Multi-scene-oriented source-load-storage aggregation scheme generation method

Abstract

The invention discloses a multi-scene-oriented source-load-storage aggregation scheme generation method. The method comprises the following steps: acquiring a plurality of feature index data; a plurality of preset scene fitness evaluation indexes; obtaining scores of a plurality of resources and / or resource combinations on each preset scene fitness evaluation index to serve as a first score group of each resource and / or resource combination; according to the multiple pieces of index data, multiple scenes and scene features corresponding to each scene are obtained through calculation; according to the multiple scenes and the corresponding scene features, a random forest algorithm is utilized to calculate weights of the multiple preset scene fitness evaluation indexes, and the weights of the multiple preset scene fitness evaluation indexes in each scene are obtained; according to the first score group of each resource and / or resource combination and the weight of a plurality of preset scene fitness evaluation indexes in each scene, calculating to obtain a fitness matrix of the plurality of scenes and the plurality of resources and / or resource combinations; and obtaining an optimal scheme under each scene according to the adaptation degree matrix of the plurality of scenes and the plurality of resources and / or resource combinations. The method has the characteristic of high resource utilization efficiency.

Description

Technical Field

[0001] This invention relates to the field of power system technology, and more specifically, to a method for generating source-load-storage aggregation schemes for multiple scenarios. Background Technology

[0002] With the widespread application of resources such as distributed photovoltaics, energy storage, and flexible loads, power systems need to match specific scenarios for these resources in order to achieve efficient resource utilization.

[0003] Traditional scenario segmentation methods often rely on only one or a few dimensions, such as weather type, load curve shape, or resource type, and fail to systematically integrate the complex coupling relationship between resource physical characteristics, external environmental conditions, and market policy rules. Their simplified segmentation methods cannot truly reflect the typical operating state formed by the interaction of multiple factors such as source, load, storage, environment, and policy, resulting in a disconnect between the constructed scenarios and actual operating needs, thus leading to low resource utilization efficiency.

[0004] Existing technology discloses a hierarchical collaborative control method for multiple aggregated distributed resource clusters based on source, load, and storage, including the following steps: classifying the control levels of multiple aggregated distributed resources based on source, load, and storage; based on the control level classification results, aggregating resources with the same control level into the same resource cluster to construct a multiple aggregated distributed resource cluster based on source, load, and storage; establishing a collaborative control model for the distributed resource cluster; and introducing a graph neural network model to solve the collaborative control model for the distributed resource cluster. This method does not consider the application scenario, which means it cannot maintain high resource utilization efficiency in multiple scenarios. Summary of the Invention

[0005] This invention addresses the shortcomings of existing technologies in terms of low resource utilization efficiency across various scenarios by providing a method for generating source-load-storage aggregation schemes for multiple scenarios. This method features high resource utilization efficiency.

[0006] The primary objective of this invention is to solve the aforementioned technical problems. The technical solution of this invention is as follows: A method for generating source-load-storage aggregation schemes for multiple scenarios includes: S1: Obtain multiple feature index data; multiple preset scene adaptability evaluation indicators; obtain the scores of multiple resources and / or resource combinations on each preset scene adaptability evaluation indicator, as the first score group for each resource and / or resource combination; S2: Based on the multiple indicator data, calculate multiple scenarios and the scenario features corresponding to each scenario; S3: Based on the multiple scenarios and corresponding scenario features, use the random forest algorithm to calculate the weights of the multiple preset scenario fitness evaluation indicators to obtain the weights of the multiple preset scenario fitness evaluation indicators under each scenario. S4: Based on the first scoring group of each resource and / or resource combination and the weights of multiple preset scenario adaptability evaluation indicators in each scenario, calculate the adaptability matrix of multiple scenarios with multiple resources and / or resource combinations. S5: Based on the fit matrix of the multiple scenarios and multiple resources and / or resource combinations, obtain the optimal solution for each scenario.

[0007] Further, in step S2, based on the multiple indicator data, multiple scenarios and scene features corresponding to each scenario are calculated, including: S201: Standardize each of the aforementioned feature index data to obtain multiple standardized feature index data; S202: Perform principal component dimensionality reduction on each of the standardized feature index data to obtain multiple dimensionality-reduced feature index data; S203: Perform cluster analysis on the multiple dimensionality-reduced feature index data to obtain multiple scenarios and the scene features corresponding to each scenario.

[0008] Furthermore, in step S203, the cluster analysis includes: S20301: The elbow method is used to process multiple dimensionality-reduced feature index data to obtain the number of clusters; S20302: Based on the number of clusters, the K-Means clustering algorithm is used to cluster the multiple dimensionality-reduced feature index data to obtain multiple scenarios and multiple feature index data corresponding to each scenario; S20303: Average the multiple feature index data corresponding to each scenario to form cluster center feature index data corresponding to multiple scenarios; S20304: Compare the cluster center feature index data corresponding to each scenario with the first preset threshold vector to obtain the scenario features corresponding to each scenario.

[0009] Furthermore, the characteristic indicator data includes: characteristic data, environmental data, and policy data for each resource.

[0010] Furthermore, the preset scenario adaptability evaluation indicators include: resource availability indicators, response speed indicators, and cost-effectiveness indicators.

[0011] Further, in step S4, an adaptation matrix is ​​calculated for multiple scenes with multiple resources and / or resource combinations, including: S401: Based on the first scoring group of each resource and / or resource combination and the weights of multiple preset scenario adaptability evaluation indicators in each scenario, calculate the comprehensive adaptability score of each resource and / or resource combination in each scenario; and combine them to obtain the first intermediate matrix; S402: Normalize the first intermediate matrix to obtain a normalized first intermediate matrix, which serves as the adaptation matrix for multiple scenarios and multiple resources and / or resource combinations.

