A high line loss area division method, device, equipment and readable storage medium

Abstract

The application discloses a high line loss area division method and device, equipment and a readable storage medium, relates to the technical field of power data processing, and comprises the following steps: acquiring line loss rate time series of a plurality of to-be-classified areas; calculating the line loss rate time series of the plurality of to-be-classified areas respectively, and obtaining the line loss rate mean value of each to-be-classified area; constructing a line loss rate box plot according to the line loss rate mean value of each to-be-classified area; determining an initial clustering centroid according to the line loss rate box plot; clustering the line loss rate time series of the plurality of to-be-classified areas through a first clustering algorithm and the initial clustering centroid, obtaining a first clustering set and a second clustering set; and dividing the to-be-classified areas corresponding to the line loss rate time series included in the first clustering set into high line loss areas. It can be seen that the application can comprehensively consider the operation characteristics of the to-be-classified areas and the correlation between the to-be-classified areas when dividing the high line loss areas, thereby improving the division accuracy of the high line loss areas.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of electric power data processing, and more particularly to a high line loss area division method, device and equipment and readable storage medium. BACKGROUND

[0002] In the process of power supply system delivering and distributing electric energy, a certain amount of active power loss and electric energy loss generated by each element in the power grid is collectively referred to as line loss, simply referred to as line loss. The percentage of line loss in power supply is referred to as line loss rate, simply referred to as line loss rate. The line loss rate is an index for measuring the level of line loss, which comprehensively reflects and embodies the level of power system planning, design, production operation and management, and is an important economic and technical index of power grid operation enterprises.

[0003] The power supply enterprise can divide the power supply area into high line loss area and reasonable line loss area to manage the line loss of the power supply area. Usually, the power supply area with line loss rate exceeding a certain threshold is directly divided into high line loss area, however, this method is too extensive and simple for determining the high line loss area, and the accuracy is not high, which is difficult to meet the requirements of the power supply enterprise for lean management of the line loss of the area. SUMMARY

[0004] The embodiments of the present application provide a high line loss area division method, device and equipment and readable storage medium, which can improve the division accuracy of the high line loss area.

[0005] Therefore, the embodiments of the present application provide a high line loss area division method, which comprises:

[0006] obtaining line loss rate time series of a plurality of to-be-classified areas;

[0007] respectively calculating the line loss rate time series of the plurality of to-be-classified areas to obtain line loss rate mean values of each to-be-classified area;

[0008] constructing a line loss rate box plot according to the line loss rate mean values of each to-be-classified area;

[0009] determining an initial clustering centroid according to the line loss rate box plot;

[0010] clustering the line loss rate time series of the plurality of to-be-classified areas by a first clustering algorithm and the initial clustering centroid to obtain a first clustering set and a second clustering set, the first clustering set corresponding to a first clustering centroid, the second clustering set corresponding to a second clustering centroid, and the line loss rate mean value corresponding to the first clustering centroid being greater than the line loss rate mean value corresponding to the second clustering centroid;

[0011] dividing the to-be-classified areas corresponding to the line loss rate time series included in the first clustering set into high line loss areas.

[0012] Optionally, the initial clustering centroids include a first initial clustering centroid and a second initial clustering centroid, and the determining the initial clustering centroids according to the line loss rate box plot includes:

[0013] If there is an abnormal point greater than the upper edge of the line loss rate box plot in the line loss rate box plot, determining a line loss rate time sequence corresponding to the abnormal point with the smallest value as the first initial clustering centroid;

[0014] If there is no abnormal point greater than the upper edge of the line loss rate box plot in the line loss rate box plot, determining a line loss rate time sequence corresponding to the point with the largest value in the line loss rate box plot as the first initial clustering centroid;

[0015] Determining a line loss rate time sequence corresponding to the median point in the line loss rate box plot as the second initial clustering centroid.

[0016] Optionally, the method further includes:

[0017] Clustering the first clustering set by a second clustering algorithm to obtain a plurality of first clustering sub-sets;

[0018] Determining a line loss rate time sequence corresponding to the centroid of each first clustering sub-set to be detected;

[0019] Analyzing each line loss rate time sequence to be detected to obtain feature data corresponding to each line loss rate time sequence to be detected;

[0020] Constructing a line loss rate area image corresponding to each first clustering sub-set according to the feature data corresponding to each line loss rate time sequence to be detected, respectively.

