A method and device for detecting abnormal line loss in a transformer area

Abstract

The present application discloses a method and device for detecting abnormal line loss in a substation, wherein the method includes: obtaining the line loss rate of the target substation every day during the inspection period; calculating the operating indicators corresponding to the target substation based on the line loss rate of the target substation; the operating indicators include trend indicators, difference indicators, volatility indicators and ratio indicators; inputting the operating indicators corresponding to the target substation into a trained classification model to obtain the line loss status corresponding to the target substation; the classification model is pre-trained using training samples of multiple categories and their corresponding line loss status; the training samples of one category include the operating indicators corresponding to the sample substations of the category; each category is clustered based on the operating indicators to obtain each sample substation; the line loss status corresponding to the target substation and the line loss factor corresponding to the line loss status are output; the line loss status corresponding to each category and its corresponding line loss factor are pre-analyzed for the operating indicators corresponding to each sample substation of each category.

Description

Technical Field

[0001] The present application relates to the technical field of line loss anomaly detection, and in particular to a method and device for detecting line loss anomaly in a transformer substation. Background Art

[0002] Line loss can comprehensively reflect the operating status of the power grid, the level of grid management and operation, the level of lean management of substations, and the level of energy conservation and loss reduction, so it is very important to accurately analyze the line loss situation.

[0003] Currently, when performing abnormal line loss analysis in a substation, the line loss rate of the substation is obtained, and then analysis is performed directly based on the level of the line loss rate, such as determining the degree of line loss, abnormal lines, substations, or large users.

[0004] However, due to the particularity of line loss rate, the line loss rate in different areas is closely related to various links such as line design, infrastructure, production, dispatching, substation operation, line maintenance, and monthly meter reading, verification, and collection. For example, in most substations with residential electricity load as the main load, line loss fluctuations are obviously related to changes in electricity consumption, and the fluctuations are large. Therefore, the reasonable range and fluctuation of line loss in each substation are also different. Areas with overall lower line loss rates do not necessarily have better line loss management than areas with high line loss rates. Therefore, the one-size-fits-all approach of directly analyzing only by line loss rate does not conform to the diverse on-site conditions, which makes the analysis results inaccurate. Summary of the Invention

[0005] Based on the above-mentioned deficiencies in the prior art, the present application provides a method and device for detecting abnormal line loss in a substation area to solve the problem that the existing technology cannot accurately analyze the line loss status in the substation area.

[0006] In order to achieve the above objectives, this application provides the following technical solutions:

[0007] The first aspect of the present application provides a method for detecting abnormal line loss in a transformer area, comprising:

[0008] Obtain the daily line loss rate of the target area during the inspection period;

[0009] Calculating the operating indicators corresponding to the target substation based on the daily line loss rate of the target substation during the inspection period; wherein the operating indicators include trend indicators, difference indicators, volatility indicators, and ratio indicators;

[0010] Inputting the operating indicators corresponding to the target substation into a pre-trained classification model to obtain the line loss status corresponding to the target substation; wherein the classification model is pre-trained using training samples of multiple categories and the line loss status corresponding to each category; the training samples of one category include the operating indicators corresponding to each sample substation belonging to the category; each category is obtained by clustering each sample substation based on the operating indicators corresponding to each sample substation; and the line loss status corresponding to each category is obtained by analyzing the operating indicators corresponding to each sample substation belonging to the category;

[0011] The line loss status corresponding to the target substation and the line loss factor corresponding to the line loss status are output; wherein the line loss factor corresponding to the line loss status is obtained in advance based on the analysis of the operating indicators corresponding to each of the sample substations belonging to the category corresponding to the line loss status.

[0012] Optionally, in the above-mentioned method for detecting abnormal line loss in a substation, after obtaining the daily line loss rate of the target substation during the inspection period, the method further includes:

[0013] Determining whether any of the line loss rates is a negative value or a null value within the inspection period; wherein, if it is determined whether the line loss rate is not a negative value or a null value within the inspection period, calculating the operating indicator corresponding to the target substation based on the daily line loss rate of the target substation within the inspection period;

[0014] If it is determined that any of the line loss rates within the inspection period is negative or null, then a feedback is given that the target substation is abnormal.

