correlation characteristics between multiple microgrids and multiple typical main grid interaction scenarios.

[0070] In some embodiments, S105 specifically includes:

[0071] Determine multiple optimization indicators and the corresponding indicator weights for each optimization indicator. Among them, the multiple optimization indicators can at least include a cost indicator, an interaction matching indicator, and a self - balance indicator of the micro - grid cluster. The cost indicator focuses on the equipment procurement, installation, operation, maintenance, and possible decommissioning costs of the micro - grid cluster, etc.; The interaction matching indicator evaluates the interaction performance and matching degree between the micro - grid cluster and the large power grid, which includes the matching in aspects such as power balance, voltage stability, and frequency stability; The self - balance indicator of the micro - grid cluster focuses on the self - sufficiency ability of the micro - grid cluster when operating independently (island mode), that is, the ability of each micro - grid within the micro - grid cluster to achieve power supply - demand balance through coordinated control. Algorithms such as the analytic hierarchy process, order relation analysis method, entropy weight method, and multiple correlation coefficient method can be used to determine the corresponding indicator weights for each optimization indicator;

[0072] Generate multiple configuration optimization schemes. Among them, the configuration optimization scheme can include the configuration optimization of photovoltaic power generation, wind power generation, controllable power sources, energy storage devices, etc. in the micro - grid. Multiple methods such as random generation and heuristic search can be used to generate multiple configuration optimization schemes;

[0073] For each configuration optimization scheme, based on the operation correlation characteristics among multiple micro - grids, determine the scores of multiple optimization indicators of the configuration optimization scheme in each typical large - power - grid interaction scenario;

[0074] Through the genetic algorithm, based on the scores of multiple optimization indicators of each configuration optimization scheme in each typical large - power - grid interaction scenario, generate the optimal configuration optimization scheme;

[0075] Based on the optimal configuration optimization scheme, optimize the configuration of multiple micro - grids included in the micro - grid cluster.

[0076] Specifically, the scores of the configuration optimization scheme in the optimization indicators can be determined according to the following process:

[0077] I. Determine the components of the micro - grid and their costs

[0078] 1. Photovoltaic power station:

[0079] Evaluate the total installed capacity of the photovoltaic power station.

[0080] Calculate the investment cost of the photovoltaic power station according to the unit price and required quantity of photovoltaic panels.

[0081] Consider the maintenance cost of the photovoltaic power station, including cleaning, repair, and replacement of damaged panels, etc.

[0082] 2. Wind power generation:

[0083] Evaluate the total installed capacity of wind turbines: First, determine the scale and expected power generation of the wind power project. Based on the wind resources and terrain conditions, select the appropriate model and quantity of wind turbines. Calculate the total installed capacity of all wind turbines, which is the sum of the rated powers of all generators.

[0084] Calculate the investment cost of wind turbines: Obtain the unit price of the selected wind turbines, which usually includes the costs of the generator body, tower, blades, control system, etc. Calculate the total purchase cost of the wind turbines according to the required quantity. Add the infrastructure construction costs, such as land fees, foundation construction, and transmission line laying. Consider the installation and commissioning costs, including hoisting, wiring, and commissioning operation.

[0085] Consider the maintenance cost of wind power generation: Daily maintenance costs, including regular inspections, cleaning of blades, lubrication of bearings, and inspection of the electrical system. Repair costs, when the wind turbine fails or components are damaged, repairs or replacements are required. Replacement costs, as the usage time increases, some key components (such as blades, gearboxes, generators) may need to be replaced.

[0086] 3. Controllable power source:

[0087] Determine the total power and type of the controllable power source (such as diesel generator, gas turbine, etc.).

[0088] Calculate its investment cost according to the unit capacity cost of the controllable power source.

[0089] Consider the fuel cost, maintenance cost, and operation cost of the controllable power source.

[0090] 4. Energy storage device:

[0091] Determine the rated power and total energy storage capacity of the energy storage device.

[0092] Calculate its investment cost according to the unit energy cost and unit capacity cost of the energy storage device.

[0093] Consider the maintenance cost of the energy storage device, including battery replacement, system upgrade, etc.

[0094] 5. Inverter:

[0095] Select a suitable inverter according to the power requirements of the photovoltaic power station and the controllable power source.