[0021] Optionally, the second clustering algorithm includes a K-means improved clustering algorithm, and the clustering the first clustering set by the second clustering algorithm to obtain a plurality of first clustering sub-sets includes:

[0022] Calculating a contour coefficient according to each line loss rate time sequence included in the first clustering set;

[0023] Determining a number of clusters according to the contour coefficient;

[0024] Clustering the first clustering set by the K-means improved clustering algorithm and the number of clusters to obtain a plurality of first clustering sub-sets.

[0025] Optionally, the analyzing each line loss rate time sequence to be detected to obtain feature data corresponding to each line loss rate time sequence to be detected includes:

[0026] The trend of each line loss rate time series to be detected is obtained by performing a test on each line loss rate time series to be detected through the Mann-Kendall method.

[0027] The variance and the mean of each line loss rate time series to be detected are obtained by performing calculation on each line loss rate time series to be detected.

[0028] The embodiment of the application further provides a high line loss area division device, comprising:

[0029] An acquisition unit is configured to acquire line loss rate time series of a plurality of areas to be classified;

[0030] A calculation unit is configured to perform calculation on the line loss rate time series of the plurality of areas to be classified respectively, and obtain a line loss rate mean of each area to be classified;

[0031] A first construction unit is configured to construct a line loss rate box plot according to the line loss rate mean of each area to be classified;

[0032] A first determination unit is configured to determine an initial clustering center according to the line loss rate box plot;

[0033] A first clustering unit is configured to perform clustering on the line loss rate time series of the plurality of areas to be classified through a first clustering algorithm and the initial clustering center, and obtain a first clustering set and a second clustering set, the first clustering set corresponds to a first clustering center, the second clustering set corresponds to a second clustering center, and the line loss rate mean corresponding to the first clustering center is greater than the line loss rate mean corresponding to the second clustering center;

[0034] A division unit is configured to divide an area to be classified corresponding to a line loss rate time series included in the first clustering set into a high line loss area.

[0035] Optionally, the initial clustering center comprises a first initial clustering center and a second initial clustering center, and the first determination unit is specifically configured to:

[0036] If there is an abnormal point greater than an upper edge of the line loss rate box plot in the line loss rate box plot, a line loss rate time series corresponding to the abnormal point with the smallest value is determined as the first initial clustering center;

[0037] If there is no abnormal point greater than the upper edge of the line loss rate box plot in the line loss rate box plot, a line loss rate time series corresponding to a point with the largest value in the line loss rate box plot is determined as the first initial clustering center;

[0038] A line loss rate time series corresponding to a median point in the line loss rate box plot is determined as the second initial clustering center.

[0039] Optionally, the device further comprises:

[0040] a second clustering unit, configured to cluster the first clustering set by a second clustering algorithm to obtain a plurality of first clustering sub-sets;

[0041] a second determining unit, configured to determine a to-be-detected line loss rate time sequence corresponding to a center of each first clustering sub-set;

[0042] an analyzing unit, configured to analyze each to-be-detected line loss rate time sequence to obtain feature data corresponding to each to-be-detected line loss rate time sequence;

[0043] a second constructing unit, configured to construct a line loss rate transformer area image corresponding to each first clustering sub-set according to the feature data corresponding to each to-be-detected line loss rate time sequence.

[0044] Embodiments of the present application further provide a computer device, comprising a memory, a processor and a bus system;

[0045] The memory is configured to store programs.

[0046] The processor is configured to execute the programs in the memory to implement the various methods described above.

[0047] The bus system is configured to connect the memory and the processor to enable the memory and the processor to communicate.

[0048] Embodiments of the present application further provide a computer readable storage medium storing instructions, which, when executed on a computer, enable the computer to perform the various methods described above.

