Novel power distribution network comprehensive evaluation method based on cross validation method

Abstract

The invention discloses a novel power distribution network comprehensive evaluation method based on a cross validation method, and belongs to the technical field of power distribution management systems, and the method comprises the steps: S1, obtaining the initial weight of an evaluation index in a hierarchical structure model based on an analytic hierarchy process; s2, acquiring a plurality of groups of combined weights based on the initial weight; s3, based on a first historical data set of the novel power distribution network, screening out an optimal combination weight from the plurality of groups of combination weights by using a cross validation method, and updating the hierarchical structure model based on the optimal combination weight to obtain an optimal hierarchical structure model; and S4, carrying out stability verification on the optimal hierarchical structure model based on a second historical data set of the novel power distribution network, if the stability verification is passed, carrying out comprehensive evaluation based on the optimal hierarchical structure model, and if the stability verification is not passed, taking the optimal combination weight as the initial weight, and executing S2. The problem that the accurate weight of each evaluation index is difficult to obtain is solved, and the comprehensive evaluation efficiency and the accuracy of the comprehensive evaluation result are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of distribution management systems, and particularly to a new comprehensive evaluation method for distribution networks based on the cross-validation method. Background Art

[0002] With the continuous progress and expansion of the power system, reasonable planning of the distribution network is of utmost importance. In the traditional distribution network planning process, only reliability and economy are taken as the planning criteria, and the various capabilities of the new distribution network are not comprehensively evaluated. Existing technologies such as the comprehensive evaluation method and system for the AC-DC hybrid low-voltage distribution network grid scheme with the publication number CN117578472A include collecting comprehensive evaluation index data and establishing a comprehensive evaluation index system for the low-voltage distribution network grid scheme; based on the analytic hierarchy process, establishing a comprehensive evaluation system for the AC-DC hybrid low-voltage distribution network; calculating the comprehensive evaluation value through the comprehensive evaluation system to optimize the distribution network grid scheme. The above scheme establishes a comprehensive evaluation system for the AC-DC hybrid low-voltage distribution network through the analytic hierarchy process, quantifies and compares different low-voltage distribution network grid schemes, and improves the comprehensiveness and reliability of the evaluation results. However, when calculating the comprehensive evaluation value through the comprehensive evaluation system established by the analytic hierarchy process in the above scheme, since the weights of the evaluation indexes in the comprehensive evaluation system are obtained through expert evaluation in the analytic hierarchy process, there are subjective errors in the weights of the evaluation indexes. Summary of the Invention

[0003] Aiming at the problem that it is difficult to obtain accurate weights for each evaluation index in the prior art, the present invention provides a new comprehensive evaluation method for distribution networks based on the cross-validation method. Taking the initial weights obtained by the analytic hierarchy process as a benchmark, several groups of combined weights are obtained, reducing the blindness of obtaining combined weights by the cross-validation method. Then, through the obtained several groups of combined weights, the cross-validation method is used to accurately select the combined weights from several groups of combined weights, and then the accurate weights of each evaluation index are obtained, solving the problem of difficult to obtain accurate weights for each evaluation index, and significantly improving the comprehensive evaluation efficiency and the accuracy of the comprehensive evaluation results.

[0004] To solve the above technical problems, the present invention provides a new comprehensive evaluation method for distribution networks based on the cross-validation method, including the following steps: S1: Construct a hierarchical structure model considering low-carbon constraints based on the analytic hierarchy process, and obtain the initial weights of the evaluation indexes in the hierarchical structure model; S2: Set weight combination criteria based on the initial weights, and obtain several groups of combined weights based on the weight combination criteria; S3: Based on the first historical data set of the new distribution network, use the cross-validation method to screen out the optimal combined weights from several groups of combined weights, and update the hierarchical structure model based on the optimal combined weights to obtain the optimal hierarchical structure model; S4: Validate the stability of the optimal hierarchical structure model based on the second historical dataset of the new distribution network. If the stability validation passes, conduct a comprehensive evaluation based on the optimal hierarchical structure model. If the stability validation fails, use the optimal combined weights as the initial weights and execute S2.