[0096] Calculate the investment cost of the inverter.

[0097] Consider the maintenance cost of the inverter.

[0098] II. Calculate the annual average construction cost

[0099] Add up the investment costs of the above - mentioned parts to obtain the total construction cost of the micro - grid.

[0100] Calculate the annual average construction cost of the micro - grid according to the service life of the equipment and the discount rate.

[0101] III. Evaluate the annual operating cost

[0102] 1. Operating cost of controllable power sources:

[0103] Calculate its fuel consumption and operating cost according to the output power and operating time of the controllable power source.

[0104] Consider the maintenance cost of the controllable power source.

[0105] 2. Operating cost of energy storage devices:

[0106] Calculate its operating cost according to the charge - discharge power and operating time of the energy storage device.

[0107] Consider the maintenance cost of the energy storage device.

[0108] 3. Operating cost of loads:

[0109] Calculate the operating cost of the load according to the load demand of the micro - grid and the electricity purchase price.

[0110] Consider the regulation cost of controllable loads and the electricity consumption cost of users.

[0111] IV. Calculate the total cost over the entire life cycle

[0112] Add up the annual average construction cost and the annual operating cost to obtain the total cost over the entire life cycle of the micro - grid.

[0113] Under the configuration optimization plan, the higher the total cost over the entire life cycle of the micro - grid, the lower the score of the configuration optimization plan in terms of the cost index.

[0114] For each configuration optimization plan, use power system simulation software or data analysis tools to conduct simulation and data analysis on the configuration optimization plan. By simulating the operation of the microgrid under different typical large grid interaction scenarios, collect the simulated operation data. Based on the simulated operation data of different typical large grid interaction scenarios, determine the interaction performance and matching degree between the microgrid group operating in grid-connected mode and the large grid under different typical large grid interaction scenarios, which includes the matching in aspects such as power balance, voltage stability, and frequency stability, so as to determine the score of the configuration optimization plan in the interaction matching index. At the same time, based on the simulated operation data of different typical large grid interaction scenarios, determine the electric energy that the microgrid group operating in grid-connected mode needs to purchase from the large grid, the electric energy discarded or lacking inside the microgrid group operating in island mode. The more electric energy that the microgrid group operating in grid-connected mode needs to purchase from the large grid, and the more electric energy discarded or lacking inside the microgrid group operating in island mode, the lower the score of the configuration optimization plan in the self-balancing index of the microgrid group. Among them, the operating correlation characteristics between multiple microgrids are used to guide the generation of the electric energy exchange strategy inside the microgrid group.

[0115] After obtaining the scores of multiple optimization indexes of each configuration optimization plan under each typical large grid interaction scenario, optimization algorithms such as genetic algorithms can be used to search for the optimal configuration optimization plan. The genetic algorithm is an optimization algorithm that simulates natural selection and genetic mechanisms. It continuously iteratively optimizes the individuals in the solution space through operations such as selection, crossover, and mutation, and finally finds an approximate optimal solution. In this process, each configuration optimization plan is regarded as an individual, and its score is used as the fitness function value to guide the search process. Finally, based on the optimal configuration optimization plan obtained by the genetic algorithm, actual configuration optimization is carried out on multiple microgrids in the microgrid group. The optimized microgrid group will be able to better adapt to different typical large grid interaction scenarios and improve the overall energy efficiency, stability, and reliability.

[0116] S106. Obtain the real-time operation information of the large grid and the real-time operation information of multiple microgrids included in the microgrid group.

[0117] In some embodiments, obtaining the real-time operation information of multiple microgrids includes:

[0118] Based on the photovoltaic output correlation coefficient and wind turbine output correlation coefficient between any two microgrids, predict the future photovoltaic output information and future wind turbine output information of each microgrid;

[0119] Based on the load correlation coefficient between any two microgrids, predict the future load information of each microgrid.

[0120] Specifically, the future photovoltaic output information, future wind turbine output information, and future load information of each microgrid can be predicted according to the following process:

[0121] S1061. Determine the microgrids with positive photovoltaic output correlation and the microgrids with negative photovoltaic output correlation for each microgrid based on the photovoltaic output correlation coefficients between any two microgrids. Among them, the photovoltaic output correlation coefficient between the microgrids with positive photovoltaic output correlation and the microgrid is greater than the positive photovoltaic output correlation coefficient threshold, and the photovoltaic output correlation coefficient between the microgrids with negative photovoltaic output correlation and the microgrid is less than the negative photovoltaic output correlation coefficient threshold.

