Distribution network team evaluation index differentiation analysis system based on k-means clustering

Abstract

The present invention discloses a distribution network team evaluation index differential analysis system based on k-means clustering, comprising the following steps: obtaining first data of the distribution network, the first data comprising a user satisfaction evaluation data set, a power outage recovery time data set and third data; constructing a differential analysis module based on k-means clustering according to the first data, and obtaining four clusters for each differential analysis module; screening the distribution network team through the first differential analysis module, and inputting the second differential analysis module to perform secondary clustering screening to obtain a high-quality distribution network team and a power outage recovery time abnormal team; combining the third data to determine whether the power outage recovery time abnormal team belongs to a high-quality distribution network team. The present invention solves the problem of distortion and misjudgment of the differential analysis module in the prior art by clustering the user satisfaction evaluation data and the power outage recovery time data respectively.

Description

Technical Field

[0001] The present invention relates to the field of electric power technology, and more specifically, to a distribution network team evaluation index differentiation analysis system based on k-means clustering. Background Art

[0002] Distribution network team evaluation indicators are important data used to evaluate the performance of distribution network teams, including user satisfaction, power outage recovery time data, third-party data, customer complaint rate, power supply reliability, etc. The differentiation analysis module can be used to make appropriate rewards and punishments, dispatch and allocate personnel and funds for the distribution network team through the differences in distribution network team evaluation indicators.

[0003] However, when conducting a substantial differentiation analysis of the distribution network team, since the user satisfaction index is a commonly used data source, but the user satisfaction index is often falsely reported, when the distribution network team is actually working, they may ask customers for better user satisfaction. Therefore, the resulting differentiation analysis module may be distorted and misjudged when analyzing the distribution network team. Summary of the invention

[0004] In order to overcome the shortcomings of the prior art, the present invention provides a distribution network team evaluation index differentiation analysis system based on k-means clustering, which has the advantage of avoiding distortion and misjudgment of differential system results.

[0005] To achieve the above-mentioned purpose, the present invention provides the following technical solutions: a distribution network team evaluation index differential analysis system based on k-means clustering, comprising the following steps: obtaining first data of the distribution network, the first data including a user satisfaction evaluation data set, a power outage recovery time data set and third data, the third data including equipment aging impact rate and geographical complexity; constructing a differential analysis module based on k-means clustering according to the first data, the differential analysis module including a first differential analysis module and a second differential analysis module, each differential analysis module obtains four clusters; screening the distribution network team through the first differential analysis module, and inputting the second differential analysis module to perform secondary clustering screening to obtain a high-quality distribution network team and a power outage recovery time abnormal team; judging whether the power outage recovery time abnormal team belongs to the high-quality distribution network team in combination with the third data.

[0006] As a preferred technical solution of the present invention, the distribution network team is screened by the first differential analysis module and input into the second differential analysis module for secondary clustering screening, specifically: the distribution network team with high user satisfaction is obtained after screening by the first differential analysis module, and the distribution network team with high user satisfaction is input into the second differential analysis module, and the distribution network team is judged in which cluster of the second differential analysis module according to the power outage recovery time data; the excellent and good results obtained by the second differential analysis module are separately presented, and statistical methods are used to determine whether the power outage recovery time data significantly deviates from the average data within the cluster.

[0007] As a preferred technical solution of the present invention, the secondary clustering screening is performed to obtain a high-quality distribution network team, specifically: the power outage recovery duration data in the general and difference clusters in the second differential analysis module are marked as inaccurate user satisfaction data; the marked data is removed from the first differential analysis module, and the distribution network team after the removal is used as the high-quality distribution network team.

[0008] As a preferred technical solution of the present invention, the combination of the third data to determine whether the team with abnormal power outage recovery time belongs to a high-quality distribution network team is specifically: obtaining the third data in the general and difference clusters in the second differential analysis module; combining the equipment aging impact rate and geographical complexity, calculating the rationality of the power outage recovery time data, and judging the distribution network team with a high rationality of the power outage recovery time data as a high-quality distribution network team.

[0009] As a preferred technical solution of the present invention, the rationality of the power outage recovery time data is specifically as follows: the impact rate of equipment aging exceeds the average impact rate of the entire region, and the rationality of the power outage recovery time data of the distribution network is high when the urban terrain is non-flat; the impact rate of equipment aging is lower than the average impact rate of the entire region, or the rationality of the power outage recovery time data of the distribution network is low when the urban terrain is flat.