[0012] Furthermore, in step S401, the formula for calculating the overall fit score for each scenario is as follows:

[0013] This indicates the total number of the first scoring group representing resources and / or resource combinations. Indicates the sequence number of the first scoring group of resources and / or resource combinations. This represents the weight of the a-th preset scenario adaptability evaluation index. This represents the a-th value in the first rating group of resources and / or resource combinations.

[0014] Further, in step S5, based on the fit matrix of the multiple scenarios and multiple resources and / or resource combinations, the optimal solution for each scenario is obtained, including: S501: Based on the fit matrix of the multiple scenarios and multiple resources and / or resource combinations, select the resources and / or resource combinations in each scenario whose fit is greater than the second preset threshold, and use them as the first resource group for each scenario. S502: A multi-objective optimization algorithm is used to optimize and generate the solution for the first resource group of each scenario, so as to obtain the optimal solution for each scenario.

[0015] Furthermore, in step S502, the multi-objective optimization algorithm includes: S50201: Based on the first resource group of each scenario, generate multiple resource allocation genes, scheduling strategy genes, and economic genes corresponding to each scenario; S50202: Generate multiple chromosomes corresponding to each scenario based on the multiple resource allocation genes, scheduling strategy genes, and economic genes corresponding to each scenario; S50203: Using a genetic algorithm, optimize multiple chromosomes for each scenario to obtain the optimal chromosome for each scenario; S50204: Based on the optimal chromosome for each scenario, obtain the optimal solution for each scenario.

[0016] A source-load-storage aggregation solution generation system for multiple scenarios includes: Data acquisition module: acquires multiple feature index data; multiple preset scene adaptability evaluation indicators; acquires the scores of multiple resources and / or resource combinations on each preset scene adaptability evaluation indicator, as the first score group for each resource and / or resource combination; Scene calculation module: Based on the multiple indicator data, calculate multiple scenes and the scene features corresponding to each scene; Weight calculation module: Based on the multiple scenarios and corresponding scenario features, the random forest algorithm is used to calculate the weights of the multiple preset scenario fitness evaluation indicators to obtain the weights of the multiple preset scenario fitness evaluation indicators under each scenario; Adaptability matrix calculation module: Based on the first scoring group of each resource and / or resource combination and the weights of multiple preset scenario adaptability evaluation indicators in each scenario, the adaptability matrix of multiple scenarios with multiple resources and / or resource combinations is calculated. Solution generation module: Based on the adaptability matrix of the multiple scenarios and multiple resources and / or resource combinations, obtain the optimal solution for each scenario.

[0017] Compared with the prior art, the beneficial effects of the present invention are: This invention calculates the scene characteristics of each scene by using multiple feature index data, which can comprehensively cover the differences in the coupling of multiple characteristics of source, load and storage, and comprehensively consider multiple factors to more accurately evaluate the aggregation characteristics of resources in different scenes.

[0018] By calculating the weights of multiple preset scenario adaptability evaluation indicators for each scenario, and combining the scores of multiple resources and / or resource combinations on each preset scenario adaptability evaluation indicator, the degree of matching between resource capabilities and scenario requirements is quantified, the applicability of resources in different scenarios is accurately assessed, and a scientific basis for resource aggregation is provided.

[0019] In summary, this invention features high resource utilization efficiency. Attached Figure Description

[0020] Figure 1 The flowchart is provided for Example 1, which describes a method for generating source-load-storage aggregation schemes for multiple scenarios.

[0021] Figure 2 The flowchart provided in Example 1 shows the calculation of multiple scenarios and the scene features corresponding to each scenario.

[0022] Figure 3 The flowchart for cluster analysis provided in Example 1.

[0023] Figure 4 The flowchart for calculating the fitness matrix provided in Example 1 is shown.

[0024] Figure 5 The flowchart for obtaining the optimal solution in each scenario is provided in Example 1.

[0025] Figure 6 The flowchart is for the multi-objective optimization algorithm provided in Example 1.

[0026] Figure 7 This is a schematic diagram of a source-load-storage aggregation scheme generation method for multiple scenarios provided in Example 1. Detailed Implementation

[0027] The accompanying drawings are for illustrative purposes only and should not be construed as limiting the scope of this patent. To better illustrate this embodiment, some parts in the accompanying drawings may be omitted, enlarged, or reduced, and do not represent the actual product dimensions; It will be understood by those skilled in the art that certain well-known structures and their descriptions may be omitted in the accompanying drawings.

[0028] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0029] Example 1 like Figure 1 As shown, a method for generating source-load-storage aggregation schemes for multiple scenarios includes: S1: Obtain multiple feature index data; multiple preset scene adaptability evaluation indicators; obtain the scores of multiple resources and / or resource combinations on each preset scene adaptability evaluation indicator, as the first score group for each resource and / or resource combination; S2: Based on the multiple indicator data, calculate multiple scenarios and the scenario features corresponding to each scenario; S3: Based on the multiple scenarios and corresponding scenario features, use the random forest algorithm to calculate the weights of the multiple preset scenario fitness evaluation indicators to obtain the weights of the multiple preset scenario fitness evaluation indicators under each scenario. S4: Based on the first scoring group of each resource and / or resource combination and the weights of multiple preset scenario adaptability evaluation indicators in each scenario, calculate the adaptability matrix of multiple scenarios with multiple resources and / or resource combinations. S5: Based on the fit matrix of the multiple scenarios and multiple resources and / or resource combinations, obtain the optimal solution for each scenario.

[0030] It should be noted that by collecting characteristic data of distributed photovoltaic, energy storage, flexible loads, and other resources, as well as environmental and policy data, multi-source data is integrated into a multi-dimensional characteristic indicator system, constructing a comprehensive multi-dimensional characteristic indicator system. This not only fully covers the differences in the coupling of multiple characteristics of source, load, and storage, but also solves the problem that traditional methods rely on a single dimension and cannot comprehensively consider multiple factors. This provides a richer data foundation for subsequent scenario division and resource aggregation characteristic analysis, and helps to more accurately evaluate the aggregation characteristics of resources in different scenarios, thereby optimizing resource allocation and scheduling strategies.