[0049] The embodiment of the present application provides a high line loss area division method, comprising the following steps: acquiring line loss rate time series of a plurality of to-be-classified areas; calculating the line loss rate time series of the plurality of to-be-classified areas respectively to obtain line loss rate mean values of each to-be-classified area; constructing a line loss rate box plot according to the line loss rate mean values of each to-be-classified area; determining initial clustering centroids according to the line loss rate box plot; clustering the line loss rate time series of the plurality of to-be-classified areas through a first clustering algorithm and the initial clustering centroids to obtain a first clustering set and a second clustering set, the first clustering set corresponds to first clustering centroids, the second clustering set corresponds to second clustering centroids, and the line loss rate mean value corresponding to the first clustering centroid is greater than the line loss rate mean value corresponding to the second clustering centroid; and dividing the to-be-classified areas corresponding to the line loss rate time series included in the first clustering set into high line loss areas. It can be seen that the present application can construct a line loss rate box plot of to-be-classified areas, and divide the to-be-classified areas into high line loss areas based on a clustering algorithm, so that the operation characteristics of to-be-classified areas and the relevance between to-be-classified areas can be comprehensively considered when high line loss areas are divided, thereby improving the division accuracy of high line loss areas, and further improving the management level of the line loss management method of a power supply enterprise. BRIEF DESCRIPTION OF DRAWINGS

[0050] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without any creative effort on the basis of the provided drawings.

[0051] Figure 1 A flowchart of a high line loss area division method provided by the embodiment of the present application;

[0052] Figure 2 A schematic diagram of a box plot provided by the embodiment of the present application;

[0053] Figure 3 A structural schematic diagram of a high line loss area division device provided by the embodiment of the present application. DETAILED DESCRIPTION

[0054] The technical solutions in the embodiments of the present application will be described clearly and completely in the following with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without any creative effort belong to the scope of protection of the present application.

[0055] In the process of power supply system delivering and distributing electric energy, a certain amount of active power loss and electric energy loss generated by each element in the power grid is collectively referred to as line loss, simply referred to as line loss. The percentage of line loss in power supply is referred to as line loss rate, simply referred to as line loss rate. The line loss rate is an index for measuring the level of line loss, which comprehensively reflects and embodies the level of power system planning, design, production operation and management, and is an important economic and technical index of power grid operating enterprises.

[0056] The power supply enterprise can divide the power supply area into high line loss area and reasonable line loss area to manage the line loss of the power supply area. Usually, the power supply area with line loss rate exceeding a certain threshold is directly divided into high line loss area, however, this method is too extensive and simple for determining the high line loss area, and when the line loss rates of the power supply areas to be divided are not significantly different, the high line loss area cannot be accurately divided, the accuracy is not high, and it is difficult to meet the requirements of the power supply enterprise for lean management of the line loss of the area.

[0057] Therefore, in order to solve the above problems, the embodiment of the present application provides a high line loss area division method, device and equipment and readable storage medium, which can improve the division accuracy of the high line loss area.

[0058] Please refer to Figure 1 The safety authentication method provided by the embodiment of the present application includes the following steps.

[0059] S101, obtaining line loss rate time series of a plurality of to-be-classified areas.

[0060] In this embodiment, the line loss rate time series of a plurality of to-be-classified areas can be obtained first. It can be understood that a plurality of to-be-classified areas can be divided at the same time, first, the historical line loss rate data of a plurality of to-be-classified areas can be obtained, and the historical line loss rate data can be constructed into line loss rate time series in units of days, so as to obtain the line loss rate time series of each to-be-classified area. Specifically, if there are n to-be-classified areas in the region, the line loss rate time series of the i th area can be represented as l i =(l i1 ,l i2 ,...,l im ).

[0061] S102, respectively calculating the line loss rate time series of the plurality of to-be-classified areas, to obtain the line loss rate mean value of each to-be-classified area.

[0062] In this embodiment, after obtaining the line loss rate time series of a plurality of to-be-classified areas, the line loss rate time series of a plurality of to-be-classified areas can be calculated respectively, to obtain the line loss rate mean value of each to-be-classified area. Specifically, the line loss rate sequence l i =(l i1 ,li2 ,...,l im (average)

[0063] value This yields the average line loss rate for each transformer substation to be classified, which can then be used to construct a sequence.

[0064] S103. Construct a box plot of line loss rate based on the average line loss rate of each substation to be classified.