[0122] S1062. For each microgrid, based on the photovoltaic output of the microgrid's positive photovoltaic output correlation microgrids in multiple historical time periods of the current power generation cycle, the photovoltaic output of the microgrid in multiple historical time periods of the current power generation cycle, and the photovoltaic output of the microgrid's negative photovoltaic output correlation microgrids in multiple historical time periods of the current power generation cycle, predict the future photovoltaic output information of the microgrid through a photovoltaic output prediction model. Among them, the photovoltaic output prediction model can be a long short-term memory network model.

[0123] The method for predicting the future wind turbine output information and future load information of each microgrid is similar to the method for predicting the future photovoltaic output information of each microgrid, and will not be elaborated here.

[0124] S107. Determine the real-time interaction requirements between the large power grid and the microgrids based on the real-time operation information of the large power grid and the real-time operation information of multiple microgrids included in the microgrid group.

[0125] In some embodiments, S107 specifically includes:

[0126] Predict the real-time interaction requirements of the large power grid based on the real-time operation information of the large power grid. Specifically, the real-time operation information of the large power grid includes, but is not limited to, key parameters such as voltage, current, frequency, power factor, and load demand. The real-time interaction requirements of the large power grid can be predicted through a first demand prediction model based on the real-time operation information of the large power grid. The first demand prediction model takes into account influencing factors such as the historical operation data of the power grid, weather conditions, and economic activities, so as to be able to more accurately predict the power demand of the large power grid in a future time period. The first demand prediction model can be a long short-term memory network model;

[0127] For each microgrid group, predict the real-time interaction requirements of the microgrid group based on the future photovoltaic output information, future wind turbine output information, and future load information of each microgrid included in the microgrid group. For example, the real-time interaction requirements of each microgrid group can be predicted through a second demand prediction model based on the future photovoltaic output information, future wind turbine output information, and future load information of each microgrid included in the microgrid group. Among them, the second demand prediction model can be a long short-term memory network model.

[0128] S108. Determine the microgrid coordination strategy based on the real-time interaction requirements between the large power grid and the microgrids and the operation correlation characteristics among multiple microgrids.

[0129] In some embodiments, S108 specifically includes:

[0130] Establish a policy generation model based on the Deep Q-Network (DQN);

[0131] Based on the real-time interaction requirements between the large power grid and the microgrids and the operation correlation characteristics among multiple microgrids, determine the microgrid coordination strategy through the policy generation model.

[0132] Specifically, the policy generation model may include:

[0133] Input layer: Receive the real-time interaction requirements of the large power grid, the real-time interaction requirements of each microgrid group, and the complementary microgrid groups of each microgrid group as input information.

[0134] Hidden layer: Process and extract features from the input information through a multi-layer neural network.

[0135] Output layer: Output the Q value (i.e., the expected return) corresponding to each possible microgrid coordination strategy. The Q value reflects the long-term benefits that the system can achieve after executing a certain strategy. In the policy generation model, the greater the operation complementary coefficient between the microgrid group and the complementary microgrid group, the higher the Q value corresponding to the microgrid coordination strategy.

[0136] Train the policy generation model using historical data or simulation data. During the training process, the policy generation model will try different microgrid coordination strategies and adjust its strategy selection according to the actual returns (such as power quality, system stability, economy, etc.). Through continuous iteration and optimization, the policy generation model can learn one or more optimal microgrid coordination strategies.

[0137] After the policy generation model is trained, it can be used to determine the microgrid coordination strategy according to the real-time interaction requirements between the large power grid and the microgrids and the operation correlation characteristics among the microgrids.

[0138] Input the real-time interaction requirements of the large power grid, the real-time interaction requirements of each microgrid group, and the complementary microgrid groups of each microgrid group into the policy generation model. The policy generation model calculates the Q value of each possible microgrid coordination strategy according to the input information. Select the strategy with the highest Q value as the current microgrid coordination strategy. In the microgrid coordination strategy, the greater the operation complementary coefficient between the microgrid group and the complementary microgrid group, the higher the Q value corresponding to the microgrid coordination strategy, and the higher the matching degree between the electric energy exchanged between the microgrid group and the large power grid and the real-time interaction requirements of the large power grid, the higher the Q value corresponding to the microgrid coordination strategy.

