County power distribution network multi-dimensional planning index flexible construction and evaluation method, system, device and medium

Abstract

The invention relates to the technical field of power distribution network planning evaluation, and discloses a county power distribution network multi-dimensional planning index flexible construction and evaluation method, system, device and medium, comprising the steps of analyzing operation characteristics of a county power distribution network fusing various flexible resources, and combining energy endowment, flexible resource configuration and digitization degree characteristics of different county power distribution networks to obtain a multi-dimensional planning index of the county power distribution network; dividing time-space resource fusion county power distribution network planning scenes, and constructing multi-dimensional planning evaluation indexes of different scenes; a data driving method is adopted, importance and correlation between a planning scheme and indexes are considered, a game theory is adopted to comprehensively sort the indexes, multi-dimensional planning evaluation indexes based on data driving are constructed, and the planning scheme is evaluated through an entropy weight method. According to the method, a multi-dimensional planning evaluation index system based on data driving can be constructed for different planning scenes, evaluation is carried out by using an entropy weight method, and finally scene, difference and flexibility evaluation of county power distribution network planning is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of distribution network planning and evaluation, and particularly to a method, system, device and medium for flexibly constructing and evaluating multi-dimensional planning indicators of a county-level distribution network. Background Art

[0002] A scientific and reasonable distribution network planning and evaluation method can effectively evaluate the impact of distributed power source access on the power supply reliability of the system, optimize the power source layout, optimize the distribution network structure and equipment configuration, evaluate the system flexibility and new energy consumption capacity, evaluate the digital and intelligent level and the safe and stable control ability of the system, identify the weak links in the system, provide a scientific decision-making basis for the distribution network planning and construction, ensure the feasibility and effectiveness of the planning scheme, and provide comprehensive reference for decision-makers.

[0003] In this context, the planning of county-level distribution networks not only needs to comprehensively reflect the characteristics of multiple aspects such as technology, economy, environment and society, but also must be highly flexible and digital and intelligent to adapt to the dynamic development needs of different counties. The necessity of flexibly constructing the planning index system is mainly reflected in the following aspects: First, it can comprehensively evaluate the operation characteristics and planning effects of the distribution network; Second, it ensures that the index system can be differentially and dynamically adjusted according to different scenarios to meet the specific needs of the distribution network in different counties and different development stages. Summary of the Invention

[0004] In view of the above existing problems, the present invention is proposed. Therefore, the present invention provides a method for flexibly constructing and evaluating multi-dimensional planning indicators of a county-level distribution network to solve the above problems.

[0005] To solve the above technical problems, the present invention provides the following technical solutions: In the first aspect, the present invention provides a method for flexibly constructing and evaluating multi-dimensional planning indicators of a county-level distribution network, including: analyzing and integrating the operation characteristics of a county-level distribution network with multiple flexibility resources, and combining the energy endowments, flexible resource configurations, and digitalization degree characteristics of different county-level distribution networks to divide the planning scenarios of the county-level distribution network with spatio-temporal resource integration; Based on the results of different planning scenario divisions, according to the complete index system for county-level distribution network planning and evaluation, construct multi-dimensional planning evaluation indicators for different scenarios; Based on the multi-dimensional planning evaluation indicators for different scenarios, adopt a data-driven method, take into account the importance and relevance between the planning scheme and the indicators, and comprehensively sort the indicators using game theory to construct data-driven multi-dimensional planning evaluation indicators; Based on the data-driven multi-dimensional planning evaluation indicators, evaluate the planning scheme by the entropy weight method.

[0006] As a preferred solution of a method for flexibly constructing and evaluating multi-dimensional planning indicators of a county-level distribution network according to the present invention, wherein: the division of the county-level distribution network planning scenario for spatio-temporal resource integration includes:

[0007] The division of the county-level distribution network planning scenario for spatio-temporal resource integration includes dividing the energy endowment, flexible resource allocation, and digitalization degree of different power supply areas or different development periods of power supply areas; The energy endowment divides energy resources into energy input type and energy output type. The energy input type means that the installed capacity of centralized and distributed energy in the county-level distribution network planning scheme is small, and the energy output type refers to that the installed capacity of centralized and distributed energy in the county-level distribution network planning scheme is large; The flexible resource allocation means the installed capacity size of flexible resources in the county-level distribution network planning scheme; The digitalization degree means the degree of digitalization configuration in the county-level distribution network planning scheme.

[0008] As a preferred solution of a method for flexibly constructing and evaluating multi-dimensional planning indicators of a county-level distribution network according to the present invention, wherein: constructing multi-dimensional planning evaluation indicators for different scenarios includes: The complete evaluation index system for the county-level distribution network planning includes multiple primary indicators, secondary indicators, and tertiary indicators in multiple aspects such as clean and low-carbon, safe and abundant, economic and efficient, flexible and intelligent, and supply-demand coordination; According to the complete evaluation index system for the county-level distribution network planning, combined with the characteristic data of different planning scenarios, the planning evaluation indicators for different scenarios are constructed differentially, and the planning evaluation indicators for different scenarios are screened out from the planning evaluation index system.

[0009] As a preferred solution of a method for flexibly constructing and evaluating multi-dimensional planning indicators of a county-level distribution network according to the present invention, wherein: constructing multi-dimensional planning evaluation indicators based on data-driven includes: Adopt a data-driven method to preliminarily screen the potential indicators of each scenario through the maximum correlation minimum redundancy algorithm, maintain the maximum correlation between the indicators and the target variable, and minimize the redundant information between the indicators; Evaluate the importance of the screened indicators through the random forest algorithm to handle the non-linear relationship between the indicators; Through the game theory optimization model, combine the correlation and importance to sort and screen the planning indicators, and allocate and dynamically adjust the indicator weights.

[0010] As a preferred solution of a method for flexibly constructing and evaluating multi-dimensional planning indicators of a county-level distribution network according to the present invention, wherein: preliminarily screening the potential indicators of each scenario through the maximum correlation minimum redundancy algorithm, maintaining the maximum correlation between the indicators and the target variable, and minimizing the redundant information between the indicators includes: Input dataset for indicator and target variable c, perform data standardization and missing value processing; Calculate each feature With the target variable Mutual information between , expressed as: , in, represents the joint probability distribution, and represents the marginal probability distribution; Calculate mutual information between features , evaluate the redundancy between features, expressed as: , Perform MIQ criterion to calculate the correlation between features and target variables , expressed as: , Perform redundancy measurement calculation and calculate the average mutual information between the feature and the features in the selected feature set S , expressed as: , in, Indicates the size of the selected feature set; In each iteration, the feature with the maximum MIQ score is selected and added to the set of selected features to obtain the mRMR importance score. , expressed as: , in, Represents the total number of indicators that are effectively ranked by mRMR, and is the indicator The above steps are repeated until the preset number of features is reached or the MIQ score is lower than the set threshold.