[0065] In this embodiment, after obtaining the average line loss rate of each transformer substation to be classified, a sequence composed of the average line loss rates of each transformer substation to be classified can be used. Construct a box plot of line loss rate.

[0066] It is understandable that, such as Figure 2 As shown, a box plot is a statistical graph used to display the dispersion of a set of data. The steps for drawing a box plot of line loss rate may include:

[0067] 1. Draw a number line, with the unit of measurement consistent with the unit of the data batch, the starting point slightly smaller than the minimum value, and the length slightly longer than the range of the sequence composed of the mean of the line loss rate.

[0068] 2. Draw a rectangular box, with the positions of its two ends corresponding to the upper and lower quartiles (Q3 and Q1) of the sequence composed of the mean line loss rates. Draw a line segment inside the rectangular box at the median (Xm) position as the median line.

[0069] 3. Draw two line segments identical to the median at Q3+1.5IQR and Q1-1.5IQR. These two segments are outlier cutoff points, called the inner limits. Draw two line segments at Q3+3IQR and Q1-3IQR, called the outer limits. Points outside the inner limits represent outliers. Outliers between the inner and outer limits are mild outliers, while those outside the outer limits are extreme outliers. The interquartile range (IQR) = Q3 - Q1.

[0070] 4. Draw a line segment from each end of the rectangular box outwards to the farthest point that is not an outlier. This represents the distribution range of normal values ​​for this batch of data.

[0071] 5. Mild outliers can be marked with "〇" and extreme outliers with "*". Data points with the same value should be marked side by side on the same data line, while data points with different values ​​should be marked on different data lines.

[0072] S104. Determine the initial cluster centroids based on the line loss rate box plot.

[0073] In this embodiment, after the line loss rate box plot is constructed, the initial clustering centroids can be determined according to the line loss rate box plot, so as to cluster the line loss rate time series of the to-be-classified transformer areas. It can be understood that, since the to-be-classified transformer areas need to be divided into high line loss transformer areas and reasonable line loss transformer areas, two initial clustering centroids can be determined according to the line loss rate box plot, that is, the initial clustering centroids can include a first initial clustering centroid and a second initial clustering centroid.

[0074] In a possible implementation, if there is an abnormal point greater than the upper edge of the line loss rate box plot in the line loss rate box plot, the line loss rate time series corresponding to the abnormal point with the minimum value is determined as the first initial clustering centroid; if there is no abnormal point greater than the upper edge of the line loss rate box plot in the line loss rate box plot, the line loss rate time series corresponding to the point with the maximum value in the line loss rate box plot is determined as the first initial clustering centroid; and the line loss rate time series corresponding to the median point in the line loss rate box plot is determined as the second initial clustering centroid. It can be understood that the initial clustering centroids can be determined according to the specific situation of the constructed line loss rate box plot. If there is an abnormal point greater than the upper edge in the line loss rate box plot, the line loss rate time series corresponding to the abnormal point with the minimum value can be selected as the first initial clustering centroid; if there is no abnormal point greater than the upper edge in the line loss rate box plot, the line loss rate time series corresponding to the point with the maximum value in the line loss rate box plot can be selected as the first initial clustering centroid. In addition, the median point in the line loss rate box plot can be directly determined as the second initial clustering centroid. By clustering according to the initial clustering centroids determined according to the line loss rate box plot, the to-be-classified transformer areas can be more reasonably divided into high line loss transformer areas and reasonable line loss transformer areas.

[0075] S105, clustering the line loss rate time series of the plurality of to-be-classified transformer areas by using a first clustering algorithm and the initial clustering centroids, to obtain a first clustering set and a second clustering set.

[0076] In this embodiment, the line loss rate time series of the plurality of to-be-classified areas can be clustered by a first clustering algorithm and initial clustering centroids to obtain a first clustering aggregation and a second clustering set, wherein the first clustering set corresponds to the first clustering centroids, the second clustering set corresponds to the second clustering centroids, and the line loss rate mean value corresponding to the first clustering centroids is greater than the line loss rate mean value corresponding to the second clustering centroids. It can be understood that the first clustering algorithm can be a K-means clustering algorithm, the distance of the line loss rate time series of each to-be-classified area to the initial clustering centroid can be calculated respectively, and the line loss rate time series of each to-be-classified area can be assigned to the clustering set corresponding to the initial clustering centroid with the closest distance. After the assignment of all line loss rate time series corresponding to the to-be-classified areas is completed, the centroids of the two clustering sets are recalculated and compared with the clustering centroids of the previous time. If the clustering centroids change, the process is repeated until the centroids no longer change, and the final clustering result, i.e., the first clustering aggregation and the second clustering aggregation, is output, wherein the line loss rate mean value of the first clustering centroids corresponding to the first clustering set is greater than the line loss rate mean value of the second clustering centroids corresponding to the second clustering set.