[0010] Compared with the prior art, the present invention has the following beneficial effects:

[0011] 1. The present invention uses the k-means clustering method to make two different differential analysis modules for the collected user satisfaction evaluation data set and the power outage recovery time data set, and in each differential analysis module, the data is divided into four clusters, and then the distribution network team clusters are subjected to differential analysis, and the distribution network power outage recovery time in the excellent and good clusters obtained by the first differential analysis module is compared and analyzed in the second differential analysis module. If the comparison result falls into the general and poor in the second differential analysis module, it is marked as abnormal data, especially in the case of poor, which indicates that the distribution network team has asked customers for good reviews during the operation. At this time, the data is marked and removed to obtain the final differential analysis module, thereby avoiding distortion and misjudgment of the differential analysis results. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 is a flow chart of the present invention; DETAILED DESCRIPTION

[0013] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0014] Embodiment 1, as Figure 1 As shown, the present invention provides a distribution network team evaluation index differentiation analysis system based on k-means clustering, comprising the following steps:

[0015] S1 obtains first data of the distribution network, the first data including a user satisfaction evaluation data set, a power outage recovery time data set and third data;

[0016] S2 constructs a differential analysis module based on k-means clustering according to the first data, the differential analysis module includes a first differential analysis module and a second differential analysis module, and each differential analysis module obtains four clusters;

[0017] S3 screens the distribution network teams through the first differentiation analysis module, and inputs them into the second differentiation analysis module for secondary clustering screening to obtain high-quality distribution network teams and teams with abnormal power outage recovery time;

[0018] S4 combines the third data to determine whether the team with abnormal power outage recovery time belongs to a high-quality distribution network team.

[0019] Acquire first data of the distribution network, the first data of the distribution network including a user satisfaction evaluation data set, a power outage restoration time data set and third data;

[0020] Among them, the first data of the distribution network is specifically:

[0021] The data collection module is used to collect data, and also includes a failure rate data set, wherein the user satisfaction evaluation data set includes several groups of satisfaction evaluation data, and the satisfaction evaluation data is a scoring evaluation of 1-10, where 1 is unsatisfied and 10 is satisfied;

[0022] By adopting the scoring evaluation method, users can evaluate and score services more conveniently, and data collection is more convenient. The collection of failure rate data and power outage recovery time data is collected, and the control center is used for unified data extraction.

[0023] Based on k-means clustering, the first differentiation analysis module and the second differentiation analysis module are constructed according to the user satisfaction evaluation data set and the power outage recovery duration data set respectively. The specific process is as follows:

[0024] By using the k-means clustering method, two different differential analysis modules are made for the collected user satisfaction evaluation data set and power outage recovery time data set, and in each differential analysis module, the data is divided into four clusters, and then the distribution network team clusters are subjected to differential analysis. The distribution network power outage recovery time in the excellent and good clusters obtained by the first differential analysis module is compared and analyzed in the second differential analysis module. If the comparison result falls into the general and poor in the second differential analysis module, it is marked as abnormal data, especially in the case of poor, which indicates that the distribution network team has asked customers for good reviews during the operation. At this time, the data is marked and eliminated to obtain the final high-quality distribution network team, thereby avoiding distortion and misjudgment of the differential analysis results.

[0025] The first differentiation analysis module is constructed based on the user satisfaction evaluation data set, which includes four clusters in the first differentiation module;

[0026] The second differentiation analysis module is constructed based on the power outage restoration duration data set, which includes four clusters in the second differentiation module;

[0027] The first differentiation analysis module and the second differentiation analysis module are constructed based on k-means clustering, specifically:

[0028] The first data of the distribution network collected is used by the clustering module to divide the distribution network team into four clusters, namely excellent, good, general and poor. The specific steps are as follows:

[0029] Initialize cluster centers ,in is the cluster center of the i-th cluster;

[0030] For each data point Calculate the distance from each cluster center , and assign it to the nearest cluster. The distance calculation formula is:

[0031]

[0032] Recalculate the center of each cluster , the update formula is:

[0033]

[0034] Repeat the above steps until the cluster center no longer changes;

[0035] in, is the set of all data points in the i-th cluster; is the number of all data points in the i-th cluster; is the jth eigenvalue of the i-th cluster center; represents a data point, which represents the feature vector of the distribution network team; is the jth eigenvalue of the data point; q is the characteristic dimension of the data point; is the distance between the data point and the cluster center. The smaller the distance, the closer the data point is to the cluster center.

[0036] By using the k-means clustering method to cluster data, it is possible to quickly organize and classify data, making the entire system more convenient.