[0011] As a preferred solution of the flexible construction and evaluation method of multi-dimensional planning indicators of a county distribution network described in the present invention, the importance of the selected indicators is evaluated by a random forest algorithm, and the nonlinear relationship between the indicators is processed, including: Bootstrap sampling is performed on the training set to randomly generate k sample subsets Si and the corresponding decision tree Ti; When a node splits, m features are randomly selected from m features. try features as candidate feature sets; Calculate the optimal splitting point of the candidate features for each node, calculate the mean squared error of the parent node using different splitting thresholds, and at the same time calculate the mean squared errors of the left and right child nodes. The splitting gain is calculated as follows: , Select the feature with the largest splitting gain and the corresponding splitting point. Select the feature with the largest reduction in mean squared error for splitting, and then calculate the importance of a single feature in a single tree, which is expressed as: , Calculate the average importance of the feature in the entire forest, which is expressed as: , Obtain the importance of the feature through normalization processing and output the sorted feature importance.

[0012] As a preferred solution of the method for flexibly constructing and evaluating multi-dimensional planning indicators of a county-level distribution network according to the present invention, the evaluation of the planning scheme by the entropy weight method includes: Set the index types, where the index types include benefit-type indicators, cost-type indicators, and intermediate-type indicators. Divide the planning evaluation indicators according to the definitions of each index type, and perform data preprocessing on the cost-type indicators and intermediate-type indicators; Introduce a probability matrix by the entropy weight method, calculate the information entropy value of the information contained in each indicator, and calculate the redundancy and index weight of each level of indicators based on the information entropy value.

[0013] In a second aspect, the present invention provides a system for flexibly constructing and evaluating multi-dimensional planning indicators of a county-level distribution network, including: A division module for analyzing and integrating the operating characteristics of a county-level distribution network with multiple flexibility resources, and combining the energy endowment, flexible resource configuration, and digitalization degree characteristics of different county-level distribution networks to divide the planning scenarios of the county-level distribution network with spatio-temporal resource integration; A first construction module for constructing multi-dimensional planning evaluation indicators for different scenarios based on the division results of different planning scenarios and according to the complete index system for evaluating the county-level distribution network planning. The complete index system for evaluating the county-level distribution network planning includes multiple first-level indicators, second-level indicators, and third-level indicators in multiple aspects such as clean and low-carbon, safe and abundant, economic and efficient, flexible and intelligent, and supply-demand coordination; A second construction module for constructing data-driven multi-dimensional planning evaluation indicators based on the multi-dimensional planning evaluation indicators of different scenarios, using a data-driven method, taking into account the importance and correlation between the planning scheme and the indicators, and comprehensively sorting the indicators using game theory; An evaluation module for evaluating the planning scheme by the entropy weight method based on the data-driven multi-dimensional planning evaluation indicators.

[0014] In a third aspect, the present invention provides a computer device, comprising: a memory and a processor;

[0015] The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the method for flexibly constructing and evaluating multi-dimensional planning indicators of a county-level distribution network are implemented.

[0016] In a fourth aspect, the present invention provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the method for flexibly constructing and evaluating multi-dimensional planning indicators of a county-level distribution network.

[0017] Compared with the prior art, the beneficial effects of the present invention are as follows: The present invention adopts a multi-stage and multi-dimensional comprehensive evaluation method to evaluate the county-level distribution network planning scheme. This method not only considers the complexity and diversity of the operation of the distribution network, but also integrates advanced data analysis technologies and machine learning algorithms. Based on in-depth analysis of the operating characteristics of the county-level distribution network integrating various flexibility resources, combined with the energy endowment, flexible resource allocation, digitalization degree and other characteristics of different county-level distribution networks, the spatio-temporal resource integration planning scenarios are divided; for different planning scenarios, according to the distribution system planning evaluation index system and based on expert experience, the planning evaluation indexes of different scenarios are screened; on the basis of the scenario-based multi-dimensional planning evaluation indexes, a data-driven method is adopted, taking into account the importance and relevance between the planning scheme and the indexes, and game theory is used for comprehensive ranking; a data-driven multi-dimensional planning evaluation index system is constructed, and the entropy weight method is used for evaluation, finally realizing the scenario-based, differential and flexible evaluation of the county-level distribution network planning. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0019] Figure 1 It is a schematic diagram of the overall process of the method for flexibly constructing and evaluating multi-dimensional planning indicators of a county-level distribution network according to an embodiment of the present invention; Figure 2 It is a diagram of the ranking of the importance of the object-scenario two mRMR screening features of the method for flexibly constructing and evaluating multi-dimensional planning indicators of a county-level distribution network provided by an embodiment of the present invention; Figure 3The object - scenario two random forest screening feature importance ranking diagram of a method for flexibly constructing and evaluating multi - dimensional planning indicators of a county - level distribution network provided by an embodiment of the present invention; Figure 4 The object - scenario two game theory screening feature importance ranking diagram of a method for flexibly constructing and evaluating multi - dimensional planning indicators of a county - level distribution network provided by an embodiment of the present invention; Figure 5 The object - scenario two entropy weight method overall score diagram of a method for flexibly constructing and evaluating multi - dimensional planning indicators of a county - level distribution network provided by an embodiment of the present invention; Figure 6 The object two - scenario six mRMR screening feature importance ranking diagram of a method for flexibly constructing and evaluating multi - dimensional planning indicators of a county - level distribution network provided by an embodiment of the present invention; Figure 7 The object two - scenario six random forest screening feature importance ranking diagram of a method for flexibly constructing and evaluating multi - dimensional planning indicators of a county - level distribution network provided by an embodiment of the present invention; Figure 8 The object two - scenario six game theory screening feature importance ranking diagram of a method for flexibly constructing and evaluating multi - dimensional planning indicators of a county - level distribution network provided by an embodiment of the present invention; Figure 9 The object two - scenario six entropy weight method overall score diagram of a method for flexibly constructing and evaluating multi - dimensional planning indicators of a county - level distribution network provided by an embodiment of the present invention; Figure 10 The complete index system for the planning evaluation of a county - level distribution network of a method for flexibly constructing and evaluating multi - dimensional planning indicators of a county - level distribution network provided by an embodiment of the present invention Figure 1 ; Figure 11 The complete index system for the planning evaluation of a county - level distribution network of a method for flexibly constructing and evaluating multi - dimensional planning indicators of a county - level distribution network provided by an embodiment of the present invention Figure 2 。 Detailed implementation manners

[0020] To make the above - mentioned objects, features, and advantages of the present invention more obvious and understandable, the following will describe the detailed implementation manners of the present invention with reference to the accompanying drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0021] Example 1, refer to Figure 1 、 Figures 10 - 11As well as Tables 1 - 4, an embodiment of the present invention provides a method for flexibly constructing and evaluating multi - dimensional planning indicators for a county - level distribution network, including: S101, analyze the operating characteristics of the county - level distribution network integrating various flexibility resources, and combine the energy endowments, flexible resource allocations, and digitalization degree characteristics of different county - level distribution networks to divide the spatio - temporal resource - integrated county - level distribution network planning scenarios; S102, based on the results of different planning scenario divisions, construct multi - dimensional planning evaluation indicators for different scenarios according to the complete evaluation index system of the county - level distribution network planning; S103, based on the multi - dimensional planning evaluation indicators for different scenarios, adopt a data - driven method, consider the importance and relevance between the planning scheme and the indicators, and comprehensively sort the indicators using game theory to construct data - driven multi - dimensional planning evaluation indicators; S104, based on the data - driven multi - dimensional planning evaluation indicators, evaluate the planning scheme through the entropy weight method.

