Multi-objective optimization method and system for business expansion scheme of power supply load access drive

Abstract

The invention discloses a multi-objective optimization method and system for a business expansion scheme of power supply load access driving, and relates to the related technical field of power distribution network optimization, and the method comprises the steps: constructing a large user load feature library through a historical power utilization data set; performing load mode rule mining to obtain a large user load mode identification rule base, and determining a new user load mode according to new user installation parameter traversal matching; performing power grid business expansion scheme analysis in combination with the target power grid characteristic data, and generating a baseline business expansion scheme parameter threshold; and constructing multiple optimization objectives of the business expansion scheme, carrying out global solution optimization, and determining a target power grid business expansion scheme. The technical problems that in the prior art, a business expansion scheme lacks a data-driven decision, user load characteristics are difficult to accurately match and multi-conflict targets cannot be collaboratively optimized are solved, and the technical effects that efficient configuration of power grid resources is achieved, scientificity and accuracy of the business expansion scheme are improved, and safety and stability of power grid operation are improved are achieved.

Description

Technical Field

[0001] This invention relates to the technical field of power distribution network optimization, specifically to a multi-objective optimization method and system for power load access-driven business expansion schemes. Background Technology

[0002] As the power grid load structure becomes increasingly complex, the electricity consumption behavior of large industrial and commercial users has a significant impact on grid operation. These users have complex load characteristics and diverse electricity demands. Their grid connection process not only affects the reliability of their own electricity supply but also directly impacts the stability, energy efficiency, and resource utilization of the power grid. Traditional business expansion scheme design lacks in-depth analysis of user load characteristics and multi-objective collaborative optimization, easily leading to problems such as unreasonable grid connection capacity, low energy efficiency, and increased operating costs. In the current context of smart grids and the energy internet, massive amounts of historical electricity consumption data provide a data foundation for analyzing user electricity consumption behavior and mining load characteristic patterns. This allows for refined modeling of large user loads, identification of their typical electricity consumption patterns, and the development of more scientific and personalized business expansion schemes. However, how to extract effective load characteristics from massive data, construct accurate load pattern identification rules, and achieve multi-objective optimization of business expansion schemes remains a challenge in current power planning. Currently, there is a lack of end-to-end optimization from load characteristic analysis to business expansion scheme generation, making it impossible to determine a business expansion scheme that balances safety, economy, and energy efficiency.

[0003] Therefore, current technologies suffer from technical problems such as a lack of data-driven decision-making in business expansion solutions, difficulty in accurately matching user load characteristics, and inability to collaboratively optimize multiple conflicting objectives. Summary of the Invention

[0004] This application provides a multi-objective optimization method and system for power load access-driven business expansion schemes, which solves the technical problems in the prior art of business expansion schemes lacking data-driven decision-making, being difficult to accurately match user load characteristics, and being unable to coordinately optimize multiple conflicting objectives. It achieves the technical effects of realizing efficient allocation of power grid resources, improving the scientificity and accuracy of business expansion schemes, and enhancing the safe and stable performance of power grid operation.

[0005] This application provides a multi-objective optimization method for power load access-driven business expansion schemes. The method includes: performing feature clustering analysis on historical electricity consumption datasets to construct a large user load feature library; performing load pattern rule mining on the large user load feature library to obtain a large user load pattern identification rule library; performing traversal matching on the large user load pattern identification rule library according to the new user application parameters to determine the new user load pattern; combining target power grid characteristic data to perform power grid business expansion scheme analysis on the new user load pattern to generate baseline business expansion scheme parameter thresholds; constructing multiple optimization objectives for business expansion schemes; and performing global optimization on the baseline business expansion scheme parameter thresholds according to the multiple optimization objectives to determine the target power grid business expansion scheme.

[0006] In a possible implementation, the multi-objective optimization method for the power load access-driven business expansion scheme further performs the following processing: cleaning the historical electricity consumption dataset to obtain an available electricity consumption dataset; extracting and normalizing the load features of the available electricity consumption dataset according to data application standards to obtain a standard user load feature set; performing cluster analysis on the standard user load feature set to obtain a user load feature cluster set; and matching and mapping the available electricity consumption dataset to the user load feature cluster set for labeling and integration to construct the large user load feature library.

[0007] In a possible implementation, the multi-objective optimization method for the power load access-driven business expansion scheme further performs the following processing: Based on the dimensional characteristics of the standard user load feature set, a minimum number of samples is preset; the nearest neighbor distance between each data point in the standard user load feature set and the minimum number of samples is calculated, and a k-distance sorting graph is generated based on the nearest neighbor distance; the slope inflection point of the k-distance sorting graph is used as the target neighborhood radius; cluster analysis and verification optimization are performed on the standard user load feature set based on the minimum number of samples and the target neighborhood radius to obtain the user load feature cluster set.