[0077] In S106, the to-be-classified areas corresponding to the line loss rate time series included in the first clustering set are divided into high line loss areas.

[0078] In this embodiment, after the line loss rate time series of the plurality of to-be-classified areas are clustered to obtain the first clustering set and the second clustering set, the to-be-classified areas corresponding to the line loss rate time series included in the first clustering set can be divided into high line loss areas. It can be understood that, since the line loss rate mean value of the first clustering centroids corresponding to the first clustering set is greater than the line loss rate mean value of the second clustering centroids corresponding to the second clustering set, it is known that the line loss rate mean value of the line loss rate time series included in the first clustering set is generally higher than the line loss rate mean value of the line loss rate time series included in the second clustering set. At this time, the to-be-classified areas corresponding to the line loss rate time series included in the first clustering set can be divided into high line loss areas. In one possible implementation manner, the to-be-classified areas corresponding to the line loss rate time series included in the second clustering set can also be divided into reasonable line loss areas.

[0079] In one possible implementation manner, after the to-be-classified areas corresponding to the line loss rate time series included in the first clustering set are divided into high line loss areas, the first clustering set can also be clustered by a second clustering algorithm to obtain a plurality of first clustering sub-sets; the to-be-detected line loss rate time series corresponding to the centroid of each first clustering sub-set is determined; each to-be-detected line loss rate time series is analyzed to obtain feature data corresponding to each to-be-detected line loss rate time series; and the line loss area image corresponding to each first clustering sub-set is constructed according to the feature data corresponding to each to-be-detected line loss rate time series.

[0080] Understandably, after identifying high-line-loss transformer substations, a profile of these substations can be further constructed, thereby improving the management level of power supply companies' line loss management methods. This can be achieved by further clustering the line loss rate time series included in the first cluster set using a second clustering algorithm, resulting in multiple first cluster subsets. The line loss rate time series corresponding to the centroid of each first cluster subset is then identified as the line loss rate time series to be detected. By analyzing each line loss rate time series to be detected, corresponding feature data is obtained. Based on this feature data, a profile of the transformer substation corresponding to each first cluster subset can be constructed, i.e., a profile of the transformer substation corresponding to each line loss rate time series included in each first cluster subset. At this point, the profile of the transformer substation corresponding to each line loss rate time series included in each first cluster subset can be the same.

[0081] In one possible implementation, the second clustering algorithm may include a K-means++ clustering algorithm, which calculates a silhouette coefficient based on each line loss rate time series included in the first cluster set; determines the number of clusters based on the silhouette coefficients; and clusters the first cluster set using the K-means++ clustering algorithm and the number of clusters to obtain multiple first cluster subsets. It is understood that the silhouette coefficients can be calculated first based on the line loss rate time series included in the first cluster set to determine the number of clusters required. Specifically, for each line loss rate time series included in the first cluster set... x Calculate l x The average distance to all elements within the same cluster is denoted as a(l x ), used to quantify cohesion within a cluster; select l x Another cluster outside, calculate l x Find the minimum average distance between the cluster and all elements in the cluster by iterating through k-1 all other clusters. This minimum distance is denoted as b(l). x ), used to quantify the separation between clusters; for object l x The corresponding contour coefficient can be:

[0082]

[0083] The silhouette coefficient corresponding to each line loss rate time series is calculated. After calculating each line loss rate time series, the average value can be calculated to obtain the overall silhouette coefficient of the current cluster, and the clustering effect can be evaluated. Using an enumeration method, with k ranging from 2 to 8, the K-means++ clustering algorithm is performed on each k value and the first cluster set. The k with the largest silhouette coefficient is taken as the final number of clusters. After determining the number of clusters k, the first cluster set is clustered using the K-means++ clustering algorithm and the number of clusters, resulting in multiple first cluster subsets.