[0022] It should be noted that in this embodiment, by data - driven analysis of the characteristics of the county - level distribution network, integrating various flexibility resources, a scenario - based, differentiated, and multi - dimensional planning index system is constructed, realizing the transformation of planning evaluation from experience - driven to data - driven, assisting the green transformation of the county - level distribution network, and promoting the construction of a new type of distribution system.

[0023] In a preferred implementation manner, the division of the spatio - temporal resource - integrated county - level distribution network planning scenarios includes: The division of the spatio - temporal resource - integrated county - level distribution network planning scenarios includes dividing the energy endowments, flexible resource allocations, and digitalization degrees of different power supply areas or different development periods of the power supply area; The energy endowment classifies energy resources into energy - input - type and energy - output - type. The energy - input - type means that in the county - level distribution network planning scheme, the installed capacities of centralized and distributed energy are small, and the energy - output - type means that in the county - level distribution network planning scheme, the installed capacities of centralized and distributed energy are large; The flexible resource allocation represents the installed capacity size of the flexibility resources in the county - level distribution network planning scheme; The digitalization degree represents the degree of digitalization configuration in the county - level distribution network planning scheme.

[0024] Specifically, the energy endowment classifies energy resources into input - type and output - type. The energy - input - type means that in the county - level distribution network planning scheme, the installed capacities of centralized and distributed energy are small, and the energy supply of this power supply area mainly relies on the superior power grid. The energy - output - type means that in the county - level distribution network planning scheme, the installed capacities of centralized and distributed energy are large, and in addition to meeting the load demand of this power supply area, there is a certain degree of power transmission capacity.

[0025] The input and output type judgments are made by combining the energy density and load density in the county-level distribution network planning scheme, and the principles are as follows: Energy density = total installed capacity / power supply area (MW / km 2 ) Load density = maximum power supply load / power supply area (MW / km 2 ) Table 1: Input and Output Type Judgment Table for County-level Distribution Network Planning Scheme , Flexible resource allocation refers to the installed capacity of flexible resources in the county-level distribution network planning scheme. Flexible resources include distributed power sources, energy storage, electric vehicles, demand-side response resources, etc. Among them, distributed power sources cover distributed photovoltaics, wind power generation, and micro hydropower, etc.; energy storage systems include electrochemical energy storage, physical energy storage, and hydrogen energy storage; electric vehicles can be used as mobile energy storage devices and also provide regulation capabilities through the charging network; demand-side response resources, such as adjustable load devices of industrial and commercial users, smart home appliances, and traditional power generation devices with adjustable output, etc.

[0026] The degree of digitalization refers to the degree of digital configuration in the county-level distribution network planning scheme. The differential configuration requirements for digital substations, lines, and distribution transformers are shown in Table 2.

[0027] Table 2: Differential Configuration Table for Digital Substations, Lines, and Distribution Transformers , According to the grading requirements of the digitalization degree, for the digital configuration scheme of the county-level distribution network, the design requirements of key links such as substations, lines, equipment, and users should be comprehensively considered according to different levels of the power supply area, and differential configuration strategies should be formulated: For Class A+ areas, as the priority development areas, high-level configurations should be comprehensively adopted for core facilities such as digital substations, distribution lines, and medium-voltage equipment to ensure that the equipment has high automation, intelligence, and remote control capabilities, and support refined management and efficient operation; Class A areas need to maintain the same high-standard configuration as Class A+ areas, and further strengthen the coverage rate of digital distribution stations and intelligent terminal equipment to meet the high-reliability power supply requirements.

[0028] For Class B areas, medium-level configurations should be adopted for substations and lines, gradually realizing the standardized construction of distribution stations and medium-voltage equipment, while taking into account the coverage of digital user terminals to meet the basic intelligent power supply requirements within the area; Class C areas are mainly based on standardized configurations. Digital substations and lines should ensure the stability of basic functions, and at the same time appropriately configure medium-voltage equipment to achieve reliable power supply within the area; Class D areas focus on the flexible configuration of core facilities. For example, digital substations and user terminal facilities are arranged according to actual situations to ensure the digitalization capabilities of basic power supply services.

[0029] According to the above principles, the planning scenarios of the county-level distribution network can be divided into 8 typical scenarios as shown in Table 3.

[0030] Table 3: Table of 8 typical scenarios , In a preferred embodiment, the multi-dimensional planning evaluation indicators for different scenarios include: The complete index system for the evaluation of the county-level distribution network planning includes multiple first-level indicators, second-level indicators, and third-level indicators in multiple aspects such as clean and low-carbon, safe and abundant, economic and efficient, flexible and intelligent, and supply-demand coordination; According to the complete index system for the evaluation of the county-level distribution network planning, combined with the characteristic data of different planning scenarios, the planning evaluation indicators for different scenarios are constructed differentially, and the planning evaluation indicators for different scenarios are screened out from the index system.

[0031] Specifically, the complete index system for the evaluation of the distribution network planning is constructed from five aspects: clean and low-carbon, safe and abundant, economic and efficient, flexible and intelligent, and supply-demand coordination, with a total of 5 first-level indicators, 13 second-level indicators, and 55 third-level indicators. According to the complete index system for the evaluation of the distribution network planning, based on different planning scenarios and expert experience, the planning evaluation indicators for different scenarios are screened to construct the evaluation index system for the distribution system planning, and the planning evaluation indicators for different scenarios are constructed differentially.

[0032] Among them, for example, Scenario 1 is applicable to scenarios where the energy is input-based, the degree of digitization is low, the flexible resource allocation ability is limited, and it focuses on improving the power supply reliability through network improvement; Scenario 2 is applicable to scenarios where digital technology is used to improve power supply safety and flexibility, but the flexible resource allocation is still low; Scenario 3 has a relatively high flexible resource allocation and requires efficient collaborative management of flexible resources, but the digital participation degree is low; Scenario 4 is applicable to scenarios with high digitization and flexible resource collaboration, focusing on intelligent management and supply-demand matching; Scenario 5 is applicable to scenarios with the goal of low-carbonization, low flexible resource allocation, and low digitization degree; Scenario 6 is applicable to scenarios that combine low-carbonization and digitization, emphasizing intelligent low-carbon development; and Scenario 7 is applicable to scenarios with relatively high flexible resource allocation but low digitization degree; Scenario 8 is applicable to scenarios that comprehensively combine high digitization, flexible resource allocation, and low-carbonization goals.