[0008] In a possible implementation, the multi-objective optimization method for the power load access-driven business expansion scheme further performs the following processing: extracting large user electricity load datasets from the large user load feature library; performing K-means clustering and load pattern identification on the large user electricity load datasets to obtain large user load pattern data clusters; using decision trees to sequentially mine load pattern rules on the large user load pattern data clusters to obtain large user load pattern cluster rule sets; and performing confidence evaluation, filtering, and rule integration on the large user load pattern cluster rule sets to obtain the large user load pattern identification rule library.

[0009] In a possible implementation, the multi-objective optimization method for the power load access driven business expansion scheme further performs the following processing: converting the new user application parameters according to the large user load pattern recognition rule library to obtain standardized application parameters; performing fuzzy matching on the large user load pattern recognition rule library based on the standardized application parameters to obtain a matching load pattern rule set; sorting the matching load pattern rule set in descending order according to confidence level to obtain a matching load pattern sequence rule, and determining the new user load pattern according to the matching load pattern sequence rule.

[0010] In a possible implementation, the multi-objective optimization method for the power load access-driven business expansion scheme further performs the following processing: acquiring power grid structure data, operational constraint data, historical power grid operation data, and expansion planning data based on the target power grid characteristic data; performing characteristic modeling based on the power grid structure data, operational constraint data, historical power grid operation data, and expansion planning data to generate a target power grid characteristic model; and performing power grid business expansion scheme analysis on the new user load pattern according to the target power grid characteristic model to generate baseline business expansion scheme parameter thresholds.

[0011] In a possible implementation, the multi-objective optimization method for the power load access-driven business expansion scheme further performs the following processing: performing importance assessment and weighted normalization on the multiple optimization objectives of the business expansion scheme to construct the business expansion scheme optimization objective function; initializing the business expansion scheme particle space according to the baseline business expansion scheme parameter threshold; and performing global solution optimization on the business expansion scheme particle space based on the business expansion scheme optimization objective function to determine the target power grid business expansion scheme.

[0012] In a possible implementation, the multi-objective optimization method for the power load access driven business expansion scheme further performs the following processing: evaluating the fitness of the business expansion scheme particle space based on the business expansion scheme optimization objective function to obtain a business expansion scheme particle fitness set; updating the particle velocity and position of the business expansion scheme particle space according to the business expansion scheme particle fitness set until a preset termination condition is met, and finding the target power grid business expansion scheme with the largest particle fitness.

[0013] In a possible implementation, the multi-objective optimization method for the power load access driven business expansion scheme further performs the following processing: performing non-dominated sorting and congestion calculation based on the particle fitness set of the business expansion scheme to generate a Pareto optimal parameter solution set; selecting a guiding particle from the Pareto optimal parameter solution set, and introducing the guiding particle to iteratively update the particle velocity and position of the business expansion scheme particle space.

[0014] This application also provides a multi-objective optimization system for power load access-driven business expansion schemes. The system includes: an electricity consumption characteristic analysis module for performing feature clustering analysis on historical electricity consumption datasets to construct a large user load characteristic library; a load pattern determination module for performing load pattern rule mining on the large user load characteristic library to obtain a large user load pattern identification rule library, and performing traversal matching on the large user load pattern identification rule library according to the new user application parameters to determine the new user load pattern; a business expansion scheme parsing module for performing power grid business expansion scheme parsing on the new user load pattern in conjunction with target power grid characteristic data to generate baseline business expansion scheme parameter thresholds; and a power grid business expansion scheme determination module for constructing multiple optimization objectives for business expansion schemes, and performing global optimization on the baseline business expansion scheme parameter thresholds according to the multiple optimization objectives to determine the target power grid business expansion scheme.

[0015] This application proposes a multi-objective optimization method and system for power load access-driven business expansion schemes. This method constructs a large user load feature library using historical electricity consumption datasets; mines load pattern rules to obtain a large user load pattern recognition rule library; determines the new user load pattern by traversing and matching new user application parameters; analyzes the power grid business expansion scheme using target power grid characteristic data to generate baseline business expansion scheme parameter thresholds; constructs multiple optimization objectives for the business expansion scheme and performs global optimization to determine the target power grid business expansion scheme. This addresses the technical problems in existing technologies, such as the lack of data-driven decision-making in business expansion schemes, difficulty in accurately matching user load characteristics, and inability to collaboratively optimize multiple conflicting objectives. It achieves the technical effects of efficient allocation of power grid resources, improved scientific rigor and accuracy of business expansion schemes, and enhanced power grid operational safety and stability. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of this disclosure, the accompanying drawings of the embodiments of this disclosure will be briefly described below. Flowcharts are used in this application to illustrate the operations performed by the system according to the embodiments of this application. It should be understood that the preceding or following operations are not necessarily performed precisely in sequence. Instead, various steps can be processed in reverse order or simultaneously as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from these processes.

[0017] Figure 1 This is a schematic diagram of the multi-objective optimization method for the power load access driven business expansion scheme provided in the embodiments of this application.

[0018] Figure 2 This is a schematic diagram of the multi-objective optimization system structure for the power load access driven business expansion scheme provided in the embodiments of this application.