[0031] Furthermore, such as Figure 2 As shown, in step S2, based on the multiple indicator data, multiple scenarios and scene features corresponding to each scenario are calculated, including: S201: Standardize each of the aforementioned feature index data to obtain multiple standardized feature index data; S202: Perform principal component dimensionality reduction on each of the standardized feature index data to obtain multiple dimensionality-reduced feature index data; S203: Perform cluster analysis on the multiple dimensionality-reduced feature index data to obtain multiple scenarios and the scene features corresponding to each scenario.

[0032] It should be noted that by using principal component analysis to comprehensively analyze data from different dimensions, multiple typical scenarios are identified. Each typical scenario reflects a specific combination of source-load-storage aggregation characteristics and environmental policy conditions, in order to solve the problem that traditional scenario classification relies on a single dimension and cannot cover the differences in the coupling of multiple characteristics.

[0033] In step S201, the collected multi-dimensional characteristic index data are standardized, including resource characteristic data, environmental data and policy data. Data with different dimensions and ranges are converted into unified standardized values ​​to eliminate the influence of dimensions and facilitate subsequent analysis. In one specific embodiment, assume the following data is collected:

[0034] The above data is standardized by converting each characteristic index into a standardized value with a mean of 0 and a standard deviation of 1. Assuming that the mean of "photovoltaic power output fluctuation" is 0.6 and the standard deviation is 0.2, the standardized value of sample 1 is: (0.8-0.6) / 0.2=1.0.

[0035] In step S202, the covariance matrix of the standardized data is calculated to quantify the correlation between the various feature indicators; the eigenvalues ​​and eigenvectors of the covariance matrix are solved, where the eigenvalues ​​represent the variance contribution of each principal component and the eigenvectors represent the direction of the principal components; assuming that the obtained eigenvalues ​​are λ1, λ2, ..., λ8, and the corresponding eigenvectors are v1, v2, ..., v8.

[0036] Based on the magnitude of the eigenvalues, select the top few principal components whose cumulative contribution rate reaches a preset proportion to achieve data dimensionality reduction; project the standardized data onto the selected principal component directions to obtain the dimensionality-reduced data; for example: assuming the first 3 eigenvalues ​​are 1.5, 1.2, and 0.8, with a cumulative contribution rate of 90%, calculate the covariance matrix to obtain the eigenvalues ​​and eigenvectors, assuming the following: Eigenvalue 1 = 1.5, eigenvector 1 = [0.6, 0.5, 0.6] (principal component 1); Eigenvalue 2 = 1.2, eigenvector 2 = [0.4, 0.7, 0.5] (principal component 2); Eigenvalue 3 = 0.8, eigenvector 1 = [0.5, 0.4, 0.7] (principal component 3); For sample 1, assuming principal components 1, 2, and 3 are selected from the table above, the data is projected onto the principal component directions; the projection value of sample 1 is calculated as follows: Projected value 1 = [0.8, 0.6, 0.7] [0.6, 0.5, 0.6] = 0.9 Projected value 2 = [0.8, 0.6, 0.7] [0.4, 0.7, 0.5] = 0.7 Projected value 3 = [0.8, 0.6, 0.7] [0.5, 0.4, 0.7] = 0.6 Therefore, the projection values ​​of sample 1 onto principal components 1, 2, and 3 are 0.9, 0.7, and 0.6, respectively.

[0037] Sample 1 is a part of the feature index data.

[0038] Furthermore, such as Figure 3 As shown, in step S203, the cluster analysis includes: S20301: The elbow method is used to process multiple dimensionality-reduced feature index data to obtain the number of clusters; S20302: Based on the number of clusters, the K-Means clustering algorithm is used to cluster the multiple dimensionality-reduced feature index data to obtain multiple scenarios and multiple feature index data corresponding to each scenario; S20303: Average the multiple feature index data corresponding to each scenario to form cluster center feature index data corresponding to multiple scenarios; S20304: Compare the cluster center feature index data corresponding to each scenario with the first preset threshold vector to obtain the scenario features corresponding to each scenario.

[0039] Assuming the elbow appears in 3 clusters, we choose to divide the data into 3 typical scenarios.

[0040] Based on the clustering results, the data is divided into several typical scenarios, each reflecting a specific combination of source-load-storage aggregation characteristics and environmental policy conditions. Assume that the data is divided into 3 typical scenarios: Scenario 1: Samples 1, 3, and 5; Scenario 2: Samples 2 and 4; Scenario 3: Other samples.

[0041] Based on the clustering results, feature descriptions are performed for each typical scenario, including: calculating the mean of samples within the scenario across each feature dimension to form the cluster center vector for that scenario; comparing the values ​​of each dimension of the cluster center vector with preset thresholds to generate qualitative labels describing the scenario's resource characteristics, environmental conditions, and policy background; and finally outputting a structured report containing the scenario's cluster centers, sample distribution, and feature descriptions to form a scenario feature description report. For example: Scenario 1: High light intensity, low wind speed, peak industrial load, preferential subsidy policies. Scenario 2: Low light intensity, high wind speed, low commercial load, large market price fluctuations. Scenario 3: Medium light intensity, medium wind speed, mainly residential load, stable market electricity prices.

[0042] The output scenario segmentation results include cluster centers of typical scenarios, sample distribution (sample numbers included in each scenario), and scenario feature descriptions (main resource characteristics, environmental conditions, and policy background of each scenario), providing a foundation for subsequent source-load-storage aggregation characteristic analysis.