[0033] The purpose of setting different scenarios in the distribution network planning is to meet the differentiated needs of different objects in the same period, as well as the development and change needs of the same object in different periods. Different regions or users vary in terms of resource endowment, load demand, technical conditions, and policy environment. The planning objectives and key concerns will also change dynamically over time. Therefore, by constructing differentiated scenarios based on expert experience, it is possible to conduct matching evaluations for different scenarios such as traditional distribution networks, smart distribution networks, or new-type distribution networks.

[0034] According to expert experience, the screening of planning evaluation indicators for different scenarios is shown in Table 4.

[0035] Table 4: Screening Table of Planning Evaluation Indicators for Different Scenarios , , , , , In a preferred implementation, the construction of multi-dimensional planning evaluation indicators based on data-driven includes: Adopt a data-driven method to preliminarily screen the potential indicators of each scenario through the maximum correlation minimum redundancy algorithm, maintaining the maximum correlation between the indicators and the target variable and minimizing the redundant information between the indicators; Evaluate the importance of the screened indicators through the random forest algorithm to handle the non-linear relationship between the indicators; Through the game theory optimization model, combine correlation and importance to sort and screen the planning indicators, and allocate and dynamically adjust the indicator weights.

[0036] In an alternative implementation, for the maximum correlation minimum redundancy (mRMR) algorithm, mutual information, as an effective information measurement method, can not only measure the linear relationship between variables but also evaluate the non-linear relationship well. The mutual information between two random variables and is defined as: , where and are the probability densities of the random variables and respectively, and is the joint probability density of the random variables and . The definitions of maximum correlation and minimum redundancy are respectively expressed as: , , Among them, is the feature set, is the number of features in the feature set, is the target variable, is the feature and the target variable mutual information between, is the mutual information between the features in the feature set and the target variable mean value of mutual information between, is the feature and the feature mutual information between, is the mutual information between the features in the feature set.

[0037] Combining the above formula, the maximum correlation and minimum redundancy criterion is obtained, which is expressed as: , In a preferred embodiment, based on the mRMR feature selection process, the potential indicators of each scenario are preliminarily screened by the maximum correlation and minimum redundancy algorithm, and the maximum correlation between the indicators and the target variable is maintained, and the redundant information between the indicators is minimized, including: In the data input stage, data preprocessing is first performed, and the indicator input data set and the target variable c are standardized and missing value processed; Calculate the mutual information between each feature and the target variable between, is expressed as: , Among them, represents the joint probability distribution, and represent marginal probability distributions; Calculate the mutual information between features , evaluate the redundancy between features, which is expressed as: , Perform the MIQ criterion to calculate the correlation between the feature and the target variable , which is expressed as: , Perform redundancy measurement calculation, calculate the average mutual information between the feature and the features in the selected feature set S , which is expressed as: , Among them, represents the size of the selected feature set; In each iteration, the feature with the maximum MIQ score is selected, denoted as: , and the selected feature is added to the set of selected features, denoted as: , The mRMR importance score is obtained, denoted as: , where respectively represent the total number of indicators for the effective sorting of mRMR, and is the position of the indicator in the mRMR sorted list. Repeat the above steps until the preset number of features is reached or the MIQ score is lower than the set threshold. Finally, the final optimal feature subset will be selected. Among them, the preset number of features and the set threshold can be set according to the actual application scenario requirements or expert experience.

[0038] In an alternative embodiment, the basic idea of the random forest algorithm is to use multiple decision trees to train and predict the sample sequence. It can be divided into two parts: decision tree generation and forest generation. In the regression problem of this embodiment, the optimal split point is mainly determined by the mean squared error (MSE), and the importance of each feature is evaluated based on the reduction in node impurity.

[0039] Assume that the node contains samples, and its mean squared error can be expressed as: , where is the actual value of the sample, is the average value of all samples in the node ; When using the feature to split the node , the split gain can be expressed as: , where and are the number of samples in the left and right child nodes respectively, and are the mean squared errors of the corresponding child nodes; In the entire random forest, the importance evaluation of a single decision tree for the feature is expressed as: , where is the sample weight of the node , is the impurity reduction amount of this node.

[0040] In a preferred embodiment, the importance of the screened metrics is evaluated through the random forest algorithm, and the processing of the non-linear relationship between metrics includes: Perform Bootstrap sampling on the training set, randomly generate k sample subsets Si, and simultaneously generate corresponding decision trees Ti; When splitting a node, randomly select m try features from m features as the candidate feature set; Calculate the optimal split point of the candidate features for each node, calculate the mean squared error of the parent node, and try different split thresholds. Among them, the split threshold can be set according to the actual application scenario requirements or expert experience. At the same time, calculate the mean squared error of the left and right child nodes, and calculate the split gain, which is expressed as: , Select the feature with the largest split gain and the corresponding split point, select the feature with the largest reduction in mean squared error for splitting, and then calculate the importance of a single feature in a single tree, which is expressed as: , Calculate the average importance of the feature in the entire forest, which is expressed as: , Obtain the importance of feature j through normalization processing, which is expressed as: , where T is the total number of decision trees, is the sum of the importance of all features; Finally, output the feature importance ranking.

[0041] In an alternative embodiment, a game theory optimization model is constructed based on the mRMR algorithm and the random forest algorithm. Game theory uses the Shapley value for cooperative games. Formally, the definition of a coalition game is: there is a set ( ones) and a function , which maps subsets of these players to real numbers: , where represents the empty set, and the function is called the characteristic function.

[0042] The meaning of the function is as follows: If S is a coalition of players, then v(S) is called the value of the coalition , representing the total expected reward that the members of can obtain through cooperation.

[0043] The Shapley value is a way to distribute the total payoff to the participants, assuming they all cooperate. It is a "fair" distribution because it is the only distribution with certain desirable properties. According to the Shapley value, in a given coalition game the score payoff that a player gets is expressed as: , where is an arbitrary subset of all players in excluding , is the number of elements in the subset , that is, the number of other players excluding the player , is the total number of players in the set , represents the payoff brought by the coalition formed after adding the player to the set , represents the payoff brought by the coalition consisting only of the players in the set .

[0044] The specific process of the Shapley value game theory is as follows: Step 1: Data input and preprocessing Load the feature dataset from the calculation results of mRMR and random forest , fill the missing values in the dataset with the mean, and at the same time use Min-Max normalization to scale the feature values to the interval [0,1].

[0045] Step 2: Calculation of the correlation of mRMR features Feature selection, use the mRMR algorithm to sort the features and select features according to the MIQ index; Importance scoring, assign importance scores according to the selection order of the features, expressed as: , where is the total number of features, is the feature in the selection order in the mRMR algorithm.