[0019] Explanation of reference numerals in the attached diagrams: Electricity consumption characteristic analysis module 10, business expansion scheme analysis module 30, load mode determination module 20, and power grid business expansion scheme determination module 40. Detailed Implementation

[0020] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0021] This application provides a multi-objective optimization method for power load access driven business expansion schemes, such as... Figure 1 As shown, the method includes: Step S100: Perform feature clustering analysis on the historical electricity consumption dataset to construct a large user load feature database.

[0022] Step S100 further includes step S110, performing data cleaning processing on the historical electricity consumption dataset to obtain an available electricity consumption dataset; step S120, performing load feature extraction and normalization processing on the available electricity consumption dataset according to data application standards to obtain a standard user load feature set; step S130, performing cluster analysis on the standard user load feature set to obtain a user load feature cluster set; and step S140, matching and mapping the available electricity consumption dataset to the user load feature cluster set for labeling and integration to construct the large user load feature library.

[0023] Preferably, historical electricity consumption data typically contains erroneous, missing, abnormal, or redundant information. Data cleaning of historical electricity consumption datasets may include handling missing values, i.e., filling in or deleting data lost due to equipment failure or communication interruption; correcting outliers, i.e., identifying and correcting data that clearly does not conform to physical laws due to metering errors or transient interference, such as negative power or power values ​​far exceeding transformer capacity; data alignment, unifying data from different sources or with different time granularities to the same timestamp; and removing duplicate data, i.e., deleting identical data entries generated by duplicate records, thereby outputting a usable electricity consumption dataset with better data quality. According to data application standards, load features are extracted from available electricity datasets. This includes calculating various statistical features from time-series electricity data, such as daily / monthly / yearly maximum load, minimum load, average load, load rate, and other time-series features; peak-hour, valley-hour, and normal-hour electricity consumption ratios, peak-valley differences, curve fluctuations, and other curve features; as well as statistical features such as variance, standard deviation, skewness, and kurtosis. Then, normalization processing is performed, that is, the extracted feature data of different dimensions are normalized and scaled to a uniform scale through Min-Max normalization or Z-Score standardization. Finally, a standard user load feature set is output, with each user corresponding to a feature vector.

[0024] Preferably, density-based DBSCAN or K-means clustering is used to calculate the standard user load feature set. Based on the distance or density of feature vectors in multidimensional space, users with similar features are grouped into the same cluster, while dissimilar users are grouped into different clusters. This grouping of users yields multiple user load feature clusters, forming a user load feature cluster set. Each cluster represents a group of users with similar load characteristics, such as daytime-high and nighttime-low industrial users, continuously stable data center users, and peak-prone commercial users. Then, the available electricity dataset is matched and mapped to the user load feature cluster set for labeling and integration. This involves matching and mapping the label of the cluster to each user data record in the available electricity dataset, and integrating the labeled user data to ultimately construct a structured large user load feature library. This library stores the user's original / available electricity data, the user's standard load feature vector, and the user's load category label.

[0025] Furthermore, step S130 also includes step S131, setting a minimum number of samples based on the dimensional characteristics of the standard user load feature set; step S132, calculating the nearest neighbor distance between each data point in the standard user load feature set and the minimum number of samples, and generating a k-distance sorting graph based on the nearest neighbor distance; step S133, using the slope inflection point of the k-distance sorting graph as the target neighborhood radius; and step S134, performing clustering analysis and validation optimization on the standard user load feature set based on the minimum number of samples and the target neighborhood radius to obtain the user load feature cluster set.

[0026] Preferably, density-based DBSCAN clustering is used to group users with similar load characteristics. Specifically, based on the dimensionality of the standard user load feature set, a minimum sample size is preset. This minimum sample size defines the minimum number of data points required in the neighborhood of a core point, determining the minimum density threshold needed to form a dense region. The preset value of the minimum sample size is related to the dataset dimension; that is, the preset minimum sample size is not less than the number of features plus one. For example, if 10 load features are extracted from electricity consumption data, the minimum sample size is preset to 11. For each data point in the standard user load feature set, the distance to its k-th nearest neighbor is calculated, where k is the minimum sample size. This distance is called the k-distance. After calculating the k-distances of all data points, they are sorted from largest to smallest. A line graph is plotted with the sorted index as the horizontal axis and the k-distance value as the vertical axis, generating a k-distance sorted graph. The curve shows a downward trend.

[0027] Preferably, on the k-distance sorting graph, the point where the slope of the curve changes drastically is identified as the inflection point. Before the inflection point, the k-distance decreases slowly, indicating that most points are relatively close to their k nearest neighbors. After the inflection point, the k-distance increases rapidly, indicating that the data points begin to become sparse and the distances between them increase rapidly. The k-distance value corresponding to this inflection point distinguishes between dense and sparse regions in the dataset, and this value is used as the target neighborhood radius. Data points containing at least a minimum number of samples within this radius are marked as core points, thus forming clusters. Finally, the preset minimum number of samples and the target neighborhood radius are input into the DBSCAN algorithm to start execution. All core points are identified, and mutually density-reachable core points are connected to form clusters. Points that cannot be connected to any cluster are marked as noise. The clustering results are then evaluated and validated based on indicators such as cluster separation and silhouette coefficient until a reasonable clustering result is obtained, ultimately determining a high-quality user load feature cluster set.