[0084] In one possible implementation, the characteristic data corresponding to each time series of line loss rate to be detected may include the trend, variance, and mean of the line loss rate. The Mann-Kendall method can be used to examine each time series of line loss rate to be detected, obtaining the trend of the line loss rate for each time series; the variance and mean of each time series of line loss rate to be detected can be calculated. It is understood that after determining the time series of line loss rate to be detected corresponding to the centroid of each first cluster subset, the Mann-Kendall method can be used to perform trend analysis on each time series of line loss rate to be detected to obtain the trend of the line loss rate for each time series of line loss rate to be detected. Specifically, for each time series of line loss rate to be detected (X1, X2, ..., X...), the trend can be obtained by... m Define the test statistic S as follows:

[0085]

[0086] Where sign() is the sign function. When X j -X i When less than, equal to, or greater than 0, sign(X) j -X i The values ​​are -1, 0, or 1, respectively; when m >= 8, the statistic S follows a normal distribution, with an expected value E(S) = 0 and a variance V. ar (S) = m(m-1)(2m+5) / 18. The formula for calculating the trend Z of the line loss rate is as follows:

[0087]

[0088] Where Z is a positive value indicating an upward trend and a negative value indicating a downward trend.

[0089] Then, the variance and mean of each line loss rate time series to be detected can be calculated according to mathematical statistical formulas. Finally, the line loss rate trend, contrast and mean of each line loss rate time series to be detected are used as the feature data corresponding to each line loss rate time series to be detected, so as to facilitate the subsequent construction of a profile of each high line loss transformer area.

[0090] Therefore, the embodiment of the application provides a high line loss transformer area division method, which can construct a line loss rate box plot of a transformer area to be classified, and divide the transformer area to be classified into a high line loss transformer area based on a clustering algorithm, so that the operation characteristics of the transformer area to be classified and the correlation between the transformer areas to be classified are comprehensively considered when the high line loss transformer area is divided, the division accuracy of the high line loss transformer area is improved, and the management level of the power supply enterprise in line loss management is improved.

[0091] Please refer to Figure 3 The embodiment of the application provides a high line loss transformer area division device, which comprises:

[0092] The acquisition unit 301 is configured to acquire line loss rate time series of a plurality of transformer areas to be classified.

[0093] The calculation unit 302 is configured to calculate the line loss rate time series of the plurality of transformer areas to be classified respectively, and obtain a line loss rate mean value of each transformer area to be classified.

[0094] The first construction unit 303 is configured to construct a line loss rate box plot according to the line loss rate mean value of each transformer area to be classified.

[0095] The first determination unit 304 is configured to determine an initial clustering centroid according to the line loss rate box plot.

[0096] The first clustering unit 305 is configured to cluster the line loss rate time series of the plurality of transformer areas to be classified by using a first clustering algorithm and the initial clustering centroid, and obtain a first clustering set and a second clustering set, the first clustering set corresponds to a first clustering centroid, the second clustering set corresponds to a second clustering centroid, and the line loss rate mean value corresponding to the first clustering centroid is greater than the line loss rate mean value corresponding to the second clustering centroid.

[0097] The division unit 306 is configured to divide the transformer area corresponding to the line loss rate time series included in the first clustering set into a high line loss transformer area.

[0098] Optionally, the initial clustering centroid comprises a first initial clustering centroid and a second initial clustering centroid, and the first determination unit 304 is specifically configured to:

[0099] If there is an abnormal point greater than the upper edge of the line loss rate box plot in the line loss rate box plot, the line loss rate time series corresponding to the abnormal point with the smallest value is determined as the first initial clustering centroid.

[0100] If there is no abnormal point greater than the upper edge of the line loss rate box plot in the line loss rate box plot, the line loss rate time series corresponding to the point with the largest value in the line loss rate box plot is determined as the first initial clustering centroid.

[0101] determine a line loss rate time sequence corresponding to the median point in the line loss rate box plot as a second initial clustering centroid.