[0005] After adopting the above technical solution, the present invention has the following advantages: Considering that the initial weights of the evaluation indicators obtained by the analytic hierarchy process are close to the accurate weights of the evaluation indicators and there are subjective errors, while the cross-validation method can eliminate subjective errors through multiple validations, but there is blindness when using the cross-validation method to obtain combined weights, and it is difficult to obtain combined weights close to the accurate weights in a short time. Therefore, by combining the analytic hierarchy process and the cross-validation method for comprehensive evaluation of the new distribution network, the efficiency of comprehensive evaluation and the accuracy of the comprehensive evaluation results are significantly improved; by validating the stability of the obtained optimal hierarchical structure model and performing corresponding steps according to the stability validation results, the stability of the optimal hierarchical structure model is further improved; It solves the problem of difficult to obtain the accurate weights of each evaluation indicator.

[0006] Preferably, in S1, the obtaining of the initial weights of the evaluation indicators in the hierarchical structure model includes: S11: Construct a judgment matrix of the hierarchical structure model based on expert evaluation in the analytic hierarchy process; S12: Conduct normalization processing and consistency test on the judgment matrix, and obtain the initial weights of the evaluation indicators in the hierarchical structure model according to the consistency test results.

[0007] Preferably, in S12, the obtaining of the initial weights of the evaluation indicators in the hierarchical structure model according to the consistency test results includes: If the consistency test results meet the consistency requirements, the process weights obtained by the normalization processing are the initial weights. If the consistency requirements are not met, update the judgment matrix based on the mutual relationship between the data in the judgment matrix and execute S12. By conducting a consistency test on the judgment matrix, the rationality of the obtained initial weights is ensured.

[0008] Preferably, the S2 includes: S21: Take the initial weights as the benchmark, set a floating standard based on the comprehensive evaluation requirements, and obtain alternative weights of the evaluation indicators according to the floating standard; S22: Set the weight combination criterion according to the magnitude relationship between the initial weights and the principle of full coverage and uniqueness, and combine the alternative weights based on the weight combination criterion to obtain several groups of combined weights.

[0009] In this solution, a floating standard is set through comprehensive evaluation of requirements, which improves the flexibility of comprehensive evaluation. Taking the initial weight as a benchmark, alternative weights are obtained according to the floating standard, reducing the blindness of obtaining weights using the cross-validation method. Furthermore, alternative weights close to the accurate weight are obtained within a short time, improving the efficiency of comprehensive evaluation. The weight combination criterion is set based on the magnitude relationship between the initial weights and the principle of full coverage and uniqueness, and the combined weight is obtained according to the combination criterion, considering the rationality of the weights between different evaluation indicators and the integrity of the combined weight. Thus, the integrity and rationality of the combined weight are improved.

[0010] Preferably, the S3 includes: S31: Perform data cleaning, missing value filling, and outlier processing on the first historical dataset; S32: Train the hierarchical structure model based on several groups of combined weights and the first historical dataset and obtain the optimal combined weight.

[0011] Preferably, the S32 includes: S321: Input the first historical dataset and several groups of combined weights into the hierarchical structure model, calculate the first product of several groups of combined weights and the corresponding data in the first historical dataset, compare the first product with the actual result of the corresponding data in the first historical dataset, and obtain the number of successful comparisons; S322: Determine whether the number of successful comparisons of the combined weight with the highest number of successful comparisons is greater than the first preset number. If it is greater, the combined weight with the highest number of successful comparisons is the optimal combined weight. If it is less than or equal, use the combined weight with the highest number of successful comparisons as the initial weight and execute S2.

[0012] In this solution, the product of the first historical dataset and each combined weight is calculated respectively, providing a rich data basis for the training of the hierarchical structure model. Thus, the generalization ability and accuracy of the model are improved. By inputting several combined weights into the hierarchical structure model for calculation, and then taking the combined weight with the most comparison times and greater than the first preset number as the optimal combined weight, it is ensured that the selected weight shows high accuracy in multiple comparisons, improving the stability and reliability of the optimal hierarchical structure model.

[0013] Preferably, in S4, the stability verification of the optimal hierarchical structure model based on the second historical dataset of the new distribution network includes: Input the second historical dataset into the optimal hierarchical structure model, calculate the second product of the optimal combined weight and the corresponding data in the second historical dataset, match the second product with the actual result of the corresponding data in the second historical dataset. If the number of successful matches exceeds the second preset number, the stability verification passes; if the number of successful matches does not exceed the second preset number, the stability verification fails.

[0014] Preferably, in S1, the evaluation indicators at least include the distributed power consumption capacity, power supply capacity, power supply reliability, extreme disaster scenario recovery capacity, and the cost of carbon emission rights trading for the distribution network considering low-carbon constraints.

[0015] Preferably, the distributed power consumption capacity is characterized by the consumption rate, and the consumption rate where P DG is the power generation of the distributed power source, and P tol is the load demand of the distribution network.