[0028] Step S200: Perform load pattern rule mining on the large user load feature database to obtain a large user load pattern identification rule database. Then, traverse and match the large user load pattern identification rule database according to the new user application parameters to determine the new user load pattern.

[0029] Step S200 further includes step S210, extracting large user electricity load dataset from the large user load feature database; step S220, performing K-means clustering and load pattern identification on the large user electricity load dataset to obtain a large user load pattern data cluster; step S230, using decision trees to sequentially perform load pattern rule mining on the large user load pattern data cluster to obtain a large user load pattern cluster rule set; step S240, performing confidence evaluation, filtering, and rule integration on the large user load pattern cluster rule set to obtain the large user load pattern identification rule database.

[0030] Preferably, load pattern rule mining is performed on the large user load feature database. Specifically, the large user electricity load dataset is extracted from the large user load feature database, which includes the standardized load data of each large user and the corresponding load feature cluster. Then, a secondary clustering is performed on the large user electricity load dataset based on K-means clustering. That is, the K-means algorithm is used to cluster user load curves with similar shapes, thereby generating more representative typical load curve templates. Then, load pattern identification is performed, and a load pattern label is assigned to each cluster. For example, clusters with high loads during the day and low loads at night are labeled as "daytime operation type", and clusters with stable loads are labeled as "basic load type". Finally, a load pattern data cluster set with business labels is output, that is, the large user load pattern data cluster set.

[0031] Preferably, a decision tree is then used to sequentially mine load pattern rules for the large user load pattern data clusters. Specifically, the load pattern is used as the prediction target, and the user's industry category, installed capacity, production shift, and other installed parameters, as well as key features extracted from the load data, are used as input features. A decision tree is trained for each load pattern, and a set of decision rules is mined from it to form the large user load pattern cluster rule set. Then, the confidence level of each rule in the large user load pattern cluster rule set is calculated, for example, the proportion of users who are correctly classified among those who meet the rule conditions. Unreliable rules with confidence levels below a preset threshold are removed. Finally, the filtered rules are integrated, similar rules are merged, and redundant and contradictory rules are removed to form a large user load pattern recognition rule library, which is used to automatically and quickly predict the most likely load pattern of new users based on their installed parameters.

[0032] Furthermore, step S200 also includes step S250, converting the new user application parameters according to the large user load pattern identification rule library to obtain standardized application parameters; step S260, performing fuzzy matching on the large user load pattern identification rule library based on the standardized application parameters to obtain a matching load pattern rule set; step S270, sorting the matching load pattern rule set in descending order according to confidence level to obtain a matching load pattern sequence rule, and determining the new user load pattern according to the matching load pattern sequence rule.

[0033] Preferably, the new user's application parameters are format-converted according to the large user load pattern recognition rule base. This involves mapping, quantifying, and standardizing the original form data in the new user's application parameters into feature variables recognized by the large user load pattern recognition rule base, such as "industry type," "application capacity," and "planned electricity usage period." This ensures that the new user's input data is consistent with the rule base. For example, the industry type described in text is converted into the code defined by the rule base, and the unit of application capacity is standardized. This results in a feature vector that is fully aligned with the feature dimensions, data types, and value range of the rule base—that is, the standardized application parameters. Then, the standardized application parameters are traversed within the large user load pattern recognition rule base, and fuzzy matching is performed with the condition part of each rule. For numerical features, a certain deviation range is allowed; for categorical features, a similarity metric is used to calculate a matching score for each rule. Rules with matching scores exceeding a preset matching threshold are then integrated into a matching load pattern rule set, which contains rules whose condition parts have a high similarity to the new user's application parameters. When multiple rules match, the rules in the load pattern matching rule set are sorted in descending order of confidence to determine the load pattern matching sequence rules. The load pattern corresponding to the rule with the highest confidence is directly selected as the new user's pattern. Alternatively, the first N high-confidence rules are considered. If they point to the same pattern, that pattern is determined. If they point to different patterns, a weighted vote is performed to determine the new user's pattern. Finally, the new user's installation parameters are intelligently mapped to the known load patterns, and the new user's load pattern is output.

[0034] Step S300: Combine the target power grid characteristic data to analyze the new user load pattern and generate baseline power expansion scheme parameter thresholds.

[0035] Step S300 further includes step S310, obtaining grid structure data, operational constraint data, historical grid operation data, and expansion planning data based on the target grid characteristic data; step S320, performing characteristic modeling based on the grid structure data, operational constraint data, historical grid operation data, and expansion planning data to generate a target grid characteristic model; and step S330, performing grid expansion scheme analysis on the new user load pattern according to the target grid characteristic model to generate baseline expansion scheme parameter thresholds.