[0139] S109. Schedule the microgrid group based on the microgrid coordination strategy.

[0140] Figure 3 It is a schematic diagram of the modules of the microgrid coordination and optimization system that meets the requirements of friendly interaction with the large power grid as shown in some embodiments of this specification. Figure 3 As shown, the microgrid coordination and optimization system that meets the requirements of friendly interaction with the large power grid may include an operation analysis module, an interaction analysis module, a configuration optimization module, an information acquisition module, a demand determination module, a strategy formulation module, and a strategy execution module.

[0141] The operation analysis module is used to obtain the historical operation information of multiple microgrids included in the microgrid group, and determine the operation correlation characteristics among the multiple microgrids based on the historical operation information of the multiple microgrids included in the microgrid group.

[0142] The interaction analysis module is used to obtain the historical interaction information between multiple microgrids included in the microgrid group and the large power grid, and determine multiple typical large power grid interaction scenarios based on the historical interaction information between the multiple microgrids included in the microgrid group and the large power grid.

[0143] The configuration optimization module is used to optimize the configuration of multiple microgrids included in the microgrid group based on the operation correlation characteristics among the multiple microgrids and multiple typical large power grid interaction scenarios.

[0144] The information acquisition module is used to obtain the real-time operation information of the large power grid and the real-time operation information of multiple microgrids included in the microgrid group.

[0145] The demand determination module is used to determine the real-time interaction demand between the large power grid and the microgrid based on the real-time operation information of the large power grid and the real-time operation information of multiple microgrids included in the microgrid group.

[0146] The strategy formulation module is used to determine the microgrid coordination strategy based on the real-time interaction demand between the large power grid and the microgrid and the operation correlation characteristics among the multiple microgrids.

[0147] The strategy execution module is used to schedule the microgrid group based on the microgrid coordination strategy.

[0148] The microgrid coordination and optimization system that meets the requirements of friendly interaction with the large power grid can be used to execute the microgrid coordination and optimization method that meets the requirements of friendly interaction with the large power grid, which will not be elaborated here.

[0149] Finally, it should be understood that the embodiments described in this specification are only used to illustrate the principles of the embodiments of this specification. Other variations may also fall within the scope of this specification. Therefore, by way of example and not limitation, alternative configurations of the embodiments of this specification may be regarded as consistent with the teachings of this specification. Accordingly, the embodiments of this specification are not limited to the embodiments explicitly presented and described in this specification.

Claims

1. A microgrid coordination optimization method adapted to the requirements of large power grid friendly interaction, characterized in that Including: Obtaining historical operation information of multiple microgrids included in a microgrid cluster; Determining operation correlation characteristics among the multiple microgrids based on the historical operation information of the multiple microgrids included in the microgrid cluster; Obtaining historical interaction information between the multiple microgrids included in the microgrid cluster and a large power grid; Determining multiple typical large power grid interaction scenarios based on the historical interaction information between the multiple microgrids included in the microgrid cluster and the large power grid; Performing configuration optimization on the multiple microgrids included in the microgrid cluster based on the operation correlation characteristics among the multiple microgrids and the multiple typical large power grid interaction scenarios; Obtaining real-time operation information of the large power grid and future operation information of the multiple microgrids included in the microgrid cluster; Determining future interaction requirements between the large power grid and the microgrids based on the real-time operation information of the large power grid and the future operation information of the multiple microgrids included in the microgrid cluster; Determining a microgrid coordination strategy based on the future interaction requirements between the large power grid and the microgrids and the operation correlation characteristics among the multiple microgrids; Scheduling the microgrid cluster based on the microgrid coordination strategy.