[0015] Optionally, in the above-mentioned method for detecting abnormal line loss in a substation, the operation index corresponding to the target substation is calculated based on the daily line loss rate of the target substation during the inspection period, including:

[0016] Calculate the trend index corresponding to the target substation based on the daily line loss rate of the target substation during the inspection period; wherein the trend index includes an upward trend index, a downward trend index, and a monthly change rate;

[0017] Based on the daily line loss rate of the target substation during the inspection period, calculate the difference index corresponding to the target substation; wherein the difference index includes the range, the interquartile range, the mean difference between the front part of the data and the back part of the data, the difference between the mean of the segmented data and the overall mean, the modulus of the Fourier coefficient difference between the front part of the data and the back part of the data, and the fitting slope; the front part of the data and the back part of the data are the line loss rates of the two parts obtained by evenly dividing the respective line loss rates based on the time sequence;

[0018] Based on the daily line loss rate of the target substation during the inspection period, a volatility index corresponding to the target substation is calculated; wherein the volatility index includes the standard deviation of the overall data, the standard deviation of the front part of the data, the standard deviation of the back part of the data, and the coefficient of variation; the coefficient of variation is equal to the ratio of the standard deviation of the overall data to the mean of all the line loss rates;

[0019] Based on the daily line loss rate of the target substation during the inspection period, the ratio index corresponding to the target substation is calculated; wherein, the ratio index includes the ratio of the mean of the front part of the data to the total mean, the ratio of the mean of the rear part of the data to the total mean, and the correlation coefficient; the correlation coefficient refers to the correlation coefficient between the mean of each line loss rate of the target substation and the average line loss rate of similar substations.

[0020] Optionally, in the above-mentioned method for detecting abnormal line loss in a substation, calculating the trend indicator corresponding to the target substation based on the daily line loss rate of the target substation during the inspection period includes:

[0021] Sort the daily line loss rates of the target substation area during the inspection period in chronological order to obtain a line loss rate sequence of the target substation area;

[0022] For each target line loss rate in the line loss rate sequence, the mean of the moving data corresponding to the target line loss rate is calculated to obtain a moving average value corresponding to the target line loss rate; wherein the target line loss rate is the line loss rate other than the first and last line loss rates in the line loss rate sequence; and the moving data corresponding to one target line loss rate is a maximum number of no more than n line loss rates selected forward from the target line loss rate as a starting point;

[0023] Calculating the mean of the moving average values ​​corresponding to the target line loss rates to obtain the moving average value corresponding to the line loss rate sequence;

[0024] The rising trend indicator corresponding to the target station area is obtained by dividing the sum of the squares of the line loss rates in the line loss rate sequence that are smaller than the moving average value corresponding to the line loss rate sequence by the number of data segments located below the moving average value corresponding to the line loss rate sequence;

[0025] The downward trend indicator corresponding to the target station area is obtained by dividing the sum of the squares of the line loss rates in the line loss rate sequence that are greater than the moving average value corresponding to the line loss rate sequence by the number of data segments located on the moving average value corresponding to the line loss rate sequence;

[0026] The change rate of the mean value of each line loss rate in every two adjacent months is calculated respectively to obtain the monthly change rate corresponding to the target substation.

[0027] Optionally, in the above-mentioned method for detecting abnormal line loss in a substation area, calculating the difference index corresponding to the target substation area based on the daily line loss rate of the target substation area during the inspection period includes:

[0028] Calculating the difference between the maximum and minimum values ​​of each of the line loss rates to obtain the range corresponding to the target area;

[0029] Calculating the difference between the upper quartile and the lower quartile of the line loss rate sequence composed of the line loss rates sorted in time, to obtain the interquartile range corresponding to the target station area;

[0030] Calculating the mean difference between the first part of the data and the second part of the data;

[0031] Divide the line loss rate sequence into four segments on average, and subtract the mean of all line loss rates from the sum of the mean of the line loss rates in each segment to obtain the difference between the mean of the segment data and the total mean;

[0032] Calculating the modulus of the difference between the Fourier coefficients of the front portion of data and the rear portion of data;

[0033] Perform linear fitting on each of the line loss rates to obtain a fitting slope corresponding to the target station area.

[0034] Optionally, in the above-mentioned method for detecting abnormal line loss in a transformer area, the training method of the classification model includes:

[0035] Obtaining daily line loss rates of the plurality of sample areas during an inspection period;

[0036] Calculating the operating index corresponding to each sample area based on the line loss rate of each sample area on each day during the inspection period;

[0037] Clustering each of the sample areas based on the operation indicators corresponding to each of the sample areas to obtain a plurality of categories;

[0038] For each of the categories, analyzing the operating indicators corresponding to the sample substations belonging to the category, and obtaining the line loss status corresponding to the category and the line loss factor corresponding to the line loss status;

[0039] respectively taking the line loss status corresponding to each category as a label corresponding to each sample station area belonging to the category;

[0040] The classification model is iteratively trained using the operating indicators and labels corresponding to each of the sample stations.