[0043] In step S3, for each typical scenario, the scenario requirements are analyzed. Typical scenarios include, but are not limited to, power supply requirements, flexibility requirements, and economic requirements. Define scenario adaptability evaluation metrics to quantify the degree of matching between resource capabilities and scenario requirements; scenario adaptability evaluation metrics include, but are not limited to, resource availability, response speed, cost-effectiveness, etc. Based on the characteristics of different scenarios, the demand characteristics include: multi-dimensional demand and dynamic adaptability demand. Multi-dimensional demand includes: power supply demand: ensuring that resources can meet the basic power needs of the load in a specific scenario, including load size, duration, and reliability requirements; flexibility demand: resources need to have rapid response and adjustment capabilities to cope with load fluctuations and uncertainties, such as the charging and discharging speed of energy storage devices and the interruptibility of flexible loads; economic demand: the use of resources needs to be cost-effective, including investment costs, operating costs, and return levels, while considering the impact of subsidy policies and market electricity price fluctuations on economics; reliability demand: the availability and stability of resources, including equipment failure rate, reserve capacity, and maintenance cycle; environmental friendliness demand: the use of resources needs to comply with environmental protection requirements, such as carbon emission limits and the proportion of renewable energy use; market responsiveness demand: resources need to be able to adapt to changes in market signals, such as market price fluctuations and adjustments to trading rules. Dynamic adaptability includes: scenario demands change over time, market conditions, and user behavior, therefore a mechanism that can dynamically adjust weights is needed to adapt to real-time changes.

[0044] Based on machine learning algorithms such as the random forest algorithm, corresponding weights are assigned to scene adaptability evaluation indicators. Following the above example, assume the following typical scenes and resource data: Typical scenarios include: Scenario 1: High light intensity, low wind speed, peak industrial load, and preferential subsidy policies. Scenario 2: Low light intensity, high wind speed, low commercial load, and large market price fluctuations. Scenario 3: Moderate light intensity, moderate wind speed, mainly residential load, and stable market electricity prices.

[0045] Resource data includes:

[0046] Collect feature data related to scenario requirements, including resource characteristics (e.g., photovoltaic output volatility, energy storage regulation capacity), environmental conditions (e.g., solar irradiance, wind speed), and policy background (e.g., subsidy policies, market electricity price fluctuations); define the main requirements for each scenario, such as high reliability, rapid response, cost-effectiveness, flexibility, and environmental friendliness, and convert them into quantifiable labels; convert qualitative features, such as solar irradiance, wind speed, and load type, into quantitative data, and integrate all feature data into a feature matrix; convert requirement labels into numerical labels, and integrate all label data into a label vector; Initialize the random forest model and set appropriate parameters, such as the number of trees and maximum depth. Train the random forest model using feature matrices and label vectors, and evaluate the importance score of each feature using the random forest model as the weight for scene adaptability evaluation. Normalize the feature importance scores to the [0,1] interval to ensure the comparability of weights. Output the normalized weights for subsequent scene adaptability evaluation. Combine the weights and feature values ​​to generate a detailed description of each scene.

[0047] Furthermore, the characteristic indicator data includes: characteristic data, environmental data, and policy data for each resource.

[0048] In one specific embodiment, the characteristic data of each resource includes: the characteristics of resources such as distributed photovoltaic, energy storage, and flexible loads, including but not limited to the fluctuation of photovoltaic output, the regulation capability of energy storage, and the interruptibility of load; Environmental data includes, but is not limited to, meteorological conditions (such as temperature, light intensity, wind speed, etc.) and seasonal variations; Policy data, including but not limited to electricity market rules, subsidy policies, and entry thresholds.

[0049] Furthermore, the preset scenario adaptability evaluation indicators include: resource availability indicators, response speed indicators, and cost-effectiveness indicators.

[0050] Furthermore, such as Figure 4 As shown, in step S4, the fit matrix of multiple scenes with multiple resources and / or resource combinations is calculated, including: S401: Based on the first scoring group of each resource and / or resource combination and the weights of multiple preset scenario adaptability evaluation indicators in each scenario, calculate the comprehensive adaptability score of each resource and / or resource combination in each scenario; and combine them to obtain the first intermediate matrix; S402: Normalize the first intermediate matrix to obtain a normalized first intermediate matrix, which serves as the adaptation matrix for multiple scenarios and multiple resources and / or resource combinations.

[0051] In step S401, the matrix value of the first intermediate matrix is ​​the comprehensive adaptability score, with rows representing resources and / or resource combinations and columns representing scenarios.

[0052] It should be noted that for each typical scenario, the compatibility between computational resource capabilities and scenario requirements is calculated to form a compatibility matrix. This matrix is ​​then used to accurately assess the resource aggregation characteristics under different scenarios, providing a basis for multi-energy collaborative aggregation. Each element in the compatibility matrix represents the degree of matching between a specific resource or resource combination and scenario requirements. By defining scenario compatibility evaluation indicators and quantifying the degree of matching between resource capabilities and scenario requirements, the applicability of resources in different scenarios can be accurately assessed, providing a scientific basis for resource aggregation. A dynamic weight adjustment mechanism is introduced, adjusting weights in real time based on market electricity price fluctuations, resource status changes, and user demand feedback. This allows the evaluation model to dynamically adapt to changes in scenario requirements, enhancing the model's flexibility and adaptability.