[0037] Among them, the iteration is performed, specifically:

[0038] The clustering results are iterated through the iteration module until four stable clusters are obtained;

[0039] By iteratively processing the clustering results, the data can be made more regular and distributed more evenly, which can improve the overall accuracy of the differential analysis module.

[0040] The distribution network team with high user satisfaction is obtained through the first differential analysis module and input into the second differential analysis module to mark abnormal data;

[0041] In the second differentiation analysis module, the accuracy of distribution network user satisfaction is determined based on the power outage recovery time data to obtain a high-quality distribution network team.

[0042] The power outage recovery duration data in the general and poor clusters in the second differential analysis module are marked as inaccurate user satisfaction data;

[0043] The marked data are removed from the first differential analysis module, and the distribution network team after the removal constitutes a high-quality distribution network team.

[0044] Among them, the clustering of the first differentiation analysis module and the second differentiation analysis module are compared, and the high-quality distribution network team is obtained as follows:

[0045] The excellent and good results obtained in the first differential analysis module are presented separately to obtain two groups of distribution network teams with high user satisfaction. The power outage and restoration time data of these two groups of distribution networks are put into the second differential analysis module for cluster comparison to determine which cluster the distribution network team belongs to in the second differential analysis module.

[0046] The above clustering comparison is a further analysis of different clustering results to determine whether certain data is abnormal. Clustering comparison usually includes the following:

[0047] Cluster the data within the cluster again to analyze whether there is any abnormal data.

[0048] Use statistical methods (such as t-value test) to compare the differences between different clusters and evaluate the significance of the data.

[0049] The comparison process takes into account external factors (such as urbanization level, equipment aging, etc.) to further determine whether the data is reasonable.

[0050] This embodiment selects the t-value test method in the statistical method, which can help the system accurately identify the performance of the distribution network and optimize it to improve user satisfaction and the overall efficiency of the system.

[0051] When comparing the excellent and good clusters, statistical methods such as t-value test or variance analysis are used to determine whether the data with long power outage recovery time significantly deviates from the average data within the cluster. Specifically, the significance is evaluated by calculating the t-value. The t-value calculation formula is:

[0052]

[0053] Where: are the means of the data in the excellent and good clusters in the second differential analysis module, and are the standard deviations of the data in the excellent and good clusters in the second differential analysis module, and are the number of data points in the excellent and good clusters in the second differential analysis module, and t is the data deviation.

[0054] It should be noted that the larger the t value, the more significant the deviation.

[0055] If the comparison result shows that the power outage recovery time data is also in the two clusters of excellent and good in the second differential analysis module, it indicates that the user satisfaction evaluation data of this distribution network is accurate;

[0056] If the comparison result shows that the power outage recovery time data is in the general and difference clusters in the second differential analysis module, it indicates that the user satisfaction evaluation data of this distribution network is inaccurate, and then this distribution network team is eliminated to obtain the final differential analysis module.

[0057] Acquire third data in the general and difference clusters in the second differential analysis module, wherein the third data includes equipment aging impact rate and geographical complexity;

[0058] Based on the impact rate of equipment aging and geographical complexity, the rationality of the power outage recovery time data is calculated, and the distribution network team with a high rationality of the power outage recovery time data is judged as a high-quality distribution network team.

[0059] The rationality of the power outage recovery time data is calculated as follows:

[0060] The equipment aging impact rate and geographic complexity are combined with fuzzy synthesis operators to construct a two-dimensional evaluation matrix;

[0061] Compare the absolute difference between each element and the row mean in the two-dimensional evaluation matrix, and preset the resolution coefficient value;

[0062] The correlation coefficient between the third data of the distribution network and the power outage restoration data is calculated according to the grey correlation analysis formula using the absolute difference and the resolution coefficient;

[0063] The rationality of the distribution network power outage recovery time data is obtained by comparing the correlation coefficient with the standard correlation degree.

[0064] First, the fuzzy synthesis operator is used to integrate the equipment aging impact rate and geographical complexity to construct a two-dimensional evaluation matrix. The two-dimensional evaluation matrix reflects the comprehensive evaluation of the distribution network team in terms of equipment aging and geographical complexity. During the construction process, these two factors are quantified according to the principles of fuzzy mathematics, and the value of each element in the matrix is ​​obtained according to the preset synthesis rules.