[0041] Optionally, in the above-mentioned method for detecting abnormal line loss in a substation, clustering the sample substations based on the operation indicators corresponding to the sample substations to obtain a plurality of categories includes:

[0042] The density peak algorithm CFSFDP is used to calculate the operation index corresponding to each of the sample stations to obtain multiple initial cluster centers;

[0043] The K-means clustering algorithm is used to cluster the sample areas based on the multiple initial cluster centers and the operating indicators corresponding to the sample areas to obtain the multiple categories.

[0044] Optionally, in the above-mentioned substation line loss anomaly detection method, the classification model includes multiple classification sub-models, and the iterative training of the classification model using the operating indicators and corresponding labels corresponding to each of the sample substations includes:

[0045] Combining every two of the categories to obtain multiple combinations;

[0046] The operation indicators and labels corresponding to the sample stations under the two categories in each combination are used to train a classification sub-model to obtain the trained classification sub-models.

[0047] Optionally, in the above-mentioned method for detecting abnormal line loss in a substation area, inputting the operating indicator corresponding to the target substation area into a pre-trained classification model to obtain the line loss status corresponding to the target substation area includes:

[0048] Outputting the operating indicators corresponding to the target substation to each of the pre-trained classification sub-models to obtain the line loss status output by each of the classification sub-models;

[0049] The line loss state with the largest proportion among the line loss states output by each of the classification sub-models is determined as the line loss state corresponding to the target substation.

[0050] A second aspect of the present application provides a device for detecting abnormal line loss in a transformer area, comprising:

[0051] The first acquisition unit is used to obtain the line loss rate of the target area every day during the inspection period;

[0052] A first calculation unit is configured to calculate an operating index corresponding to the target substation based on the daily line loss rate of the target substation during an inspection period; wherein the operating index includes a trend index, a difference index, a volatility index, and a ratio index;

[0053] An input unit is configured to input the operating indicators corresponding to the target substation into a pre-trained classification model to obtain the line loss status corresponding to the target substation; wherein the classification model is pre-trained using training samples of multiple categories and the line loss status corresponding to each category; the training samples of one category include the operating indicators corresponding to each sample substation belonging to the category; each category is obtained by clustering each sample substation based on the operating indicators corresponding to each sample substation; and the line loss status corresponding to each category is obtained by analyzing the operating indicators corresponding to each sample substation belonging to the category.

[0054] An output unit is used to output the line loss status corresponding to the target substation and the line loss factor corresponding to the line loss status; wherein the line loss factor corresponding to the line loss status is obtained in advance based on the analysis of the operating indicators corresponding to each of the sample substations belonging to the category corresponding to the line loss status.

[0055] The present invention provides a method for detecting abnormal line loss in a substation. The method pre-clustering multiple sample substations based on their corresponding operating indicators yields multiple categories. The method then analyzes the operating indicators of each sample substation within each category to obtain the corresponding line loss status and the corresponding line loss factors. Finally, a classification model is trained using the training samples from the multiple categories and the corresponding line loss status. Therefore, when analyzing a target substation, the daily line loss rate for the target substation during an inspection period is obtained. Based on the daily line loss rate for the target substation during the inspection period, the corresponding operating indicators are calculated. These operating indicators include trend indicators, variance indicators, volatility indicators, and ratio indicators, which can more comprehensively reflect the line loss situation under different on-site conditions. The operating indicators corresponding to the target substation are then input into the pre-trained classification model to obtain the corresponding line loss status of the target substation, thereby determining the category to which the target substation belongs. Finally, the corresponding line loss status and the corresponding line loss factors are output, thereby accurately analyzing the substation's line loss status and its causes. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are merely embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without any creative work.