[0036] Preferably, power grid structure data, operational constraint data, historical power grid operation data, and expansion planning data are acquired based on target power grid characteristic data. Power grid structure data refers to the physical connections and component parameters of the power grid, such as the number, capacity, and location of main transformers in substations, the type, length, impedance, and current carrying capacity of lines, and the configuration of switching stations and distribution cabinets. Operational constraint data refers to the mandatory regulations that the power grid must adhere to for safe and stable operation, such as the allowable deviation range of bus voltage, the long-term and short-term current carrying capacity limits of transformers and lines, and the setting range of protection devices. Historical power grid operation data reflects the recent actual operating status of the power grid, such as historical voltage curves of key nodes, historical load rate data of transformers and lines, and the daily / monthly maximum load of the region, reflecting the current health status and potential bottlenecks of the power grid. Expansion planning data refers to the planned future construction or renovation plans of the power grid, such as newly commissioned transformers, new transmission channels to be built, and the upgrade and renovation schedule of the regional power grid.

[0037] Preferably, power system analysis software such as power flow calculation and short-circuit calculation is used to construct an electrical calculation model of the target power grid. Then, power grid structure data, operation constraint data, historical power grid operation data, and expansion planning data are mapped onto the electrical calculation model and characteristic modeling is performed. This ensures that the model includes the topology and component parameters of the power grid, embeds operation constraints as verification conditions for simulation calculations, and can be calibrated using historical operation data. Expansion planning data is used as the future scenario of the model, thereby ensuring that the actual operation characteristics of the power grid are reflected. Finally, a target power grid characteristic model that can be used for simulation calculations is output.

[0038] Preferably, the load patterns of new users are input into the target power grid characteristic model for power grid expansion scheme analysis. Specifically, the typical load curves of new users are superimposed onto the target power grid characteristic model to simulate the operating state of the power grid after user access, analyze whether the voltage of all nodes is within the constraints, and whether the load rate of equipment such as transformers and lines exceeds the limit; then the short-circuit capacity of the user access point is calculated to assess whether the user equipment meets the dynamic and thermal stability requirements; then, based on the analysis results, one or more feasible preliminary access schemes are determined, and the thresholds of key parameters are calculated. For example, based on the remaining capacity of transformers and lines and the impact of voltage disturbances, the maximum allowable application capacity threshold for users is determined; based on the power grid structure and power flow distribution, the optimal location of grid-connected substations, feeders, and access points is determined; based on the voltage level and reactive power balance requirements of the power grid, the power factor assessment threshold that the user side must meet is proposed; finally, the baseline business expansion scheme parameter thresholds are output, including multiple parameter feasible domain boundaries, to ensure that the generated baseline scheme is safe and reliable.

[0039] Step S400: Construct multiple optimization objectives for the business expansion scheme, and perform global optimization on the parameter thresholds of the baseline business expansion scheme according to the multiple optimization objectives to determine the target power grid business expansion scheme.

[0040] Step S400 further includes step S410, which involves evaluating the importance of the multiple optimization objectives of the business expansion scheme and performing weighted normalization to construct the business expansion scheme optimization objective function; step S420, which involves initializing the business expansion scheme particle space based on the baseline business expansion scheme parameter thresholds; and step S430, which involves globally solving and optimizing the business expansion scheme particle space based on the business expansion scheme optimization objective function to determine the target power grid business expansion scheme.

[0041] Preferably, multiple optimization objectives for business expansion schemes are constructed, which may include, but are not limited to, minimizing grid-side investment costs, minimizing grid risks, minimizing additional grid network losses, minimizing voltage fluctuations, and maximizing grid equipment utilization efficiency. These multiple optimization objectives have different dimensions and varying importance. Their importance is assessed and normalized to eliminate the influence of dimensions. Then, domain experts assign a weight coefficient to each optimization objective through hierarchical analysis to reflect its relative importance in decision-making. Each optimization objective sub-function is then constructed, and a weighted sum function is obtained by weighted summation of these sub-functions, serving as the optimization objective function for the business expansion scheme. Based on the baseline business expansion scheme parameter thresholds, an initial search range and population are set for the particle swarm optimization algorithm. Specifically, each particle represents a potential business expansion scheme, containing multiple variable parameters such as access capacity, access point location, and reactive power compensation capacity. The baseline scheme provides a safe initial point and upper and lower bounds for the parameters. Based on the baseline business expansion scheme parameter thresholds, the business expansion scheme particle space is set, i.e., the optimization search range composed of all possible business expansion schemes.

[0042] Preferably, each business expansion scheme in the particle space is substituted into the business expansion scheme optimization objective function for global optimization, i.e., the comprehensive score of each scheme is calculated. The better the objective function value, the better the scheme. Then, each particle, based on its own historical best position and the global best position of the entire particle swarm, simulates social learning, updates its own change direction, step size, and scheme parameters, so that the entire particle swarm moves towards a better scheme region. Iterative optimization is performed until the maximum number of iterations is met or the quality of the solution no longer significantly improves, then the iteration ends, and the historical best position determined by the particle swarm is output. This is the business expansion scheme parameter combination with the highest comprehensive score of the business expansion scheme optimization objective function, which is used as the final target power grid business expansion scheme, thereby improving the scientificity and accuracy of the business expansion scheme.