[0053] In S4, the fit between computing resource capabilities and scenario requirements is formed into a fit matrix, which includes the following steps: For each typical scenario, a dynamic weight adjustment mechanism is introduced, which dynamically adjusts the weights based on market electricity price fluctuations, changes in resource status, and user demand feedback to adapt to the dynamic changes in scenario requirements. When market electricity prices fluctuate significantly, the weight of market adaptability indicators is increased. For example, in some scenarios, reliability may be more important, while in other scenarios, cost-effectiveness or flexibility may be more critical. For each resource or combination of resources, evaluate its performance on various scenario suitability metrics, including but not limited to: Regarding resource availability indicators: When the actual output or response capacity of a resource reaches more than 90% of its rated capacity, it is considered to be in a callable operating state; when it is under maintenance shutdown, experiencing unplanned shutdown due to fault, or when the output is less than 30% of the rated capacity for more than 30 consecutive minutes, it is considered to be in a non-callable operating state.

[0054] Regarding the response speed indicator: Extract all adjustment command records from the resource monitoring historical database; filter out commands with adjustment ranges greater than 20% of the rated capacity, calculate the average response time of all valid commands; and plan a classification standard based on this: Level A (fast response): Δt≤30 seconds; Level B (standard response): 30 seconds<Δt≤5 minutes; Level C (slow response): Δt>5 minutes.

[0055] Regarding cost-benefit indicators: assess the economics of resources throughout their entire life cycle, including investment costs, operating costs, and revenue potential, and calculate the overall cost-benefit ratio accordingly; For each resource or resource combination, in each typical scenario, a weighted fit score is calculated based on its performance on each evaluation indicator and its corresponding weight. The calculation process is as follows: multiply the performance value of each indicator by the weight of its scenario fit evaluation indicator (as explained in step S3), and then sum the weighted values ​​of all indicators to obtain the comprehensive fit score of the resource in that scenario. This weighted fit score reflects the comprehensive ability of the resource to meet the scenario requirements in a specific scenario. The calculated fit score is then normalized to the [0,1] range. The normalization process includes subtracting the minimum value from each score and then dividing by the difference between the maximum and minimum scores. Resources with lower scores are closer to 0 after normalization, while resources with higher scores are closer to 1, thus intuitively reflecting the relative fit of resources in different scenarios. Through the above process, the weighted fit score of each resource or resource combination in each typical scenario can be obtained, providing a quantitative basis for subsequent resource aggregation characteristic analysis and optimization decisions. A nonlinear adjustment factor is introduced to adjust the resource suitability score based on the synergistic effect of resources, so as to reflect the comprehensive performance of resources in a specific scenario. All typical scenarios are represented as rows of a matrix, and all resources or combinations of resources are represented as columns of the matrix, forming a scenario-resource matrix; Enter the corresponding fit score into each cell of the matrix, and normalize the fit matrix to standardize the fit score to the [0,1] interval to facilitate comparison between different scenarios and resources; Introducing resource synergy analysis to evaluate the synergistic adaptability of resource combinations in different scenarios includes the following steps: Based on complementarity, synergy gain, and redundancy indicators, resource synergy effect indicators are defined to measure the synergistic performance of resource combinations in specific scenarios. These indicators include measuring the degree of complementarity of resources in time, space, or function; for example, the complementarity of photovoltaics and energy storage is quantified by the output volatility of photovoltaics and the regulation capability of energy storage. Synergy gains of resource combinations in specific scenarios are also measured; for example, the synergistic gains of energy storage and flexible loads can be quantified by the response speed of energy storage and the interruptibility of flexible loads. Finally, redundancy in resource combinations is measured to avoid over-allocation of resources. The synergistic fit score of the resource combination is calculated. By comprehensively considering the fit scores of each resource in the resource combination and the synergistic effect index, the overall fit score of the resource combination is obtained. Update the fit matrix by filling in the synergistic fit scores of the resource combinations into the matrix, forming a fit matrix with resource combinations; the filling steps include: Expand the matrix dimensions in the original scenario-resource matrix by adding new columns to represent resource combinations; for example, if resource 1 and resource 2 form a combination, add a new column for resource 1+2. Enter the corresponding collaboration adaptation score in each cell of the matrix; for example, if the collaboration adaptation score of resource 1+2 in scenario 1 is 0.85, then enter that value in the matrix. The updated fitness matrix is ​​normalized to ensure that all scores are within the [0,1] interval for easy comparison.

[0056] Assume the fitness matrix is ​​as follows:

[0057] Updated matrix (incorporating the combination of resources 1 and 2):

[0058] Perform multi-dimensional analysis on the adaptability matrix to extract the aggregation characteristics of resources in different scenarios; generate a resource aggregation characteristic analysis report based on the adaptability matrix; the multi-dimensional analysis of the adaptability matrix includes the following steps: The clustering analysis method based on step S203 divides resources or resource combinations into different clusters and identifies the aggregation patterns of resources under different demand types. For example: Cluster 1: mainly meets the demand for power supply, and the resource combination has high availability and reliability; Cluster 2: mainly meets the demand for flexibility, and the resource combination has rapid response and high regulation capability; Cluster 3: mainly meets the demand for economic efficiency, and the resource combination has high cost-effectiveness and market adaptability.

[0059] The adaptability matrix is ​​analyzed based on resource type (e.g., distributed photovoltaic, energy storage, flexible load, etc.) to assess the adaptability differences of different resource types in various scenarios; a time dimension is introduced to analyze the adaptability changes of resources in different time periods and identify the dynamic aggregation characteristics of resources. A real-time data calibration mechanism is introduced, which collects resource operation data and scenario demand data in real time, including but not limited to real-time output, response speed, cost-effectiveness indicators of resources, as well as real-time power demand and market electricity price information of scenarios. Based on real-time data, the fit between resource capabilities and scenario requirements is recalculated, and relevant elements in the fit matrix are updated to ensure the timeliness and accuracy of the fit assessment. The adaptability matrix is ​​continuously optimized, and the resource aggregation strategy is dynamically adjusted based on real-time data to ensure that the resource aggregation solution can adapt to the changing needs of the scenario.