[0046] Step 3: Calculation of random forest feature importance Model training: Train a random forest regression model on the preprocessed data; Importance extraction: Extract the feature importance of the model, calculated based on the reduction in mean squared error, expressed as: , ​​​​​​SHAP value calculation: Use the trained random forest model to calculate the SHAP value of each feature: , Importance evaluation: Take the average absolute value of the SHAP values as the importance of the features: , Step 4: Normalization Min-Max normalization: Normalize the feature importance obtained by each method to ensure that the importance scores are on the same scale, expressed as: , where k represents the mRMR, RF or SHAP method.

[0047] Step 5: Feature importance fusion based on Shapley values Fusion strategy: Use the average value as the fusion method to combine the three normalized importance scores, expressed as: , It should be noted that from the perspective of game theory, each feature evaluation method is regarded as a "player" in game theory, and the fused importance score is equivalent to the average of the "Shapley values" contributed by the features under different methods. By integrating the advantages of multiple feature evaluation methods, a more robust feature importance evaluation result can be obtained.

[0048] In a preferred implementation, the evaluation of the planning scheme by the entropy weight method includes: Set the index type, which includes benefit type indicators, cost type indicators, and intermediate type indicators. Divide the planning evaluation indicators according to the definitions of each index type, and perform data preprocessing on the cost type indicators and intermediate type indicators; Introduce a probability matrix by the entropy weight method, calculate the information entropy value of the information contained in each indicator, and calculate the redundancy and indicator weight of each level of indicators based on the information entropy value.

[0049] It should be noted that the entropy weighting method in this embodiment is used to objectively evaluate and determine the relative importance (i.e., weight) of each planning indicator. The evaluation process includes preprocessing the indicator data, calculating the information entropy of each indicator to measure the effectiveness of the information it provides, and determining the weight of each indicator based on this information. The direct output of the entropy weighting method is a set of objective weights that quantitatively reflect the contribution of each indicator in distinguishing different planning schemes. Subsequently, the standardized performance value of each planning scheme on each indicator is multiplied by its corresponding indicator weight and summed to obtain a final comprehensive score for each planning scheme. This comprehensive score is the final evaluation result of the planning scheme. The high or low score can clearly distinguish the overall advantages and disadvantages of different planning schemes. By further analyzing the contribution of each dimension to the score, planners can identify the strengths and weaknesses of specific schemes in specific scenarios. This comprehensive evaluation method based on objective weighting improves the scientific and targeted nature of the evaluation, providing strong support for accurate decision-making in distribution network planning.

[0050] Specifically, the planning evaluation indicators are classified and preprocessed, and the planning evaluation indicators are divided into benefit indicators, cost indicators and intermediate indicators, and the cost indicators and intermediate indicators are preprocessed; Benefit-based indicators are those whose larger values indicate better performance or benefits of the evaluation object. For example, a higher "power supply reliability" indicates better power supply service quality; a higher "renewable energy utilization rate" indicates greater utilization of renewable energy.

[0051] Cost-based indicators are those whose smaller values indicate better performance or lower costs. For example, a lower "line loss rate" indicates higher energy transmission efficiency, and a lower "total investment cost" indicates a more economical solution.

[0052] Intermediate indicators (or moderate indicators) are those whose values do not monotonically improve but rather have an optimal range or interval. Values that are too high or too low are not ideal. For these indicators, the optimal value or range can be set based on specific engineering experience, industry recommendations, or specific planning objectives.

[0053] According to the definitions of the above indicator types, each planning evaluation indicator is matched and the division of planning evaluation indicators is completed.

[0054] In an optional implementation, data preprocessing is performed on the cost-type indicators and the intermediate-type indicators to facilitate effective and reasonable comparison and analysis of the indicators; For benefit-type indicators, take their original values, and for cost-type indicators, take their reciprocals, expressed as: , For interval-type indicators , let it satisfy the following: , where is a permitted lower limit or minimum value of the indicator , and is a permitted upper limit or maximum value of the indicator . At the same time, different indicators have different dimensions, which will have a significant impact on the analysis of indicator data. It is necessary to standardize and dimensionless the indicator data to obtain normalized data. The normalization formula is expressed as: , It should be noted that by using the entropy weight method and introducing a probability matrix, the probability matrix can reflect the distribution of each indicator data value. Furthermore, the calculated information entropy value can describe the amount of information contained in each indicator. The smaller the information entropy value, the more concentrated the distribution of the indicator data, and the greater the impact on the overall goal. Therefore, the weight of this indicator is also greater. Based on the obtained information entropy value, the redundancy and indicator weight of each level of indicators can be further calculated.

[0055] Its evaluation process mainly includes: Calculate the information entropy. Based on the standardized decision matrix, calculate the proportion of the performance values of each planning scheme under each indicator, and further calculate the information entropy value of each indicator.

[0056] Determine the weight based on the information entropy calculation. According to the information entropy value, calculate the information utility value (or difference degree) of each indicator. The greater the information utility value, the greater the contribution of the indicator to distinguishing different planning schemes. Finally, normalize the information utility values of each indicator to obtain the objective weight coefficient W of each indicator in this specific scenario j .

[0057] Using the index weight W scientifically determined by the entropy weight method above j , conduct a final comprehensive effectiveness evaluation of each planning scheme. Specifically, multiply the standardized performance value Z of each planning scheme i on each (preprocessed and dimensionless) indicator j ij by its corresponding index weight W j , and then sum the weighted performance values of all indicators to obtain the final comprehensive score S of each planning scheme i i .

[0058] This comprehensive score S i is the final evaluation result of the planning scheme, which can quantitatively reflect the overall advantages and disadvantages of each planning scheme in a specific scenario, so as to realize the effective ranking and selection of different schemes.

[0059] The present invention adopts a multi-stage and multi-dimensional comprehensive evaluation method to evaluate the county-level distribution network planning scheme. This method not only considers the complexity and diversity of the operation of the distribution network, but also integrates advanced data analysis technologies and machine learning algorithms. Based on the in-depth analysis of the operation characteristics of the county-level distribution network integrating various flexibility resources, combined with the energy endowment, flexible resource allocation, digitalization degree and other characteristics of different county-level distribution networks, the spatio-temporal resource integration planning scenarios are divided; for different planning scenarios, according to the distribution system planning evaluation index system and based on expert experience, the planning evaluation indexes of different scenarios are screened; on the basis of the scenario-based multi-dimensional planning evaluation indexes, a data-driven method is adopted, taking into account the importance and relevance between the planning scheme and the indexes, and game theory is used for comprehensive ranking; a data-driven multi-dimensional planning evaluation index system is constructed, and the entropy weight method is used for evaluation, finally realizing the scenario-based, differential and flexible evaluation of the county-level distribution network planning.

[0060] The above is a schematic solution of a method for flexibly constructing and evaluating multi-dimensional planning indexes of a county-level distribution network in this embodiment. It should be noted that the technical solution of the system for flexibly constructing and evaluating multi-dimensional planning indexes of the county-level distribution network belongs to the same concept as the technical solution of the above method for flexibly constructing and evaluating multi-dimensional planning indexes of the county-level distribution network. For the details not described in detail in the technical solution of the system for flexibly constructing and evaluating multi-dimensional planning indexes of the county-level distribution network in this embodiment, reference can be made to the description of the technical solution of the above method for flexibly constructing and evaluating multi-dimensional planning indexes of the county-level distribution network.