[0102] Optionally, the apparatus further comprises:

[0103] a second clustering unit, configured to cluster the first clustering set by a second clustering algorithm to obtain a plurality of first clustering sub-sets;

[0104] a second determination unit, configured to determine a line loss rate time sequence to be detected corresponding to a centroid of each first clustering sub-set;

[0105] an analysis unit, configured to analyze each line loss rate time sequence to be detected to obtain feature data corresponding to each line loss rate time sequence to be detected;

[0106] a second construction unit, configured to construct a line loss rate transformer area portrait corresponding to each first clustering sub-set according to the feature data corresponding to each line loss rate time sequence to be detected.

[0107] Optionally, the second clustering algorithm comprises a K-means improved clustering algorithm, and the second clustering unit is specifically configured to:

[0108] calculate a silhouette coefficient according to each line loss rate time sequence included in the first clustering set;

[0109] determine a clustering number according to the silhouette coefficient;

[0110] cluster the first clustering set by the K-means improved clustering algorithm and the clustering number to obtain the plurality of first clustering sub-sets.

[0111] Optionally, the analysis unit is specifically configured to:

[0112] test each line loss rate time sequence to be detected by a Mann-Kendall method to obtain a line loss rate trend of each line loss rate time sequence to be detected;

[0113] calculate each line loss rate time sequence to be detected to obtain a variance and a mean value of each line loss rate time sequence to be detected.

[0114] It can be seen that the embodiment of the application provides a high line loss transformer area division apparatus, which can construct a line loss rate box plot of a transformer area to be classified, and divide the high line loss transformer area based on a clustering algorithm, can comprehensively consider the operation characteristics of the transformer area to be classified and the relevance between the transformer areas to be classified when dividing the high line loss transformer area, thereby improving the division accuracy of the high line loss transformer area, and further improving the management level of the power supply enterprise in line loss management.

[0115] The embodiment of the application further provides a computer device, comprising a memory, a processor and a bus system.

[0116] The memory is configured to store a program.

[0117] The processor is configured to execute the program in the memory to implement any one of the high line loss power grid partitioning methods described above.

[0118] The bus system is configured to connect the memory and the processor to enable the memory and the processor to communicate.

[0119] The embodiments of the present application also provide a computer readable storage medium storing instructions, which, when executed on a computer, cause the computer to perform any one of the high line loss power grid partitioning methods described above.

[0120] Finally, it needs to be explained that, in this document, relationship terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply that there is any such actual relationship or order between these entities or operations. Moreover, the terms "include", "contain" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, article or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or device. Without more limitations, the element defined by the statement "including a" does not exclude the presence of other identical elements in the process, method, article or device including the element.

[0121] The above description of the disclosed embodiments enables a person skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to these embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A high line loss area division method, characterized by, The method comprises: obtaining line loss rate time series of a plurality of to-be-classified transformer areas; calculating the line loss rate time series of the plurality of to-be-classified transformer areas respectively to obtain a line loss rate mean value of each to-be-classified transformer area; constructing a line loss rate box plot according to the line loss rate mean value of each to-be-classified transformer area; determining an initial clustering centroid according to the line loss rate box plot; performing clustering on the line loss rate time series of the plurality of to-be-classified transformer areas through a first clustering algorithm and the initial clustering centroid to obtain a first clustering set and a second clustering set, the first clustering set corresponding to a first clustering centroid, the second clustering set corresponding to a second clustering centroid, the line loss rate mean value corresponding to the first clustering centroid being greater than the line loss rate mean value corresponding to the second clustering centroid; dividing the to-be-classified transformer areas corresponding to the line loss rate time series included in the first clustering set into high line loss transformer areas; the initial clustering centroid comprises a first initial clustering centroid and a second initial clustering centroid, and the determining of the initial clustering centroid according to the line loss rate box plot comprises: if there is an abnormal point greater than the upper edge of the line loss rate box plot in the line loss rate box plot, determining the line loss rate time series corresponding to the abnormal point with the smallest value as the first initial clustering centroid; if there is no abnormal point greater than the upper edge of the line loss rate box plot in the line loss rate box plot, determining the line loss rate time series corresponding to the point with the largest value in the line loss rate box plot as the first initial clustering centroid; determining the line loss rate time series corresponding to the median point in the line loss rate box plot as the second initial clustering centroid.