[0016] Preferably, the power supply capacity is characterized by the voltage deviation, and the voltage deviation where V actual is the actual voltage of the distribution network node, and V c is the rated voltage of the distribution network.

[0017] Beneficial effects of this solution: Considering that the initial weights of the evaluation indicators obtained by the analytic hierarchy process are close to the accurate weights of the evaluation indicators and there are subjective errors, while the cross-validation method can eliminate subjective errors through multiple verifications, but there is blindness when using the cross-validation method to obtain the combined weights, and it is difficult to obtain the combined weights close to the accurate weights in a short time. Therefore, the comprehensive evaluation of the new distribution network is carried out by combining the analytic hierarchy process and the cross-validation method, which significantly improves the efficiency of the comprehensive evaluation and the accuracy of the comprehensive evaluation results; the weight combination criteria are set according to the magnitude relationship between the initial weights and the principle of full coverage and uniqueness, and the combined weights are obtained according to the combination criteria, considering the rationality of the weights between different evaluation indicators and the integrity of the combined weights, thereby improving the integrity and rationality of the combined weights; The products of the first historical data set and each combined weight are calculated respectively, providing a rich data basis for the training of the hierarchical structure model, thereby improving the generalization ability and accuracy of the model. By inputting several combined weights into the hierarchical structure model for calculation, and then taking the combined weight with the most comparison times and greater than the first preset number of times as the optimal combined weight, ensuring that the selected weights show high accuracy in multiple comparisons, improving the stability and reliability of the optimal hierarchical structure model. By performing stability verification on the obtained optimal hierarchical structure model and executing corresponding steps according to the stability verification results, the stability of the optimal hierarchical structure model is further improved; The problem of difficult to obtain the accurate weights of each evaluation indicator is solved. Description of the Drawings

[0018] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description of non - restrictive embodiments read in conjunction with the accompanying drawings. The drawings are only for the purpose of showing the preferred embodiments and are not considered as limiting the present invention. Also, throughout the drawings, the same reference signs are used to represent the same components.

[0019] Figure 1 This is a flowchart of a novel comprehensive evaluation method for a distribution network based on the cross - validation method of the present invention. Detailed implementation manners

[0020] To make the objectives, technical solutions, and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below in conjunction with the drawings and embodiments. It should be understood that the specific implementation manners described herein are only the best embodiments of the present invention, which are only used to explain the present invention and do not limit the protection scope of the present invention. All other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the protection scope of the present invention.

[0021] Before discussing the exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although the flowcharts describe the operations (or steps) as sequential processes, many of the operations (or steps) can be implemented in parallel, concurrently, or simultaneously. In addition, the order of the operations can be rearranged. The process can be terminated when its operations are completed, but it can also have additional steps not included in the drawings; the process can correspond to a method, function, procedure, subroutine, sub - program, and so on.

[0022] Embodiment 1: As Figure 1 shown, the novel comprehensive evaluation method for a distribution network based on the cross - validation method includes the following steps: S1: Construct a hierarchical structure model considering low - carbon constraints based on the analytic hierarchy process, and obtain the initial weights of the evaluation indicators in the hierarchical structure model.

[0023] In S1, the obtaining of the initial weights of the evaluation indicators in the hierarchical structure model includes: S11: Construct a judgment matrix of the hierarchical structure model based on expert evaluation in the analytic hierarchy process; S12: Perform normalization processing and consistency test on the judgment matrix, and obtain the initial weights of the evaluation indicators in the hierarchical structure model according to the consistency test results.

[0024] In S12, obtaining the initial weights of the evaluation indicators in the hierarchical model according to the consistency test results includes: if the consistency test result meets the consistency requirement, the process weights obtained by the normalization process are the initial weights; if it does not meet the consistency requirement, the judgment matrix is updated based on the mutual relationship between the data in the judgment matrix, and S12 is executed.

[0025] In S1, the evaluation indicators at least include the distributed power generation consumption capacity, power supply capacity, power supply reliability, extreme disaster scenario recovery capacity, and the cost of carbon emission trading in the distribution network considering low-carbon constraints.

[0026] The distributed power generation consumption capacity is characterized by the consumption rate, and the consumption rate where P DG is the power generation of the distributed power source, and P tol is the load demand of the distribution network.

[0027] The power supply capacity is characterized by the voltage deviation, and the voltage deviation where V actual is the actual voltage of the distribution network node, and V c is the rated voltage of the distribution network.