[0061] A system for flexibly constructing and evaluating multi-dimensional planning indexes of a county-level distribution network in this embodiment includes: A division module, configured to analyze the operation characteristics of the county-level distribution network integrating various flexibility resources, and combine the energy endowment, flexible resource allocation, digitalization degree characteristics of different county-level distribution networks to divide the spatio-temporal resource integration county-level distribution network planning scenarios; A first construction module, configured to construct multi-dimensional planning evaluation indexes for different scenarios according to the complete county-level distribution network planning evaluation index system based on the division results of different planning scenarios; A second construction module, configured to adopt a data-driven method based on the multi-dimensional planning evaluation indexes of different scenarios, take into account the importance and relevance between the planning scheme and the indexes, and use game theory to comprehensively rank the indexes to construct data-driven multi-dimensional planning evaluation indexes; An evaluation module, configured to evaluate the planning scheme by the entropy weight method based on the data-driven multi-dimensional planning evaluation indexes.

[0062] This embodiment also provides a computer device applicable to the situation of flexibly constructing and evaluating multi-dimensional planning indexes of a county-level distribution network, including: A memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the method for flexibly constructing and evaluating multi-dimensional planning indicators of a county-level distribution network as proposed in the above embodiments.

[0063] This embodiment also provides a storage medium, on which a computer program is stored. When the program is executed by a processor, it implements the method for flexibly constructing and evaluating multi-dimensional planning indicators of a county-level distribution network as proposed in the above embodiments.

[0064] The storage medium proposed in this embodiment and the method for flexibly constructing and evaluating multi-dimensional planning indicators of a county-level distribution network proposed in the above embodiments belong to the same inventive concept. Technical details not described in detail in this embodiment can be referred to the above embodiments, and this embodiment has the same beneficial effects as the above embodiments.

[0065] From the above description of the embodiments, those skilled in the art can clearly understand that the present invention can be implemented by means of software and necessary general-purpose hardware. Of course, it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as a floppy disk, read-only memory (ROM), random access memory (RAM), flash memory (FLASH), hard disk, or optical disc of a computer, etc., including several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods of various embodiments of the present invention.

[0066] Embodiment 2, referring to Figures 2 - 11 As well as Tables 5 - 11, this is an embodiment of the present invention, which provides a method for flexibly constructing and evaluating multi-dimensional planning indicators of a county-level distribution network. In order to verify its beneficial effects, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0067] To verify the effectiveness and practicality of the proposed method for screening distribution network performance indicators, this embodiment selects the actual distribution network data of Object 1 for case analysis. This area has typical urban-rural mixed characteristics, including highly urbanized areas and rural areas, with significant differences in new energy penetration rates, which is an ideal scenario for verifying the research method of this study. The following is the basis for selecting Object 1.

[0068] (1) The characteristics of resource endowment are prominent. For the PV power generation project below 110 kV of Object 1, the total installed capacity is only 0.038 MW, the power supply area is 148 square kilometers, and the power supply load reaches 1.323 million kW. It is a typical area where the electricity demand is greater than the local supply capacity. As a new urban area, it mainly relies on the superior power grid for power supply, and there are very few local power generation facilities. Although the wind energy resources in this area are rich (the annual utilization hours reach more than 2,200 hours), the current scale of new energy development is still limited.

[0069] Table 5: Data Sheet of Object 1 , (2) There are few flexible resources. The load is mainly commercial load and smart home, and the construction of electric vehicle charging facilities has not yet formed a scale, and the energy storage resources have not been put on the agenda.

[0070] (3) The digital and intelligent development is rapid. Through digital and intelligent construction, the comprehensive real-time monitoring of the distribution network has been realized, and the self-healing has been fully covered, greatly improving the operation reliability of the distribution network.

[0071] Based on the above analysis, the planning scheme of Object 1 is an energy input type, with flexible configuration and high digital and intelligent degree, and the corresponding planning scenario is Scenario 2.

[0072] The planning scenario of Object 1 is Scenario 2. According to expert experience, a total of 41 indicators have been selected, covering five aspects: clean and low-carbon, economic and efficient, flexible and intelligent, safe and abundant, and supply-demand coordination. These indicators comprehensively reflect the characteristics of Scenario 2.

[0073] Table 6: Data Sheet of Evaluation Indicators of Object 1 Based on Scenario 2 , , , Table 7: Target Variables of Scenario 2 , On the basis of the above construction of indicators based on Scenario 2, a data-driven method is used for dimensionality reduction of indicators; Among them, a feature evaluation system is constructed based on the mRMR algorithm theory: For the indicator set and the target variable c (initial score of the scheme), calculate the mutual information between each feature and the target variable , which is expressed as: , Among them, represents the joint probability distribution, and Represents the marginal probability distribution; Calculate the metric The mutual information between , evaluate the redundancy between features, expressed as: , Perform the MIQ criterion to calculate the correlation between features and the target variable , expressed as: , Perform the redundancy measurement calculation, calculate the average mutual information between the feature and the features in the selected feature set S , expressed as: , where, Represents the size of the selected feature set; In each iteration, select the feature with the maximum MIQ score, expressed as: , And add the selected feature to the selected feature set, expressed as: , Finally, obtain the mRMR importance score : , respectively represent the total number of metrics for the effective sorting of mRMR, and is the position of the metric in the mRMR sorted list. Repeat the above steps until the preset number of features is reached or the MIQ score is lower than the set threshold. Finally, the final preferred feature subset will be selected. Among them, the preset number of features and the set threshold can be set according to the actual application scenario requirements or expert experience. Select the features with the importance score ranked in the top 20 for display, and at the same time, use a horizontal bar chart to display the feature importance scores.

[0074] Arrange in descending order according to the importance score to obtain, such as Figure 2 The object of the flexible construction method of multi-dimensional planning indicators for the county-level distribution network based on digital means - the mRMR screening feature importance ranking chart based on scenario two.

[0075] Among them, based on the random forest algorithm, it can be obtained that: Construct a random forest regression model with optimized training process; Feature importance analysis Importance calculation: Calculate feature importance based on the splitting gain of decision trees, and adopt an ensemble method to synthesize the evaluation results of multiple trees. Standardize the feature importance scores for easy result comparison Visualization: Select the top 20 features ranked by importance for key display Add specific numerical labels to provide precise importance information, and we can get Figure 3 Figure of the importance ranking of features selected by random forest for Object 1 based on Scenario 2.

[0076] Among them, the calculation based on game theory shows that: mRMR feature selection; Random forest model evaluation; Shapley value calculation; Feature fusion and normalization, respectively normalize the scores of the three algorithms, construct a unified feature scoring data frame to ensure the comparability of scores; Multi-algorithm fusion strategy; Construct a comprehensive scoring system, select the top 20 indicators to generate the final feature importance ranking result, such as Figure 4 Figure of the importance ranking of features selected by game theory for Object 1 based on Scenario 2.