[0043] Furthermore, step S430 also includes step S431, evaluating the fitness of the particle space of the business expansion scheme based on the optimization objective function of the business expansion scheme to obtain a set of particle fitness for the business expansion scheme; step S432, updating the particle velocity and position of the particle space of the business expansion scheme according to the set of particle fitness for the business expansion scheme until a preset termination condition is met, and finding the target power grid business expansion scheme with the largest particle fitness.

[0044] Preferably, the specific parameters of each candidate business expansion scheme represented by each particle in the business expansion scheme particle space are substituted into the business expansion scheme optimization objective function for calculation to determine the fitness of the particle, thereby obtaining a business expansion scheme particle fitness set composed of the fitness of multiple business expansion scheme particles; then, the particle velocity and position of the business expansion scheme particle space are updated according to the business expansion scheme particle fitness set, that is, each particle updates its change direction and step size, as well as the candidate business expansion scheme parameters, based on its own historical best position and the global best position of the entire particle swarm, guiding the entire particle swarm to move towards a better scheme region until the preset termination condition is met, such as reaching the maximum number of iterations or the solution quality no longer significantly improving, and finally outputting the historical best position found by the particle swarm, which is the business expansion scheme parameter combination with the highest comprehensive score of the business expansion scheme optimization objective function, as the target power grid business expansion scheme with the largest particle fitness.

[0045] Furthermore, step S432 also includes performing non-dominated sorting and crowding calculation based on the particle fitness set of the business expansion scheme to generate a Pareto optimal parameter solution set; selecting a guiding particle from the Pareto optimal parameter solution set, and introducing the guiding particle to iteratively update the particle velocity and position of the particle space of the business expansion scheme.

[0046] Preferably, for a solution A, if there is no other solution B such that B is better or equal to A in all objectives and strictly superior to A in at least one objective, then A is a non-dominated solution or a Pareto optimal solution. Non-dominated sorting is performed based on the particle fitness set of the expansion scheme. This involves comparing all particles to each other, identifying all particles not dominated by any other particles, marking them as the first frontier, removing these particles, and then identifying non-dominated solutions from the remaining particles, marking them as the second frontier, and so on, until the non-dominated sorting is complete, dividing the entire particle swarm into different levels. Crowding degree is used to measure the density between a solution and its neighboring solutions within the same non-dominated frontier. Crowding degree calculations determine the distribution of optimal particles in the target space. Finally, the set of particles at the first frontier is taken as the Pareto optimal parameter solution set, also known as the Pareto frontier. Based on congestion, a particle with lower congestion is selected from the Pareto optimal parameter solution set for each particle as a guide particle. This encourages the particle to explore trade-off regions where there are fewer current solutions and they are not fully developed, thus ensuring that the Pareto front found at the end can broadly and uniformly cover all possible trade-off solutions. Finally, the guide particle is introduced to iteratively update the particle velocity and position of the business expansion scheme particle space, that is, to move towards the selected guide particle position. Finally, the best business expansion scheme is selected to improve the scientificity and accuracy of the business expansion scheme and improve the safe and stable performance of the power grid operation.

[0047] In the above text, refer to Figure 1 This paper describes in detail a multi-objective optimization method for a power load access-driven business expansion scheme according to an embodiment of the present invention. Next, we will refer to... Figure 2 This invention describes a multi-objective optimization system for a power load access driven business expansion scheme according to an embodiment of the present invention.

[0048] The power load access-driven multi-objective optimization system for business expansion schemes according to embodiments of the present invention addresses the technical problems in existing technologies, such as the lack of data-driven decision-making in business expansion schemes, difficulty in accurately matching user load characteristics, and inability to collaboratively optimize multiple conflicting objectives. It achieves the technical effects of efficient allocation of power grid resources, improved scientific rigor and accuracy of business expansion schemes, and enhanced power grid operational safety and stability. Figure 2 As shown, the multi-objective optimization system for power load access driven business expansion schemes includes: power consumption characteristic analysis module 10, business expansion scheme analysis module 30, load mode determination module 20, and power grid business expansion scheme determination module 40.

[0049] The electricity consumption feature analysis module 10 is used to perform feature clustering analysis on historical electricity consumption datasets to construct a large user load feature library; the load pattern determination module 20 is used to perform load pattern rule mining on the large user load feature library to obtain a large user load pattern identification rule library, and to traverse and match the new user's application parameters in the large user load pattern identification rule library to determine the new user's load pattern; the business expansion scheme analysis module 30 is used to analyze the new user's load pattern in conjunction with the target power grid characteristic data to generate a baseline business expansion scheme parameter threshold; the power grid business expansion scheme determination module 40 is used to construct multiple optimization objectives for business expansion schemes, and to globally solve and optimize the baseline business expansion scheme parameter threshold according to the multiple optimization objectives to determine the target power grid business expansion scheme.

[0050] The specific configuration of the electricity consumption characteristic analysis module 10 will be described in detail below. The electricity consumption characteristic analysis module 10 further includes: performing data cleaning processing on the historical electricity consumption dataset to obtain an available electricity consumption dataset; extracting and normalizing load features from the available electricity consumption dataset according to data application standards to obtain a standard user load feature set; performing cluster analysis on the standard user load feature set to obtain a user load feature cluster set; and matching and mapping the available electricity consumption dataset to the user load feature cluster set for labeling and integration to construct the large user load feature library.