2. The method of claim 1, wherein, The method further comprises: performing clustering on the first clustering set through a second clustering algorithm to obtain a plurality of first clustering sub-sets; determining a to-be-detected line loss rate time series corresponding to the centroid of each first clustering sub-set; analyzing each to-be-detected line loss rate time series to obtain feature data corresponding to each to-be-detected line loss rate time series; constructing a line loss rate transformer area image corresponding to each first clustering sub-set according to the feature data corresponding to each to-be-detected line loss rate time series respectively.

3. The method of claim 2, wherein, The second clustering algorithm comprises a K-means improved clustering algorithm, and the performing of clustering on the first clustering set through the second clustering algorithm to obtain a plurality of first clustering sub-sets comprises: calculating a contour coefficient according to each line loss rate time series included in the first clustering set; determining a clustering number according to the contour coefficient; performing clustering on the first clustering set through the K-means improved clustering algorithm and the clustering number to obtain a plurality of first clustering sub-sets.

4. The method of claim 2, wherein, The analyzing of each to-be-detected line loss rate time series to obtain feature data corresponding to each to-be-detected line loss rate time series comprises: performing testing on each to-be-detected line loss rate time series through the Mann-Kendall method respectively to obtain a line loss rate trend of each to-be-detected line loss rate time series; calculating each to-be-detected line loss rate time series to obtain a variance and a mean value of each to-be-detected line loss rate time series.

5. A high line loss area division apparatus, characterized by comprising: The device comprises: an obtaining unit, configured to obtain line loss rate time series of a plurality of to-be-classified transformer areas; The computing unit is configured to calculate line loss rate time series of the plurality of to-be-classified transformer areas respectively, and obtain a line loss rate mean value of each to-be-classified transformer area; The first constructing unit is configured to construct a line loss rate box plot according to the line loss rate mean value of each to-be-classified transformer area; The first determining unit is configured to determine an initial clustering centroid according to the line loss rate box plot; The first clustering unit is configured to cluster the line loss rate time series of the plurality of to-be-classified transformer areas by using a first clustering algorithm and the initial clustering centroid, and obtain a first clustering set and a second clustering set, the first clustering set corresponding to a first clustering centroid, and the second clustering set corresponding to a second clustering centroid, the line loss rate mean value corresponding to the first clustering centroid being greater than the line loss rate mean value corresponding to the second clustering centroid; The dividing unit is configured to divide the to-be-classified transformer area corresponding to the line loss rate time series included in the first clustering set into a high line loss transformer area. The initial clustering centroid includes a first initial clustering centroid and a second initial clustering centroid, and the first determining unit is specifically configured to: If there is an abnormal point greater than the upper edge of the line loss rate box plot in the line loss rate box plot, determine the line loss rate time series corresponding to the abnormal point with the smallest value as the first initial clustering centroid; If there is no abnormal point greater than the upper edge of the line loss rate box plot in the line loss rate box plot, determine the line loss rate time series corresponding to the point with the largest value in the line loss rate box plot as the first initial clustering centroid; Determine the line loss rate time series corresponding to the median point in the line loss rate box plot as the second initial clustering centroid.

6. The apparatus of claim 5, wherein, The device further includes: The second clustering unit is configured to cluster the first clustering set by using a second clustering algorithm, and obtain a plurality of first clustering sub-sets; The second determining unit is configured to determine a to-be-detected line loss rate time series corresponding to a centroid of each first clustering sub-set; The analysis unit is configured to analyze each to-be-detected line loss rate time series, and obtain feature data corresponding to each to-be-detected line loss rate time series; The second constructing unit is configured to construct a line loss rate transformer area image corresponding to each first clustering sub-set according to the feature data corresponding to each to-be-detected line loss rate time series.

7. A computer device, comprising: It includes: A memory, a processor, and a bus system; The memory is configured to store a program; The processor is configured to execute the program in the memory to implement the method in any one of claims 1 to 4; The bus system is configured to connect the memory and the processor to enable the memory and the processor to communicate.

8. A computer-readable storage medium, characterized in that, The instructions stored in the computer enable the computer to execute the method in any one of claims 1 to 4.

Images