[0028] In this embodiment, the hierarchical model is specifically: with the evaluation indicators as the index layer and the highest comprehensive score of the new distribution network as the target layer, each evaluation indicator has a weight vector pointing to the highest comprehensive score of the new distribution network. The specific use of the hierarchical model is as follows: the data corresponding to the evaluation indicators are output into the hierarchical model, the product of the data corresponding to the evaluation indicators and the weights of the evaluation indicators in the hierarchical model is calculated, and the new distribution network is comprehensively evaluated according to the product.

[0029] In this embodiment, when the evaluation indicators only include the distributed power generation consumption capacity, power supply capacity, power supply reliability, extreme disaster scenario recovery capacity, and the cost of carbon emission trading in the distribution network considering low-carbon constraints, the judgment matrix is of the 5*5 order, and the consistency requirement is: when the consistency test result is less than 0.1, it meets the consistency requirement; if it is greater than or equal to 0.1, it does not meet the consistency requirement. When it does not meet the consistency requirement, the judgment matrix is adjusted. Updating the judgment matrix according to the mutual relationship between the data in the judgment matrix is specifically: checking the mutual relationship between the data in the judgment matrix. If there is a situation where the weight of evaluation indicator A is greater than the weight of evaluation indicator B, the weight of evaluation indicator B is greater than the weight of evaluation indicator C, but the weight of evaluation indicator C is greater than the weight of evaluation indicator A, then the weight of evaluation indicator C is adjusted to ensure the rationality and scientificity of the weights between the evaluation indicators.

[0030] As an implementation manner, the order of the judgment matrix is adjusted accordingly according to the number of evaluation indicators. If the number of evaluation indicators increases, the order of the judgment matrix is increased accordingly; if it decreases, the order of the judgment matrix is decreased accordingly, so that the judgment matrix is applicable to different numbers of evaluation indicators, thereby improving the flexibility and applicability of the judgment matrix.

[0031] In this embodiment, the power supply reliability is characterized by the reliability margin index, and the reliability margin index where P b is the spare capacity of the distribution network, and P cl is the critical load demand. The recovery ability in the extreme disaster scenario is characterized by the recovery rate, and the recovery rate where L res (t) represents the amount of load restored within time t, represents the total load of the affected area before the disaster. The carbon emission trading cost C of the distribution network considering low-carbon constraints = C c ×(E a -E c ), where C c is the unit carbon emission trading price for the excess part of carbon emissions, E a is the actual carbon emission, and E c is the carbon emission quota allocated by the government to key emission enterprises at the beginning of the allocation period for a certain period.

[0032] S2: Based on the initial weights, set the weight combination criterion, and obtain several groups of combined weights based on the weight combination criterion.

[0033] The S2 includes: S21: Take the initial weights as the benchmark, set the floating standard based on the comprehensive evaluation requirements, and obtain the alternative weights of the evaluation indicators according to the floating standard; S22: Set the weight combination criterion according to the magnitude relationship between the initial weights and the principle of full coverage and uniqueness, and combine the alternative weights based on the weight combination criterion to obtain several groups of combined weights.

[0034] In this embodiment, setting the floating standard based on the comprehensive evaluation requirements is specifically as follows: If the focus is on improving the comprehensive evaluation efficiency, the floating standard is correspondingly reduced. That is, if the initial weights of the evaluation indicators are 0.1, 0.2, and 0.3 respectively, the floating standard can be set to 0.01. At this time, the alternative weights of the corresponding evaluation indicators are in the range of 0.1 - 0.01 to 0.1 + 0.01, 0.2 - 0.01 to 0.2 + 0.01, and 0.3 - 0.01 to 0.3 + 0.01. If the focus is on improving the comprehensive evaluation accuracy, the floating standard is correspondingly enlarged. If it is necessary to improve both efficiency and accuracy at the same time, the floating standard is set based on the initial weights according to historical data and experience. By setting the floating standard, blindly obtaining the combined weights using the cross-validation method is avoided, thereby improving the comprehensive evaluation efficiency and at the same time improving the flexibility of the comprehensive evaluation.