[0051] The specific configuration of the electricity consumption characteristic analysis module 10 will be described in detail below. The electricity consumption characteristic analysis module 10 further includes: presetting a minimum sample number based on the dimensional characteristics of the standard user load characteristic set; calculating the nearest neighbor distance between each data point in the standard user load characteristic set and the minimum sample number, and generating a k-distance sorting graph based on the nearest neighbor distance; using the slope inflection point of the k-distance sorting graph as the target neighborhood radius; and performing clustering analysis and validation optimization on the standard user load characteristic set based on the minimum sample number and the target neighborhood radius to obtain the user load characteristic cluster set.

[0052] The specific configuration of the load pattern determination module 20 will be described in detail below. The load pattern determination module 20 further includes: extracting a large user electricity load dataset from the large user load feature database; performing K-means clustering and load pattern identification on the large user electricity load dataset to obtain a large user load pattern data cluster set; using a decision tree to sequentially perform load pattern rule mining on the large user load pattern data cluster set to obtain a large user load pattern cluster rule set; and performing confidence evaluation, filtering, and rule integration on the large user load pattern cluster rule set to obtain the large user load pattern identification rule library.

[0053] The specific configuration of the load pattern determination module 20 will be described in detail below. The load pattern determination module 20 further includes: converting the new user's application parameters according to the large user load pattern identification rule library to obtain standardized application parameters; performing fuzzy matching on the standardized application parameters within the large user load pattern identification rule library to obtain a matching load pattern rule set; arranging the matching load pattern rule set in descending order of confidence level to obtain a matching load pattern sequence rule; and determining the new user's load pattern according to the matching load pattern sequence rule.

[0054] The specific configuration of the business expansion scheme analysis module 30 will be described in detail below. The business expansion scheme analysis module 30 further includes: acquiring power grid structure data, operational constraint data, historical power grid operation data, and expansion planning data based on the target power grid characteristic data; performing characteristic modeling based on the power grid structure data, operational constraint data, historical power grid operation data, and expansion planning data to generate a target power grid characteristic model; and performing power grid business expansion scheme analysis on the new user load pattern according to the target power grid characteristic model to generate baseline business expansion scheme parameter thresholds.

[0055] The specific configuration of the power grid business expansion scheme determination module 40 will be described in detail below. The power grid business expansion scheme determination module 40 further includes: performing importance assessment and weighted normalization on the multiple optimization objectives of the business expansion scheme to construct an optimization objective function for the business expansion scheme; initializing the business expansion scheme particle space based on the baseline business expansion scheme parameter thresholds; and performing global optimization on the business expansion scheme particle space based on the optimization objective function to determine the target power grid business expansion scheme.

[0056] The specific configuration of the power grid expansion scheme determination module 40 will be described in detail below. The power grid expansion scheme determination module 40 further includes: evaluating the fitness of the particle space of the expansion scheme based on the optimization objective function of the expansion scheme to obtain a set of particle fitness for the expansion scheme; updating the particle velocity and position of the particle space of the expansion scheme according to the set of particle fitness for the expansion scheme until a preset termination condition is met, and optimizing to determine the target power grid expansion scheme with the highest particle fitness.

[0057] The specific configuration of the power grid expansion scheme determination module 40 will be described in detail below. The power grid expansion scheme determination module 40 further includes: performing non-dominated sorting and congestion calculation based on the particle fitness set of the expansion scheme to generate a Pareto optimal parameter solution set; selecting a guiding particle from the Pareto optimal parameter solution set, and introducing the guiding particle to iteratively update the particle velocity and position of the particle space of the expansion scheme.

[0058] The power load access-driven business expansion scheme multi-objective optimization system provided in the embodiments of the present invention can execute the power load access-driven business expansion scheme multi-objective optimization method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0059] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A multi-objective optimization method for business expansion schemes driven by power load access, characterized in that, The method includes: Perform feature clustering analysis on historical electricity consumption datasets to construct a large user load feature database; Load pattern rule mining is performed on the large user load feature database to obtain a large user load pattern identification rule database. The new user load pattern is determined by traversing and matching the large user load pattern identification rule database according to the new user application parameters. By combining the target power grid characteristic data, the new user load pattern is analyzed to determine the power grid expansion scheme and generate baseline expansion scheme parameter thresholds. Construct multiple optimization objectives for the business expansion scheme, and perform global optimization on the parameter thresholds of the baseline business expansion scheme according to the multiple optimization objectives to determine the target power grid business expansion scheme.

2. The multi-objective optimization method for power load access driven business expansion scheme as described in claim 1, characterized in that, Construct a large user load characteristic library, including: The historical electricity consumption dataset is cleaned to obtain an available electricity consumption dataset. According to the data application standards, the available electricity dataset is subjected to load feature extraction and normalization to obtain a standard user load feature set; Cluster analysis is performed on the standard user load feature set to obtain user load feature clusters; The available electricity data set is matched and mapped to the user load feature clusters for labeling and integration to construct the large user load feature library.