[0060] Furthermore, based on the fit matrix, a resource aggregation characteristic analysis report is generated, including the following steps: Based on the resource adaptability ranking in each typical scenario, the resource or resource combination is displayed in order of its adaptability score in each scenario. Resource aggregation characteristic analysis includes the analysis of the advantages and disadvantages of resources in different scenarios, as well as the analysis of the synergistic effect of resource combinations; multi-energy collaborative aggregation strategy suggestions are proposed based on the results of the adaptability matrix analysis, suggesting strategies and implementation plans for optimizing resource aggregation in different scenarios.

[0061] Output a scenario adaptability evaluation report. The report includes resource adaptability scores, advantageous resource analysis, potential optimization directions, and targeted aggregation strategy recommendations for each typical scenario. This provides detailed evidence for subsequent source-load-storage aggregation characteristic analysis and optimization decisions. For example: Scenario 1: Resource 1 has the highest adaptability; it is recommended to prioritize resource 1. Scenario 2: Resource 2 has the highest adaptability; it is recommended to configure resource 2 to adapt to market price fluctuations. Scenario 3: Resource 3 has the highest adaptability; it is recommended to configure resource 3 to meet residential load demand.

[0062] Furthermore, in step S401, the formula for calculating the overall fit score for each scenario is as follows:

[0063] This indicates the total number of the first scoring group representing resources and / or resource combinations. Indicates the sequence number of the first scoring group of resources and / or resource combinations. This represents the weight of the a-th preset scenario adaptability evaluation index. This represents the a-th value in the first rating group of resources and / or resource combinations.

[0064] Furthermore, such as Figure 5 As shown, in step S5, based on the fit matrix of the multiple scenarios and multiple resources and / or resource combinations, the optimal solution for each scenario is obtained, including: S501: Based on the fit matrix of the multiple scenarios and multiple resources and / or resource combinations, select the resources and / or resource combinations in each scenario whose fit is greater than the second preset threshold, and use them as the first resource group for each scenario. S502: A multi-objective optimization algorithm is used to optimize and generate the solution for the first resource group of each scenario, so as to obtain the optimal solution for each scenario.

[0065] Multi-energy synergistic aggregation analysis based on the fitness matrix is ​​used to determine the optimal source-load-storage aggregation scheme in different scenarios. This includes: selecting resources or resource combinations with fitness scores higher than a set threshold for each typical scenario based on the fitness matrix, forming a candidate aggregation scheme set for that scenario; constructing a multi-objective optimization model based on the candidate aggregation scheme set, with the core objectives of minimizing comprehensive operating costs and maximizing renewable energy absorption rate; incorporating resource complementarity (e.g., quantification of complementarity indicators), economic efficiency (e.g., investment payback period, internal rate of return), and reliability (e.g., availability rate, reserve capacity) as key constraints into the model; and solving the model using a multi-objective optimization algorithm (e.g., NSGA-II) to obtain a set of optimal frontier solutions, i.e., a set of optimal aggregation schemes that do not dominate each other in terms of cost, absorption rate, and reliability. Each scheme includes: resource composition and configuration ratio (e.g., specifying which photovoltaic, energy storage, and load resources and their capacity configurations are included), and multi-timescale scheduling strategies (e.g., providing day-ahead, day-ahead, and day-ahead scheduling strategies). Operational strategies at different time scales, such as internal and real-time: energy storage charging and discharging plans or load response periods), and expected market participation strategies (e.g., describing the expected participation methods and revenue composition of the scheme in the electricity spot market, ancillary services, green electricity trading, etc.); A comprehensive evaluation is performed on each scheme in the optimal scheme set, and its comprehensive performance score is calculated based on the analytic hierarchy process (AHP); and combined with actual needs (e.g., prioritizing economic efficiency or environmental friendliness), multiple recommended aggregated schemes are output as the final decision result, and structured reports for different application scenarios are generated; The reports include: the optimal source-load-storage aggregation scheme for each typical scenario, including resource combination, configuration ratio, and scheduling strategy; resource complementarity, economic efficiency, and reliability analysis results, demonstrating the comprehensive performance of resources in different scenarios; optimization results of comprehensive operating costs and renewable energy absorption rate, quantitatively evaluating the economic efficiency and environmental friendliness of the resource aggregation scheme; implementation suggestions and risk assessments for different scenarios, providing decision support for power system planning, operation, and market transactions; S5 also includes the following steps: Define resource complementarity indicators to quantify the degree of complementarity of resources in time, space, and function; Define resource economic indicators, including but not limited to investment costs, operating costs, and return levels; Define resource reliability metrics, including but not limited to resource availability, failure rate, and backup capacity; A multi-objective optimization model is constructed with the objectives of minimizing overall operating costs and maximizing the renewable energy absorption rate, and the complementarity, economy and reliability of resources are comprehensively analyzed. Furthermore, such as Figure 6 As shown, in step S502, the multi-objective optimization algorithm includes: S50201: Based on the first resource group of each scenario, generate multiple resource allocation genes, scheduling strategy genes, and economic genes corresponding to each scenario; S50202: Generate multiple chromosomes corresponding to each scenario based on the multiple resource allocation genes, scheduling strategy genes, and economic genes corresponding to each scenario; S50203: Using a genetic algorithm, optimize multiple chromosomes for each scenario to obtain the optimal chromosome for each scenario; S50204: Based on the optimal chromosome for each scenario, obtain the optimal solution for each scenario.