[0065] Calculate the absolute difference between each element and the row mean in the two-dimensional evaluation matrix; the row mean represents the average level of the elements in each row of the matrix, and the absolute difference measures the degree of deviation of each element from this average level. Quantify the basic data of the difference between each element and the whole.

[0066] Preset a resolution coefficient value for subsequent grey relational analysis. The resolution coefficient is an important parameter in grey relational analysis, which determines the sensitivity of the correlation coefficient. Select an appropriate resolution coefficient value based on the specific requirements of the problem and the characteristics of the data.

[0067] After obtaining the absolute difference and resolution coefficient, the correlation coefficient between the third data of the distribution network and the power outage recovery data is calculated using the grey correlation analysis formula. The third data represents a comprehensive index related to power outage recovery, while the power outage recovery data is the actual power outage recovery duration or related parameters. By calculating the correlation coefficient, the similarity or difference between the third data and the power outage recovery data is quantified.

[0068] Finally, the calculated correlation coefficient is compared with the preset standard correlation degree to determine the rationality of the power outage recovery time data of the distribution network. If the correlation coefficient is greater than or equal to the standard correlation degree, it is considered that the power outage recovery time data is more reasonable; otherwise, it is less reasonable. This comparison result provides us with a basis for evaluating the effectiveness of the distribution network team in power outage recovery.

[0069] Fuzzy synthesis operator: The fuzzy synthesis operator is a computational tool used in fuzzy mathematics to process fuzzy information. It synthesizes multiple fuzzy factors (such as the impact rate of equipment aging and geographical complexity) according to specific rules to obtain a comprehensive evaluation result. In this process, the fuzzy synthesis operator can take into account the interaction and mutual influence between various factors, making the evaluation result more in line with the actual situation. In this example, the fuzzy synthesis operator is used to construct a two-dimensional evaluation matrix, which reflects the comprehensive evaluation of the distribution network team in terms of equipment aging impact and geographical complexity.

[0070] Absolute difference: Absolute difference refers to the absolute value of the difference between two values, which measures the degree of difference between the two values. In this example, the absolute difference is used to compare the difference between each element and the row mean in the two-dimensional evaluation matrix. By calculating these absolute differences, we can get the degree of deviation of each element from the row mean, thus providing basic data for subsequent grey relational analysis.

[0071] Resolution coefficient: The resolution coefficient is an important parameter in grey relational analysis, which is used to adjust the sensitivity of the correlation coefficient. The value range of the resolution coefficient is usually between 0 and 1, and its value directly affects the calculation result of the correlation coefficient and the final evaluation accuracy. In this example, we preset a resolution coefficient value so that the degree of correlation between the factors can be accurately reflected in the subsequent calculation.

[0072] Grey correlation analysis: Grey correlation analysis is a method used to deal with the analysis of the degree of correlation of multiple factors in a grey system (i.e., a system with incomplete or uncertain information). It measures the similarity or difference between factors by calculating the correlation coefficient between them. In this example, we used the grey correlation analysis formula to calculate the correlation coefficient between the third data of the distribution network (which may represent a comprehensive indicator related to power outage recovery) and the power outage recovery data. These correlation coefficients reflect the degree of correlation between the third data and the power outage recovery data, thus providing us with an important basis for judging the rationality of the power outage recovery duration data.

[0073] Standard correlation: Standard correlation is a benchmark value used to compare and judge the correlation degree of each factor in grey correlation analysis. It is usually preset based on actual needs and expert experience, and is used to compare the calculated correlation coefficient with the benchmark value to determine whether the correlation degree of each factor meets the expected standard. In this example, we compare the calculated correlation coefficient with the preset standard correlation to determine whether the rationality of the power outage recovery time data of the distribution network meets the established standard.

[0074] Through the t-value test, the excellent and good clusters of the second differential analysis module can be verified, and the clustering of the data can be double-insured to ensure the accuracy of the clustering. At the same time, the iteration results can be verified to ensure the smooth operation of the iteration. When problems occur in the iteration, the iteration can be stopped in time, and the iteration algorithm can be verified and analyzed to achieve the verification purpose; for the distribution network teams whose comparison results are excellent and good in the second differential analysis module, adaptive commendation is carried out to achieve differentiated management of distribution network teams with good services. For the distribution network teams whose comparison results are general and poor in the second differential analysis module, further analysis is carried out to correct the accuracy of the data of the differential analysis module of the overall distribution network team evaluation index to avoid data distortion and misjudgment.

[0075] Abnormal data is marked for the distribution network teams in the general and poor clusters in the first differentiation module to avoid the situation where the distribution network team asks for better evaluation in subsequent service work, and a notification of abnormal situation is issued to the distribution network team to warn the distribution network team to improve compliance in the work process.