3. The multi-objective optimization method for power load access driven business expansion scheme as described in claim 2, characterized in that, The user load feature clusters are obtained, including: Based on the dimensional characteristics of the standard user load feature set, a minimum number of samples is preset; Calculate the nearest neighbor distance between each data point in the standard user load feature set and the minimum number of samples, and generate a k-distance sorting map based on the nearest neighbor distance; The slope inflection point of the k-distance sorting graph is taken as the target neighborhood radius; Clustering analysis and validation optimization are performed on the standard user load feature set based on the minimum number of samples and the target neighborhood radius to obtain the user load feature cluster set.

4. The multi-objective optimization method for power load access driven business expansion scheme as described in claim 1, characterized in that, We obtained a large user load pattern recognition rule base, including: Extract large user electricity load datasets from the large user load feature library; Perform K-means clustering and load pattern identification on the large user electricity load dataset to obtain a large user load pattern data cluster set; A decision tree is used to sequentially mine load pattern rules for the large user load pattern data clusters to obtain a large user load pattern cluster rule set. The large user load pattern cluster rule set is subjected to confidence evaluation, screening, and rule integration to obtain the large user load pattern recognition rule base.

5. The multi-objective optimization method for power load access driven business expansion scheme as described in claim 1, characterized in that, Determine the load pattern for new users, including: The new user installation parameters are converted according to the large user load pattern recognition rule base to obtain standardized installation parameters. Based on the standardized application parameters, a fuzzy matching is performed within the large user load pattern recognition rule base to obtain a matching load pattern rule set; The matching load pattern rule set is sorted in descending order of confidence level to obtain the matching load pattern sequence rule, and the new user load pattern is determined according to the matching load pattern sequence rule.

6. The multi-objective optimization method for power load access driven business expansion scheme as described in claim 1, characterized in that, Generate baseline business expansion plan parameter thresholds, including: Based on the target power grid characteristic data, obtain power grid structure data, operation constraint data, historical power grid operation data, and capacity expansion planning data; Based on the power grid structure data, operation constraint data, historical power grid operation data, and expansion planning data, characteristic modeling is performed to generate a target power grid characteristic model. The new user load pattern is analyzed according to the target power grid characteristic model to generate baseline power expansion scheme parameter thresholds.

7. The multi-objective optimization method for power load access driven business expansion scheme as described in claim 1, characterized in that, Determine the target power grid expansion plan, including: The importance of the multiple optimization objectives of the business expansion scheme is evaluated and weighted normalized to construct the optimization objective function of the business expansion scheme; Initialize the business expansion scheme particle space based on the baseline business expansion scheme parameter thresholds; Based on the objective function of the business expansion scheme, the global solution of the business expansion scheme particle space is performed to find the optimal solution and determine the target power grid business expansion scheme.

8. The multi-objective optimization method for power load access driven business expansion scheme as described in claim 7, characterized in that, Based on the objective function of the business expansion scheme, a global solution is performed in the particle space of the business expansion scheme to determine the target power grid business expansion scheme, including: Based on the optimization objective function of the business expansion scheme, the fitness of the particle space of the business expansion scheme is evaluated to obtain the particle fitness set of the business expansion scheme. The particle velocity and position of the business expansion scheme are updated according to the particle fitness set of the business expansion scheme until a preset termination condition is met, and the target power grid business expansion scheme with the largest particle fitness is determined.

9. The multi-objective optimization method for power load access driven business expansion scheme as described in claim 8, characterized in that, According to the particle fitness set of the expansion scheme, the particle velocity and position of the particle space of the expansion scheme are updated, including: Based on the particle fitness set of the aforementioned business expansion scheme, non-dominated sorting and crowding calculation are performed to generate the Pareto optimal parameter solution set; A guiding particle is selected from the Pareto optimal parameter solution set, and the guiding particle is introduced to iteratively update the particle velocity and position of the particle space of the expansion scheme.

10. A multi-objective optimization system for business expansion schemes driven by power load access, characterized in that, The system is used to implement the multi-objective optimization method for power load access driven business expansion schemes as described in any one of claims 1 to 9, and the system includes: The electricity consumption characteristic analysis module is used to perform feature clustering analysis on historical electricity consumption datasets and build a large user load characteristic library; The load pattern determination module is used to perform load pattern rule mining on the large user load feature library to obtain a large user load pattern identification rule library, and to traverse and match the large user load pattern identification rule library according to the new user application parameters to determine the new user load pattern. The business expansion scheme analysis module is used to analyze the new user load pattern of the power grid business expansion scheme by combining the target power grid characteristic data, and generate baseline business expansion scheme parameter thresholds. The power grid business expansion scheme determination module is used to construct multiple optimization objectives for the business expansion scheme, and to perform global optimization on the parameter thresholds of the baseline business expansion scheme according to the multiple optimization objectives to determine the target power grid business expansion scheme.

Images