[0077] Among them, the calculation based on the entropy weight method gives Figure 5 Scores obtained from the evaluation of the planning scheme for Object 1 based on Scenario 2: Among them, the score of Planning Scheme 1-1 in Scenario 2 is 0.8275, the score of Planning Scheme 1-2 is 0.9079, the score of Planning Scheme 1-3 is 0.9504, and the score of Planning Scheme 1-4 is 0.8946.

[0078] To verify the effectiveness and practicality of the proposed screening method for distribution network performance indicators, this embodiment selects the actual distribution network data of Object 2 for case analysis; For Object 2 selected in this example, the following are the basis for selecting the planning scenarios: (1) It has significant resource endowment advantages. Object 2 makes full use of its unique wind and light resources to build an energy pattern of "wind, light, water, fire, and hydrogen energy" multi-energy complementarity. In terms of energy supply, relying on the local rich coal resources, it has built multiple large-scale thermal power plants with sufficient installed capacity to ensure the supply of basic loads. At the same time, it has built multiple photovoltaic power generation bases, making full use of the barren mountain and slope resources, and layout and construction of centralized wind farms using the local high-quality wind resources, and a number of small hydropower stations have been built relying on the river system to provide regulation capacity for regional power supply. In recent years, it has actively explored the development of the hydrogen energy industry, layout of hydrogen production, hydrogen storage and hydrogen energy application facilities, further enriching the form of energy supply. The total installed capacity reaches 2855MW, the power supply area is 4056.7 square kilometers, the maximum power supply load is 661,200 kilowatts, among which the energy density reaches 0.7037MW / km 2 , and the load density reaches 0.162 MW / km 2, is a typical output-oriented region. At present, a diversified energy supply system based on coal power and with rapid development of new energy has been formed, which not only ensures the safety and reliability of power supply, but also promotes the clean and low-carbon transformation of the energy structure.

[0079] Table 8: Data Sheet of Object Two , (2) There are few flexible resources. The load is mainly for traditional industrial electricity consumption. The construction of electric vehicle charging facilities has not yet formed a scale, and energy storage resources have not been put on the agenda.

[0080] (3) The digital and intelligent construction is steadily advancing. The comprehensive real-time monitoring of the distribution network has been achieved, and the distribution automation technology route has been differentially deployed, effectively improving the operation reliability and power supply quality of the distribution network.

[0081] Based on the above analysis, the planning scheme for Object Two is energy output-oriented, with flexible configuration and high digital and intelligent level. The corresponding planning scenario is Scenario Six.

[0082] The planning scenario for Object Two is Scenario Six. According to expert experience, a total of 25 indicators have been selected, covering five aspects: clean and low-carbon, economic and efficient, flexible and intelligent, safe and abundant, and supply-demand coordination. These indicators comprehensively reflect the characteristics of Scenario Six.

[0083] Table 9: Data Sheet of Evaluation Indicators for Object Two Based on Scenario Six , , Table 10: Target Variables of Scenario Six , Based on the above indicator construction and data tables, a data-driven method is used for dimensionality reduction of indicators; Among them, a feature evaluation system is constructed based on the mRMR algorithm theory: Rank in descending order according to the importance score Select the top 20 features with the highest importance for display, and at the same time use a horizontal bar chart to display the importance scores of the features. As a result, Figure 6 Importance Ranking Diagram of mRMR-Screened Features for Object Two Based on Scenario Six.

[0084] Among them, based on the random forest algorithm, it can be obtained that: Construct a random forest regression model and optimize the training process; Feature importance analysis and importance calculation: Calculate the feature importance based on the splitting gain of the decision tree, adopt an integrated method to comprehensively evaluate the results of multiple trees, and standardize the feature importance scores for easy comparison and result visualization: Select the top 20 features with the highest importance for key display and add specific numerical labels to provide accurate importance information. As a result, Figure 7The importance ranking graph of features screened by the random forest of Object 2 based on Scenario 6.

[0085] Among them, the calculation based on game theory shows that: mRMR feature selection; random forest model evaluation; Shapley value calculation; feature fusion and normalization, respectively normalize the scores of the three algorithms, construct a unified feature scoring data frame to ensure the comparability of scores; multi-algorithm fusion strategy; construct a comprehensive scoring system, select the top 20 indicators to generate the final feature importance ranking result, such as Figure 8 The importance ranking graph of features screened by the game theory of Object 2 based on Scenario 6.

[0086] Among them, the calculation based on the entropy weight method gives Figure 9 The scores obtained by Object 2 based on the entropy weight method calculation of Scenario 6: among them, the score of Planning Scheme 2-1 is 0.8628, the score of Planning Scheme 2-2 is 0.893, the score of Planning Scheme 2-3 is 0.8887, and the score of Planning Scheme 2-4 is 0.9295.

[0087] After screening based on the algorithm, the optimal planning scheme of Object 1 based on Scenario 2 is 2-3, with a score of 0.9504; the optimal planning scheme of Object 2 based on Scenario 6 is 6-4, with a score of 0.9295, as shown in the final index table 9.

[0088] Table 11: Final Index Construction Table , It should be noted that through the in-depth analysis of the distribution network planning in two different counties of Object 1 and Object 2, it shows that the method of the present invention can realize the flexible construction and evaluation of the scenario-based and differentiated indicators of the distribution network planning in different counties.

[0089] In summary, the present invention proposes a flexible construction of multi-dimensional planning indicators for the distribution network in counties by means of digitalization. On the basis of in-depth analysis of the characteristics of the distribution network in counties integrating various flexibility resources, combined with the characteristics of energy endowment, flexible resource allocation, digitalization degree, etc. of the distribution network in different counties, the spatio-temporal resource integration planning scenarios are divided; for different planning scenarios, according to the distribution system planning evaluation index system and based on expert experience, the planning evaluation indicators for different scenarios are screened; on the basis of the multi-dimensional planning evaluation indicators based on scenarios, a data-driven method is adopted, taking into account the importance and relevance between the planning scheme and the indicators, and game theory is used for comprehensive ranking; a multi-dimensional planning evaluation index system based on data driving is constructed, and the entropy weight method is used for evaluation, and finally the analysis of the scenario-based, differential and flexible planning of the distribution network in counties is realized.

[0090] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered within the scope of the claims of the present invention.

Claims

1. A flexible construction and evaluation method for multi-dimensional planning indicators of a county-level distribution network, characterized in that Including: Analyze and integrate the operating characteristics of the county-level distribution network with various flexibility resources, and divide the planning scenarios of the county-level distribution network with spatio-temporal resource integration by combining the energy endowment, flexible resource allocation, and digitalization degree characteristics of different county-level distribution networks. Based on the results of different planning scenario divisions, construct multi-dimensional planning evaluation indicators for different scenarios according to the complete index system for the evaluation of county-level distribution network planning. Based on the multi-dimensional planning evaluation indicators for different scenarios, adopt a data-driven method, take into account the importance and correlation between the planning scheme and the indicators, and comprehensively sort the indicators using game theory to construct data-driven multi-dimensional planning evaluation indicators. Based on the data-driven multi-dimensional planning evaluation indicators, evaluate the planning scheme by the entropy weight method.