[0066] A genetic algorithm is used to solve a multi-objective optimization model. Chromosomes in the genetic algorithm are defined as resource allocation and scheduling strategies. A chromosome consists of multiple genes, each corresponding to a decision variable. The organizational relationships include: resource allocation genes represent the allocation ratio or quantity of a certain resource (such as distributed photovoltaic, energy storage, or flexible loads), for example, a gene value is an integer representing the quantity of resources; scheduling strategy genes represent the scheduling strategy of resources, such as charging / discharging time, power output, etc., for example, a gene value is a time series representing the power output or charging / discharging state of resources at different time periods; economic genes represent economic parameters related to resources, such as investment cost, operating cost, etc., for example, a gene value is a floating-point number representing the unit cost of resources; and reliability genes represent reliability parameters related to resources, such as availability, failure rate, etc., for example, a gene value is a floating-point number between [0,1] representing the availability probability of resources.

[0067] A set of chromosomes is randomly generated, each representing a potential resource allocation and scheduling strategy. The value of each gene is randomly generated based on the range of its corresponding decision variable. The fitness of each chromosome is calculated according to defined resource complementarity, economy, and reliability indicators. Based on the fitness, the best chromosomes are selected to enter the next generation, using methods such as roulette wheel selection and tournament selection. Two chromosomes are randomly selected, and some genes are exchanged to generate a new chromosome. One or more genes in a chromosome are randomly selected, and their values ​​are changed. The mutation operation can be random mutation or mutation based on a certain probability distribution. A new population is generated through selection, crossover, and mutation operations. The fitness evaluation, selection, crossover, and mutation operations are repeated until the termination condition is met (such as reaching the maximum number of iterations or fitness convergence). The chromosome with the highest fitness is selected from the final population as the optimal resource allocation and scheduling strategy.

[0068] A dynamic scheduling strategy optimization mechanism is introduced. By monitoring the resource operation status and changes in scenario demand in real time, including real-time resource output, market electricity price fluctuations, and user demand feedback, the adaptability of resources is reassessed based on real-time data, and the adaptability matrix is ​​dynamically updated. Based on the updated adaptability matrix, resource allocation and scheduling strategies are dynamically adjusted to ensure that the resource aggregation scheme can adapt to real-time changing scenario requirements. Following the above embodiment, assume there are three types of resources: photovoltaic, energy storage, and flexible loads. Each resource has four genes: allocation ratio (0.2, 0.3, 0.5), scheduling strategy (time series), economics (cost), and reliability (availability); the chromosome structure is represented as follows: [0.2,[10,20,30],0.05,0.95,0.3,[40,50,60],0.06,0.90,0.5,[70,80,90],0.07,0.85]; Among them, photovoltaic: allocation ratio 0.2, scheduling strategy [10,20,30], economic efficiency 0.05, reliability 0.95. Energy storage: allocation ratio 0.3, scheduling strategy [40,50,60], economic efficiency 0.06, reliability 0.90. Flexible load: allocation ratio 0.5, scheduling strategy [70,80,90], economic efficiency 0.07, reliability 0.85. Through genetic algorithm, the gene values ​​are continuously optimized to find the optimal resource allocation and scheduling strategy.

[0069] S5 also includes the following steps: A model-data two-layer adaptive linkage mechanism is introduced to analyze the current characteristics of the data in real time and evaluate the performance status of the deployed model; Based on the analysis results, dynamically adjust the combination of data features input to the model and the internal configuration of the model itself, including: updating model parameters, adjusting the structural parameters of the model, and adjusting the combined weights of multiple model outputs; After adjustments, verify the model's effectiveness; once verification is successful, enable the model. The adjusted model is used to process the data, output the results, and record the adjustment process, so as to automatically adapt to different models, different data, and different input data types.

[0070] It should be noted that by selecting resources or resource combinations with high suitability scores and analyzing the complementarity between resources, the optimal source-load-storage aggregation scheme can be determined under different scenarios, thereby improving resource utilization efficiency and system performance. Introducing a dynamic scheduling strategy optimization mechanism allows for dynamic adjustment of resource configuration and scheduling strategies based on real-time data, ensuring that the resource aggregation scheme can adapt to real-time changing scenario requirements. By optimizing resource configuration and scheduling strategies, the scientific rigor and adaptability of source-load-storage aggregation can be effectively improved, providing strong support for the efficient operation of the power system, renewable energy consumption, and market transactions.

[0071] like Figure 7 As shown, a source-load-storage aggregation solution generation system for multiple scenarios includes: Data acquisition module: acquires multiple feature index data; multiple preset scene adaptability evaluation indicators; acquires the scores of multiple resources and / or resource combinations on each preset scene adaptability evaluation indicator, as the first score group for each resource and / or resource combination; Scene calculation module: Based on the multiple indicator data, calculate multiple scenes and the scene features corresponding to each scene; Weight calculation module: Based on the multiple scenarios and corresponding scenario features, the random forest algorithm is used to calculate the weights of the multiple preset scenario fitness evaluation indicators to obtain the weights of the multiple preset scenario fitness evaluation indicators under each scenario; Adaptability matrix calculation module: Based on the first scoring group of each resource and / or resource combination and the weights of multiple preset scenario adaptability evaluation indicators in each scenario, the adaptability matrix of multiple scenarios with multiple resources and / or resource combinations is calculated. Solution generation module: Based on the adaptability matrix of the multiple scenarios and multiple resources and / or resource combinations, obtain the optimal solution for each scenario.

[0072] The same or similar labels correspond to the same or similar parts; The terms used to describe positional relationships in the accompanying drawings are for illustrative purposes only and should not be construed as limiting this patent. Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the implementation of the present invention. Those skilled in the art can make other variations or modifications based on the above description. It is neither necessary nor possible to exhaustively describe all embodiments here. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the claims of the present invention.