[0076] By directly marking the distribution network teams in the two clusters of general and poor in the second differentiation analysis module, and further analyzing the third data in the future, differentiated management of distribution network teams in different clusters can be achieved, avoiding the situation where the established third data standards deviate too much from the actual service capacity, so as to ensure the targeted and inclusive nature of differentiated management.

[0077] Acquire third data in the general and difference clusters in the second differential analysis module, wherein the third data includes equipment aging impact rate and geographical complexity;

[0078] Based on the impact rate of equipment aging and geographical complexity, the rationality of the power outage recovery time data is calculated, and the distribution network team with a high rationality of the power outage recovery time data is judged as a high-quality distribution network team.

[0079] The equipment aging impact rate is specifically:

[0080]

[0081] Where: and Respectively The aging degree and power outage times of each data point, and are the average values ​​of equipment aging degree and power outage times, The impact rate of equipment aging on the number of power outages in the area where the distribution network team is located. The total number of data points involved in the calculation of the equipment aging impact rate.

[0082] The geographic complexity is obtained by:

[0083] Acquiring geographic feature data, wherein the geographic feature data includes vegetation coverage density, land use rate, and average terrain slope;

[0084] The geographic feature data are mapped to three-dimensional space points, and the geometric distance and topological distance of each data to the origin are calculated based on a comprehensive evaluation of multiple indicators to obtain the geographic complexity.

[0085] In the above process, geometric distance refers to the straight-line distance from geographic feature data points (such as vegetation coverage density, land use rate, and three-dimensional space points mapped to average terrain slope) to the origin in three-dimensional space, reflecting the absolute position relationship of the data points. Topological distance takes into account the connectivity and relative position between data points, and measures the structural or layout complexity of geographic feature data in three-dimensional space.

[0086] The methods for obtaining geometric distance and topological distance are based on existing mature technologies, which are widely used in fields such as multidimensional space analysis and graph theory. Geometric distance is usually obtained by calculating the straight-line distance between spatial point coordinates, while topological distance relies on the path search algorithm in graph theory to measure the relative position and connectivity between data points. Since these technologies are quite mature and widely used, their specific methods of obtaining them will not be described in detail here.

[0087] The calculation formula of geographic complexity is as follows:

[0088]

[0089] Where: is the geographical complexity, is the geometric distance of the i-th geographic feature data, is the topological distance of the i-th geographic feature data.

[0090] It should be noted that the smaller the value of geographical complexity, the higher the difficulty of human operation.

[0091] When comparing the distribution teams located in the general and poor clusters, if the equipment aging impact rate is lower than the average impact rate of the entire region and the urban terrain is flat, it indicates that the service of this distribution network needs to be improved.

[0092] The standard for whether the urban terrain is flat is determined by the ratio of the urban area to the total urban area. If the ratio is greater than or equal to the first threshold, the urban terrain is flat, and if the ratio is less than the first threshold, the urban terrain is complex.

[0093] The comparison of the distribution teams in the general and poor clusters is as follows:

[0094] If the city where the power distribution team is located has a low urbanization level or complex terrain, this distribution network team will be marked as a concern;

[0095] By establishing the corresponding third data standard in the cluster, and separately analyzing the distribution network teams in the two general and poor clusters obtained by the second differentiation analysis module, a comprehensive analysis is performed in combination with the urbanization level and urban terrain of the area where the distribution network team is located. If the distribution network team is located in an area with a low urbanization level or a relatively complex urban terrain, the distribution network team will be marked as a concern, indicating that the work of the distribution network is difficult and needs attention, and additional resources need to be dispatched to the distribution network team. Based on this mark, resources are allocated to the distribution network team to improve the electricity experience of users in areas with a low urbanization level or a relatively complex urban terrain, improve the overall distribution network user satisfaction evaluation, and realize targeted resource allocation in areas with a low urbanization level or a relatively complex urban terrain.

[0096] By associating the third data with the equipment aging situation, it is possible to verify and predict the subsequent corresponding costs of replacing and repairing equipment in the area where the distribution network team is located, so as to achieve the rationality of resource allocation distribution among the overall distribution network team.