[0035] In this embodiment, setting the weight combination criterion according to the magnitude relationship between the initial weights and the principle of full coverage and uniqueness is specifically as follows: If the initial weight of evaluation indicator A is greater than the initial weight of evaluation indicator B, when obtaining the combined weights from the alternative weights, it is necessary to always ensure that the weight of evaluation indicator A is greater than the weight of evaluation indicator B. The principle of full coverage and uniqueness means that the weights in the combined weights cover each evaluation indicator and are unique. For example: If the alternative weights of evaluation indicator A are 0.1, 0.2, 0.3, the alternative weights of evaluation indicator B are 0.15, 0.23, 0.4, and the initial weight of evaluation indicator A is greater than the initial weight of evaluation indicator B, then the obtained several combined weights are (0.2, 0.15), (0.3, 0.15), (0.3, 0.23). By obtaining the combined weights through the weight combination criterion, the rationality of the weights between different evaluation indicators and the integrity of the combined weights are considered, thereby improving the integrity and rationality of the combined weights.

[0036] S3: Based on the first historical data set of the new distribution network, use the cross-validation method to screen out the optimal combined weights from several groups of combined weights, and update the hierarchical structure model based on the optimal combined weights to obtain the optimal hierarchical structure model.

[0037] The said S3 includes: S31: Perform data cleaning, missing value filling, and outlier processing on the first historical data set; S32: Train the hierarchical structure model based on several groups of combined weights and the first historical data set and obtain the optimal combined weights.

[0038] The said S32 includes: S321: Input the first historical data set and several groups of combined weights into the hierarchical structure model, calculate the first product of several groups of combined weights and the corresponding data in the first historical data set, compare the first product with the actual result of the corresponding data in the first historical data set, and obtain the number of successful comparisons; S322: Determine whether the comparison success count of the combination weight with the highest comparison success count is greater than a first preset count. If it is greater, then the combination weight with the highest comparison success count is the optimal combination weight. If it is less than or equal, then use the combination weight with the highest comparison success count as the initial weight and execute S2.

[0039] In this embodiment, the first preset count is set according to user requirements. If it is necessary to improve the accuracy of comprehensive evaluation, the first preset count is increased accordingly. The data in the first historical dataset is only data related to evaluation indicators, which avoids inputting hybrid data into the hierarchical result model and thus affecting the accuracy of the output result of the hierarchical structure model. At the same time, the data related to evaluation indicators in the first historical dataset involves different dimensions and different periods, providing a rich data basis for the training of the hierarchical structure model, thereby improving the generalization ability and accuracy of the model. By inputting several combination weights into the hierarchical structure model for calculation, the combination weight with the most comparison times and greater than the first preset count is used as the optimal combination weight, ensuring that the selected weight shows high accuracy in multiple comparisons and improving the stability and reliability of the optimal hierarchical structure model.

[0040] S4: Perform stability verification on the optimal hierarchical structure model based on the second historical dataset of the new distribution network. If the stability verification passes, perform comprehensive evaluation based on the optimal hierarchical structure model. If the stability verification fails, use the optimal combination weight as the initial weight and execute S2.

[0041] In S4, the performing stability verification on the optimal hierarchical structure model based on the second historical dataset of the new distribution network includes: Input the second historical dataset into the optimal hierarchical structure model, calculate the second product of the optimal combination weight and the corresponding data in the second historical dataset, and match the second product with the actual result of the corresponding data in the second historical dataset. If the number of successful matches exceeds a second preset count, the stability verification passes. If the number of successful matches does not exceed the second preset count, the stability verification fails.

[0042] In this embodiment, the second preset number of times is set according to user requirements. If it is necessary to improve the accuracy of comprehensive evaluation, the first preset number of times is increased accordingly. The data in the second historical dataset includes not only the data related to the evaluation index but also the data unrelated to the evaluation index, so as to test the ability of the optimal hierarchical structure model to screen the data related to the evaluation index. At the same time, the data in the second historical dataset involves different dimensions and different cycles, providing a rich data basis for the training of the hierarchical structure model, and then testing the generalization ability of the optimal hierarchical structure model. By inputting several combined weights and the second historical data into the optimal hierarchical structure model for calculation, and then taking the combined weight with the most comparison times and greater than the first preset number of times as the optimal combined weight, it ensures that the selected weight shows high accuracy in multiple comparisons. While comprehensively testing the optimal hierarchical structure model, it also improves the stability and reliability of the optimal hierarchical structure model.

[0043] The specific implementation manners described above are the preferred implementation manners of the novel distribution network comprehensive evaluation method based on the cross-validation method of the present invention, and do not limit the specific implementation scope of the present invention. The scope of the present invention includes but is not limited to this specific implementation manner. All equivalent changes made according to the shape and structure of the present invention are within the protection scope of the present invention.