Claims

1. A method for generating source-load-storage aggregation schemes for multiple scenarios, characterized in that, include: S1: Obtain data for multiple feature indicators; Multiple preset scene adaptability evaluation metrics; obtain the scores of multiple resources and / or resource combinations on each preset scene adaptability evaluation metric, and use them as the first score group for each resource and / or resource combination; S2: Based on the multiple indicator data, calculate multiple scenarios and the scenario features corresponding to each scenario; S3: Based on the multiple scenarios and corresponding scenario features, use the random forest algorithm to calculate the weights of the multiple preset scenario fitness evaluation indicators to obtain the weights of the multiple preset scenario fitness evaluation indicators under each scenario. S4: Based on the first scoring group of each resource and / or resource combination and the weights of multiple preset scenario adaptability evaluation indicators in each scenario, calculate the adaptability matrix of multiple scenarios with multiple resources and / or resource combinations. S5: Based on the fit matrix of the multiple scenarios and multiple resources and / or resource combinations, obtain the optimal solution for each scenario.

2. The method for generating source-load-storage aggregation schemes for multiple scenarios according to claim 1, characterized in that, In step S2, based on the multiple indicator data, multiple scenarios and scene features corresponding to each scenario are calculated, including: S201: Standardize each of the aforementioned feature index data to obtain multiple standardized feature index data; S202: Perform principal component dimensionality reduction on each of the standardized feature index data to obtain multiple dimensionality-reduced feature index data; S203: Perform cluster analysis on the multiple dimensionality-reduced feature index data to obtain multiple scenarios and the scene features corresponding to each scenario.

3. The method for generating source-load-storage aggregation schemes for multiple scenarios according to claim 2, characterized in that, In step S203, cluster analysis includes: S20301: The elbow method is used to process multiple dimensionality-reduced feature index data to obtain the number of clusters; S20302: Based on the number of clusters, the K-Means clustering algorithm is used to cluster the multiple dimensionality-reduced feature index data to obtain multiple scenarios and multiple feature index data corresponding to each scenario; S20303: Average the multiple feature index data corresponding to each scenario to form cluster center feature index data corresponding to multiple scenarios; S20304: Compare the cluster center feature index data corresponding to each scenario with the first preset threshold vector to obtain the scenario features corresponding to each scenario.

4. The method for generating a source-load-storage aggregation scheme for multiple scenarios according to any one of claims 1 to 3, characterized in that, The characteristic indicator data includes: characteristic data, environmental data, and policy data for each resource.

5. The method for generating source-load-storage aggregation schemes for multiple scenarios according to claim 1, characterized in that, The preset scenario adaptability evaluation indicators include: resource availability indicators, response speed indicators, and cost-effectiveness indicators.

6. The method for generating source-load-storage aggregation schemes for multiple scenarios according to claim 1, characterized in that, In step S4, the fit matrix between multiple scenes and multiple resources and / or resource combinations is calculated, including: S401: Based on the first scoring group of each resource and / or resource combination and the weights of multiple preset scenario adaptability evaluation indicators in each scenario, calculate the comprehensive adaptability score of each resource and / or resource combination in each scenario; and combine them to obtain the first intermediate matrix; S402: Normalize the first intermediate matrix to obtain a normalized first intermediate matrix, which serves as the adaptation matrix for multiple scenarios and multiple resources and / or resource combinations.

7. The method for generating source-load-storage aggregation schemes for multiple scenarios according to claim 6, characterized in that, In step S401, the formula for calculating the overall fit score for each scenario is as follows: This indicates the total number of the first scoring group representing resources and / or resource combinations. Indicates the sequence number of the first scoring group of resources and / or resource combinations. This represents the weight of the a-th preset scenario adaptability evaluation index. This represents the a-th value in the first rating group of resources and / or resource combinations.

8. The method for generating source-load-storage aggregation schemes for multiple scenarios according to claim 1, characterized in that, In step S5, based on the fit matrix of the multiple scenarios and multiple resources and / or resource combinations, the optimal solution for each scenario is obtained, including: S501: Based on the fit matrix of the multiple scenarios and multiple resources and / or resource combinations, select the resources and / or resource combinations in each scenario whose fit is greater than the second preset threshold, and use them as the first resource group for each scenario. S502: A multi-objective optimization algorithm is used to optimize and generate the solution for the first resource group of each scenario, so as to obtain the optimal solution for each scenario.

9. The method for generating source-load-storage aggregation schemes for multiple scenarios according to claim 8, characterized in that, In step S502, the multi-objective optimization algorithm includes: S50201: Based on the first resource group of each scenario, generate multiple resource allocation genes, scheduling strategy genes, and economic genes corresponding to each scenario; S50202: Generate multiple chromosomes corresponding to each scenario based on the multiple resource allocation genes, scheduling strategy genes, and economic genes corresponding to each scenario; S50203: Using a genetic algorithm, optimize multiple chromosomes for each scenario to obtain the optimal chromosome for each scenario; S50204: Based on the optimal chromosome for each scenario, obtain the optimal solution for each scenario.

10. A source-load-storage aggregation scheme generation system for multiple scenarios, applied to the generation method described in any one of claims 1 to 9, characterized in that, include: Data acquisition module: Acquires data from multiple feature indicators; Multiple preset scene adaptability evaluation metrics; obtain the scores of multiple resources and / or resource combinations on each preset scene adaptability evaluation metric, and use them as the first score group for each resource and / or resource combination; Scene calculation module: Based on the multiple indicator data, calculate multiple scenes and the scene features corresponding to each scene; Weight calculation module: Based on the multiple scenarios and corresponding scenario features, the random forest algorithm is used to calculate the weights of the multiple preset scenario fitness evaluation indicators to obtain the weights of the multiple preset scenario fitness evaluation indicators under each scenario; Adaptability matrix calculation module: Based on the first scoring group of each resource and / or resource combination and the weights of multiple preset scenario adaptability evaluation indicators in each scenario, the adaptability matrix of multiple scenarios with multiple resources and / or resource combinations is calculated. Solution generation module: Based on the adaptability matrix of the multiple scenarios and multiple resources and / or resource combinations, obtain the optimal solution for each scenario.

Images