[0097] For the distribution team, which is at the impact rate of equipment aging and is in a flat urban terrain, it is still located in the worse cluster of the second differential analysis module. The work difficulty is verified for the distribution network team. If it is found that the work difficulty is not high, it is necessary to find the reasons for the long power outage recovery time data. According to the 4MIE criteria, the reasons are searched for personnel, equipment, raw materials, working methods and environment respectively. If there is an unavoidable situation that causes the power outage recovery time data to be too long, the working time is eliminated separately and the differential analysis module is re-evaluated.

[0098] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0099] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A distribution network team evaluation index differentiation analysis system based on k-means clustering, characterized by: The following steps are involved: A data acquisition module is used to acquire first data of the distribution network, the first data includes a user satisfaction evaluation data set, a power outage recovery time data set and third data, the third data includes an equipment aging impact rate and geographic complexity, and the geographic complexity is obtained based on the point spacing of the geographic feature data in three-dimensional space; A clustering analysis module, used to construct a differential analysis module according to the first data and based on k-means clustering, wherein the differential analysis module includes a first differential analysis module and a second differential analysis module; The secondary clustering module is used to screen the distribution network team through the first differentiation analysis module, and input the second differentiation analysis module to perform secondary clustering screening to obtain high-quality distribution network teams and teams with abnormal power outage recovery time, specifically: The distribution network team with high user satisfaction is screened out through the first differentiation analysis module, and is input into the second differentiation analysis module. According to the power outage recovery time data, it is determined in which cluster of the second differentiation analysis module the distribution network team is located; The excellent and good results obtained by the second differential analysis module are presented separately, and statistical methods are used to determine whether the power outage recovery time data significantly deviates from the average data within the cluster; The power outage recovery duration data in the general and poor clusters in the second differential analysis module are marked as inaccurate user satisfaction data; The marked data is removed from the first differential analysis module, and the distribution network team after the removal is regarded as a high-quality distribution network team; The in-depth analysis module is used to conduct multi-indicator evaluation on the team with abnormal power outage restoration time in combination with third-party data to determine whether it is a high-quality distribution network team.

2. The distribution network team evaluation index differentiation analysis system based on k-means clustering as claimed in claim 1, characterized in that: The third data is combined to perform a multi-index evaluation on the power outage recovery time abnormal team to determine whether it belongs to a high-quality distribution network team, specifically: Obtaining third data in the general and difference clusters in the second differential analysis module; Combined with the third data, the rationality of the power outage recovery time data is calculated, and the distribution network team with a high rationality of the power outage recovery time data is judged as a high-quality distribution network team.

3. The distribution network team evaluation index differentiation analysis system based on k-means clustering as claimed in claim 2, characterized in that: The rationality of the power outage recovery time data is calculated as follows: The equipment aging impact rate and geographic complexity are combined with fuzzy synthesis operators to construct a two-dimensional evaluation matrix; Compare the absolute difference between each element and the row mean in the two-dimensional evaluation matrix, and preset the resolution coefficient value; The correlation coefficient between the third data of the distribution network and the power outage restoration data is calculated according to the grey correlation analysis formula using the absolute difference and the resolution coefficient; The rationality of the distribution network power outage recovery time data is obtained by comparing the correlation coefficient with the standard correlation degree.

4. The distribution network team evaluation index differentiation analysis system based on k-means clustering as claimed in claim 3, characterized in that: The equipment aging impact rate is specifically: Where: and Respectively The aging degree and power outage times of each data point, and are the average values ​​of equipment aging degree and power outage times, The impact rate of equipment aging on the number of power outages in the area where the distribution network team is located. The total number of data points involved in the calculation of the equipment aging impact rate.

5. The distribution network team evaluation index differentiation analysis system based on k-means clustering as claimed in claim 4, characterized in that: The geographic complexity is obtained by: Acquiring geographic feature data, wherein the geographic feature data includes vegetation coverage density, land use rate, and average terrain slope; Map geographic feature data to three-dimensional space points, calculate the geometric distance and topological distance of each data to the origin based on multi-index comprehensive evaluation, and obtain geographic complexity; The calculation formula of geographic complexity is as follows: Where: is the geographical complexity, is the geometric distance of the i-th geographic feature data, is the topological distance of the i-th geographic feature data.

6. The distribution network team evaluation index differentiation analysis system based on k-means clustering as claimed in claim 5, characterized in that: The statistical method used to determine whether the power outage recovery time data significantly deviates from the average data within the cluster is as follows: Where: and are the means of the data in the excellent and good clusters in the second differential analysis module, and are the standard deviations of the data in the excellent and good clusters in the second differential analysis module, and are the number of data points in the excellent and good clusters in the second differential analysis module, and t is the data deviation.

Images