Claims

1. A new comprehensive evaluation method for distribution network based on cross-validation method, characterized in that: The following steps are involved: S1: Based on the analytic hierarchy process, a hierarchical model considering low-carbon constraints is constructed, and the initial weights of the evaluation indicators in the hierarchical model are obtained; S2: setting a weight combination criterion based on the initial weights, and obtaining a plurality of groups of combination weights based on the weight combination criterion; S3: Based on the first historical data set of the new distribution network, the optimal combination weight is selected from several groups of combination weights using the cross-validation method, and the hierarchical structure model is updated based on the optimal combination weight to obtain the optimal hierarchical structure model; S4: Perform stability verification on the optimal hierarchical structure model based on the second historical data set of the new distribution network. If the stability verification passes, perform a comprehensive evaluation based on the optimal hierarchical structure model. If the stability verification fails, use the optimal combination weight as the initial weight and execute S2.

2. According to the new distribution network comprehensive evaluation method based on cross-validation method according to claim 1, it is characterized in that: In S1, the initial weights of the evaluation indicators in the hierarchical model are obtained, including: S11: Constructing a judgment matrix of a hierarchical structure model based on expert evaluation in the analytic hierarchy process; S12: normalize and perform consistency test on the judgment matrix, and obtain the initial weights of the evaluation indicators in the hierarchical model according to the consistency test results.

3. The new distribution network comprehensive evaluation method based on cross-validation method according to claim 2 is characterized in that: In S12, the step of obtaining the initial weights of the evaluation indicators in the hierarchical model according to the consistency test results includes: If the consistency check result meets the consistency requirement, the process weight obtained by the normalization processing is the initial weight. If it does not meet the consistency requirement, the judgment matrix is ​​updated based on the relationship between the data in the judgment matrix, and S12 is executed.

4. The novel comprehensive evaluation method for distribution network based on cross-validation method according to claim 1 is characterized in that: The S2 includes: S21: taking the initial weight as a benchmark, setting a floating standard based on comprehensive evaluation requirements, and obtaining the candidate weights of the evaluation indicators according to the floating standard; S22: setting the weight combination criterion according to the size relationship between the initial weights and the principle of full coverage and uniqueness, and combining the candidate weights based on the weight combination criterion to obtain a plurality of groups of combined weights.

5. The novel comprehensive evaluation method for distribution network based on cross-validation method according to claim 1 is characterized in that: The S3 includes: S31: performing data cleaning, missing value filling and outlier processing on the first historical data set; S32: Training the hierarchical structure model based on several groups of combination weights and the first historical data set and obtaining the optimal combination weights.

6. The novel comprehensive evaluation method for distribution network based on cross-validation method according to claim 5 is characterized in that: The S32 includes: S321: inputting the first historical data set and several groups of combined weights into the hierarchical structure model, calculating first products of the several groups of combined weights and corresponding data in the first historical data set, comparing the first product with actual results of corresponding data in the first historical data set, and obtaining the number of successful comparisons; S322: Determine whether the number of successful comparisons of the combination weight with the highest number of successful comparisons is greater than a first preset number. If so, the combination weight with the highest number of successful comparisons is the optimal combination weight. If so, the combination weight with the highest number of successful comparisons is used as the initial weight and S2 is executed.

7. The novel distribution network comprehensive evaluation method based on cross-validation method according to claim 1 is characterized in that: In S4, the stability verification of the optimal hierarchical structure model based on the second historical data set of the new distribution network includes: The second historical data set is input into the optimal hierarchical model, the second product of the optimal combination weight and the corresponding data in the second historical data set is calculated, and the second product is matched with the actual results of the corresponding data in the second historical data set. If the number of successful matches exceeds the second preset number, the stability verification passes; if the number of successful matches does not exceed the second preset number, the stability verification fails.

8. The novel comprehensive evaluation method for distribution network based on cross-validation method according to claim 1 is characterized in that: In S1, the evaluation indicators include at least the distributed power supply absorption capacity, power supply capacity, power supply reliability, extreme disaster scenario recovery capability and the distribution network carbon emission rights trading cost considering low-carbon constraints.

9. The novel distribution network comprehensive evaluation method based on cross-validation method according to claim 8 is characterized in that: The absorption capacity of the distributed power source is characterized by the absorption rate. Among them, P DG is the power generation of distributed generation, P tol is the load demand of the distribution network.

10. The novel comprehensive evaluation method for distribution network based on cross-validation method according to claim 8 is characterized in that: The power supply capability is characterized by a voltage deviation. Among them, V actual is the actual voltage of the distribution network node, V c is the rated voltage of the distribution network.

Images