2. The microgrid coordinated optimization method for adapting to the friendly interaction requirements of large power grids according to claim 1, wherein The historical operation information of the microgrid at least includes the photovoltaic output, wind turbine output and load of the microgrid in multiple historical time periods; Determining operation correlation characteristics among the multiple microgrids based on the historical operation information of the multiple microgrids included in the microgrid cluster, including: For any two microgrids, calculating the photovoltaic output correlation coefficient of the two microgrids based on the photovoltaic outputs of the two microgrids in multiple historical time periods, calculating the wind turbine output correlation coefficient of the two microgrids based on the wind turbine outputs of the two microgrids in multiple historical time periods, calculating the load correlation coefficient of the two microgrids based on the loads of the two microgrids in multiple historical time periods, calculating the comprehensive output correlation coefficient of the two microgrids based on the photovoltaic output correlation coefficient and the wind turbine output correlation coefficient of the two microgrids, performing weighted summation on the photovoltaic output correlation coefficient and the wind turbine output correlation coefficient of the two microgrids as the comprehensive output correlation coefficient of the two microgrids, and calculating the operation complementary coefficient of the two microgrid groups according to the following formula: ； Among them, is the operation complementary coefficient between the $i$-th microgrid group and the $j$-th microgrid group, is the comprehensive output correlation coefficient between the $m$-th microgrid included in the $i$-th microgrid group and the $n$-th microgrid included in the $j$-th microgrid group, is the load correlation coefficient between the $m$-th microgrid included in the $i$-th microgrid group and the $n$-th microgrid included in the $j$-th microgrid group, is the total number of microgrids included in the $i$-th microgrid group, is the total number of microgrids included in the $j$-th microgrid group; Grouping the multiple microgrids based on the comprehensive output correlation coefficient and the load correlation coefficient of any two microgrids to determine multiple microgrid groups; For any two microgrid groups, determining the operation complementary coefficient of the two microgrid groups based on the comprehensive output correlation coefficient and the load correlation coefficient of any one microgrid included in one microgrid group and any one microgrid included in the other microgrid group; Determining the complementary microgrid group of each microgrid group based on the operation complementary coefficient of any two microgrid groups.

3. The microgrid coordinated optimization method for adapting to the friendly interaction requirements of large power grids according to claim 2, wherein, Grouping the multiple microgrids based on the comprehensive output correlation coefficient and the load correlation coefficient of any two microgrids to determine multiple microgrid groups, including: Determining a first k value according to the comprehensive output correlation coefficient of any two microgrids; Based on the first k value, clustering the multiple microgrids based on the comprehensive output correlation coefficient of any two microgrids through a clustering algorithm to determine multiple microgrid clusters; For each microgrid cluster, determine the second k value corresponding to the microgrid cluster according to the load correlation coefficient between any two microgrids included in the microgrid cluster. Based on the second k value corresponding to the microgrid cluster, cluster the multiple microgrids included in the microgrid cluster based on the load correlation coefficient between any two microgrids included in the microgrid cluster through a clustering algorithm, and determine multiple microgrid groups included in the microgrid cluster.

4. The microgrid coordinated optimization method for meeting the friendly interaction requirements of large power grids according to claim 3, characterized in that, Determining the first k value according to the comprehensive output correlation coefficient between any two microgrids includes: Obtain multiple groups of first samples, where the first sample consists of the comprehensive output correlation coefficient between any two sample microgrids included in the sample microgrid group and the optimal first k value; For each first sample, sort the comprehensive output correlation coefficients between any two sample microgrids included in the sample microgrid group from smallest to largest to generate a comprehensive output correlation coefficient sequence, perform variational mode decomposition on the comprehensive output correlation coefficient sequence to generate multiple sequence components, and determine the eigenvalues of the multiple sequence component factors; Based on the eigenvalues of the multiple sequence component factors corresponding to each first sample and the optimal first k value, determine multiple target sequence component factors from the multiple sequence component factors; Based on the multiple target sequence component factors and multiple groups of first samples, generate multiple groups of second samples, where the second sample includes the eigenvalues of the multiple target sequence component factors and the optimal first k value; Based on the multiple target sequence component factors and multiple groups of second samples, establish and train a first k value prediction model; Sort the comprehensive output correlation coefficients between any two microgrids from smallest to largest to generate a comprehensive output correlation coefficient sequence, perform variational mode decomposition on the comprehensive output correlation coefficient sequence to generate multiple sequence components, and determine the eigenvalues of the multiple target sequence component factors; Determine the first k value through the first k value prediction model based on the eigenvalues of the multiple target sequence component factors.