2. The multi-dimensional planning index flexible construction and evaluation method for a county-level distribution network according to claim 1, characterized in that Dividing the planning scenarios of the county-level distribution network with spatio-temporal resource integration includes: The division of the planning scenarios of the county-level distribution network with spatio-temporal resource integration includes dividing the energy endowment, flexible resource allocation, and digitalization degree of different power supply areas or different development periods of power supply areas. The energy endowment classifies energy resources into energy input type and energy output type. The energy input type means that the installed capacity of centralized and distributed energy in the county-level distribution network planning scheme is small, and the energy output type means that the installed capacity of centralized and distributed energy in the county-level distribution network planning scheme is large. The flexible resource allocation represents the installed capacity of flexibility resources in the county-level distribution network planning scheme. The digitalization degree represents the degree of digital configuration in the county-level distribution network planning scheme.

3. The flexible construction and evaluation method for multi-dimensional planning indicators of a county-level distribution network according to claim 2, characterized in that Constructing multi-dimensional planning evaluation indicators for different scenarios includes: The complete index system for the evaluation of county-level distribution network planning includes multiple first-level indicators, second-level indicators, and third-level indicators in aspects such as clean and low-carbon, safe and abundant, economic and efficient, flexible and intelligent, and supply-demand coordination. According to the complete index system for the evaluation of county-level distribution network planning, combined with the characteristic data of different planning scenarios, differentially construct the planning evaluation indicators for different scenarios, and screen out the planning evaluation indicators for different scenarios from the planning evaluation index system.

4. A flexible construction and evaluation method for multi-dimensional planning indicators of a county-level distribution network according to claim 1, characterized in that, Constructing data-driven multi-dimensional planning evaluation indicators includes: Adopt a data-driven method to preliminarily screen the potential indicators for each scenario through the maximum correlation and minimum redundancy algorithm, maintain the maximum correlation between the indicators and the target variable, and minimize the redundant information between the indicators. Evaluate the importance of the screened indicators through the random forest algorithm to handle the non-linear relationship between the indicators. Through the game theory optimization model, combine the correlation and importance, sort and screen the planning indicators, and allocate and dynamically adjust the weights of the indicators.

5. The multi-dimensional planning index flexible construction and evaluation method for a county-level distribution network according to claim 4, characterized in that Preliminarily screening the potential indicators for each scenario through the maximum correlation and minimum redundancy algorithm, maintaining the maximum correlation between the indicators and the target variable, and minimizing the redundant information between the indicators includes: For the index input data set and the target variable c, perform data standardization and missing value processing; Calculate each feature with the target variable the mutual information between , expressed as: ， Among them, represents the joint probability distribution, and represents the marginal probability distribution; Calculate the mutual information between features , evaluate the redundancy between features, expressed as: ， Calculate the correlation between the features calculated by the MIQ criterion and the target variable , expressed as: ， Perform redundancy metric calculation to calculate the average mutual information between the feature and the features in the selected feature set S, which is expressed as: , expressed as: ， Among them, represents the size of the selected feature set; In each iteration, the feature with the maximum MIQ score is selected, and the selected feature is added to the set of selected features to obtain the mRMR importance score , denoted as: ， Among them, respectively represent the total number of indicators for effective sorting of mRMR, which are indicators in the position of the mRMR sorted list. Repeat the above steps until the preset number of features is reached or the MIQ score is lower than the set threshold.

6. The multi-dimensional planning index flexible construction and evaluation method for a county-level distribution network according to claim 4, wherein Evaluating the importance of the screened indicators through the random forest algorithm to handle the non-linear relationship between the indicators includes: Perform Bootstrap sampling on the training set, randomly generate k sample subsets Si, and simultaneously generate corresponding decision trees Ti. When a node splits, randomly select m features from m features as a candidate feature set; try features as a candidate feature set; Calculate the optimal splitting point of candidate features for each node. Calculate the mean squared error of the parent node using different splitting thresholds, and at the same time calculate the mean squared errors of the left and right child nodes. The splitting gain is calculated as follows: ， Select the feature with the largest splitting gain and the corresponding splitting point. Select the feature with the largest reduction in mean squared error for splitting, and then calculate the importance of a single feature in a single tree, which is expressed as: ， Calculate the average importance of the feature in the entire forest, which is expressed as: ， Obtain the importance of the feature through normalization processing and output the sorted feature importance.

7. The multi-dimensional planning index flexible construction and evaluation method for a county-level distribution network according to claim 6, characterized in that Evaluate the planning scheme by the entropy weight method, including: Set the index types, which include benefit-type indicators, cost-type indicators, and intermediate-type indicators. Divide the planning evaluation indicators according to the definitions of each index type, and perform data preprocessing on the cost-type indicators and intermediate-type indicators; Introduce a probability matrix by the entropy weight method, calculate the information entropy value of the information contained in each indicator, and calculate the redundancy and indicator weights of each level of indicators based on the information entropy value.

8. A flexible construction and evaluation system for multi-dimensional planning indicators of a county-level distribution network, which applies a flexible construction and evaluation method for multi-dimensional planning indicators of a county-level distribution network according to any one of claims 1 to 7, characterized in that, Include: A partitioning module for analyzing and integrating the operating characteristics of the county-level distribution network with multiple flexibility resources, and combining the energy endowment, flexible resource allocation, and digitalization degree characteristics of different county-level distribution networks to perform spatio-temporal resource integration and county-level distribution network planning scenario partitioning; A first construction module for constructing multi-dimensional planning evaluation indicators for different scenarios based on the partitioning results of different planning scenarios and according to the complete county-level distribution network planning evaluation index system. The complete county-level distribution network planning evaluation index system includes multiple first-level indicators, second-level indicators, and third-level indicators in multiple aspects such as clean and low-carbon, safe and abundant, economic and efficient, flexible and intelligent, and supply-demand coordination; A second construction module for constructing data-driven multi-dimensional planning evaluation indicators based on the multi-dimensional planning evaluation indicators of different scenarios. Using a data-driven method, taking into account the importance and correlation between the planning scheme and the indicators, and comprehensively sorting the indicators using game theory; An evaluation module for evaluating the planning scheme by the entropy weight method based on the data-driven multi-dimensional planning evaluation indicators.

9. A computer device, characterized in that, Include: A memory and a processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of a method for flexibly constructing and evaluating multi-dimensional planning indicators for a county-level distribution network according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that, It stores computer-executable instructions, and when the computer-executable instructions are executed by the processor, the steps of a method for flexibly constructing and evaluating multi-dimensional planning indicators for a county-level distribution network according to any one of claims 1 to 7 are implemented.

Images