5. The microgrid coordinated optimization method for adapting to the friendly interaction requirements of large power grids according to any one of claims 2-4, characterized in that Based on the historical interaction information between the multiple microgrids included in the microgrid group and the large power grid, determine multiple typical large power grid interaction scenarios, including: Based on the historical interaction information between the multiple microgrids included in the microgrid group and the large power grid, determine multiple large power grid interaction scenarios; Calculate the scenario similarity between any two large power grid interaction scenarios; Cluster the multiple large power grid interaction scenarios based on the scenario similarity between any two large power grid interaction scenarios; According to the clustering result, determine multiple typical large power grid interaction scenarios.

6. The microgrid coordination optimization method for meeting the friendly interaction requirements of large power grids according to claim 5, characterized in that, Based on the operation correlation characteristics between the multiple microgrids and the multiple typical large power grid interaction scenarios, perform configuration optimization on the multiple microgrids included in the microgrid group, including: Determine multiple optimization indicators and the index weights corresponding to each optimization indicator; Generate multiple configuration optimization schemes; For each configuration optimization scheme, based on the operation correlation characteristics between the multiple microgrids, determine the scores of the multiple optimization indicators of the configuration optimization scheme under each typical large power grid interaction scenario; Through the genetic algorithm, based on the scores of the multiple optimization indicators of each configuration optimization scheme under each typical large power grid interaction scenario, generate an optimal configuration optimization scheme; Based on the optimal configuration optimization scheme, perform configuration optimization on the multiple microgrids included in the microgrid group.

7. The microgrid coordination and optimization method for meeting the requirements of friendly interaction with large power grids according to claim 6, characterized in that, Obtain the future operation information of multiple microgrids, including: Predict the future photovoltaic output information and future wind turbine output information of each microgrid based on the correlation coefficients of photovoltaic output and wind turbine output between any two microgrids. Predict the future load information of each microgrid based on the load correlation coefficients between any two microgrids.

8. The microgrid coordinated optimization method for adapting to the friendly interaction requirements of large power grids according to claim 7, wherein Determine the future interaction requirements between the large power grid and the microgrids based on the real-time operation information of the large power grid and the future operation information of multiple microgrids included in the microgrid cluster, including: Predict the future interaction requirements of the large power grid based on the real-time operation information of the large power grid. For each microgrid group, predict the future interaction requirements of the microgrid group based on the future photovoltaic output information, future wind turbine output information, and future load information of each microgrid included in the microgrid group.

9. The microgrid coordination optimization method for adapting to the friendly interaction requirements of large power grids according to claim 8, characterized in that Determine the microgrid coordination strategy based on the future interaction requirements between the large power grid and the microgrids and the operation correlation characteristics among multiple microgrids, including: Establish a policy generation model based on the deep Q-learning network. Determine the microgrid coordination strategy through the policy generation model based on the future interaction requirements between the large power grid and the microgrids and the operation correlation characteristics among multiple microgrids.

10. A microgrid coordinated optimization system adapted to the friendly interaction requirements of large power grids, characterized in that, Used to execute the microgrid coordination optimization method for adapting to the friendly interaction requirements of the large power grid described in any one of claims 1-9, including: An operation analysis module, used to obtain the historical operation information of multiple microgrids included in the microgrid cluster, and determine the operation correlation characteristics among multiple microgrids based on the historical operation information of multiple microgrids included in the microgrid cluster. An interaction analysis module, used to obtain the historical interaction information between multiple microgrids included in the microgrid cluster and the large power grid, and determine multiple typical large power grid interaction scenarios based on the historical interaction information between multiple microgrids included in the microgrid cluster and the large power grid. A configuration optimization module, used to perform configuration optimization on multiple microgrids included in the microgrid cluster based on the operation correlation characteristics among multiple microgrids and multiple typical large power grid interaction scenarios. An information acquisition module, used to obtain the real-time operation information of the large power grid and the future operation information of multiple microgrids included in the microgrid cluster. A demand determination module, used to determine the future interaction requirements between the large power grid and the microgrids based on the real-time operation information of the large power grid and the future operation information of multiple microgrids included in the microgrid cluster. A policy formulation module, used to determine the microgrid coordination strategy based on the future interaction requirements between the large power grid and the microgrids and the operation correlation characteristics among multiple microgrids. A policy execution module, used to schedule the microgrid cluster based on the microgrid coordination strategy.

Images