[0057] Figure 1 A flow chart of a method for detecting abnormal line loss in a transformer area provided in an embodiment of the present application;

[0058] Figure 2 A flowchart of calculating the operating index corresponding to the target area provided in an embodiment of the present application;

[0059] Figure 3 A flowchart of calculating a trend indicator corresponding to a target area provided in an embodiment of the present application;

[0060] Figure 4 A flowchart of calculating a difference index corresponding to a target area provided in an embodiment of the present application;

[0061] Figure 5 A flowchart of a classification model training method provided in an embodiment of the present application;

[0062] Figure 6 A training diagram of a classification model provided in an embodiment of the present application;

[0063] Figure 7 A flow chart of a method for determining the line loss status corresponding to a target substation provided in an embodiment of the present application;

[0064] Figure 8 A schematic structural diagram of a device for detecting abnormal line loss in a transformer substation provided in an embodiment of the present application. DETAILED DESCRIPTION

[0065] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0066] In this application, relational terms such as first and second, etc., are used only 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 "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus comprising the element.

[0067] The present application embodiment provides a method for detecting abnormal line loss in a transformer area, such as Figure 1 As shown, the following steps are included:

[0068] S101. Obtain the daily line loss rate of the target substation during the inspection period.

[0069] The daily line loss rate is the quotient obtained by dividing the difference between the daily electricity sales and the daily power supply by the daily power supply.

[0070] Optionally, in order to ensure that the data of the target substation can be used for subsequent calculations, in another embodiment of the present application, after executing step S101, line loss anomaly is further determined based on the grid rules, that is, whether the line loss is abnormal. Specifically, after executing step S101, the following may be further included:

[0071] Determine whether any line loss rate is negative or null during the inspection period.

[0072] If no line loss rate is negative or null during the inspection period, it indicates that the line loss in the target substation is normal, and step S102 is executed. If any line loss rate is negative or null during the inspection period, the target substation is abnormal.

[0073] It's important to note that when errors occur in the substation file relationships for certain users, a user's meter that doesn't belong to a normal substation can be recorded in the file. This can cause the substation's electricity consumption to exceed the energy value recorded by the gateway meter, resulting in negative line loss. Furthermore, when the gateway meter collection device malfunctions, substation electricity consumption data cannot be obtained, resulting in a null line loss rate. Therefore, substations with negative or null line loss rates during the observation period are included in the abnormal list.

[0074] S102. Calculate the operating index corresponding to the target substation based on the daily line loss rate of the target substation during the inspection period.

[0075] The operating indicators include trend indicators, difference indicators, volatility indicators, and ratio indicators. It should be noted that directly analyzing line loss using the line loss rate is not accurate. Therefore, in this application, various operating indicators are further calculated based on the line loss rate to facilitate subsequent analysis based on the operating indicators.

[0076] It should be noted that trend indicators reflect the changing trends in line loss rates. Variation indicators refer to the differences in line loss rates over different time periods. Volatility indicators reflect fluctuations in line loss rates. Ratio indicators are other indicators used to compare line loss rates through ratios. The specific indicators included in each category can be customized based on your needs.

[0077] Optionally, in another embodiment of the present application, an implementation of step S102 is as follows: Figure 2As shown, the following steps are included:

[0078] S201. Calculate a trend indicator corresponding to the target substation based on the daily line loss rate of the target substation during the inspection period.

[0079] Among them, trend indicators include upward trend indicators, downward trend indicators and monthly change rates.

[0080] First of all, it should be noted that the execution order of each step in the embodiment of the present application is only one optional method. Since the calculation of each type of operating indicator is completely independent, other execution orders can also be used to calculate each type of operating indicator, which should all fall within the protection scope of this application.

[0081] Alternatively, as Figure 3 As shown, an implementation of step S201 includes the following steps:

[0082] S301. For each target line loss rate in the line loss rate sequence, calculate the mean of the moving data corresponding to the target line loss rate to obtain a moving average value corresponding to the target line loss rate.

[0083] Among them, the target line loss rate is the other line loss rates except the first and last line loss rates in the line loss rate sequence. The moving data corresponding to a target line loss rate is the maximum number of line loss rates not greater than n selected forward from the line loss rate, that is, for a line loss rate, if the number of line loss rates before the line loss rate is greater than or equal to n, then the nth line loss rate is selected forward. If the number of line loss rates before the line loss rate is less than n, then all line loss rates before the line loss rate are selected. Therefore, the moving average corresponding to a target line loss rate is the average of the multiple line loss rates before it, which can be specifically expressed as:

[0084]

[0085] Wherein, 2≤t≤n-1; r represents the line loss rate.

[0086] S302: Calculate the mean of the moving average values ​​corresponding to each target line loss rate to obtain the moving average value corresponding to the line loss rate sequence.

[0087] S303: Divide the sum of the squares of the line loss rates in the line loss rate sequence that are smaller than the moving average value corresponding to the line loss rate sequence by the number of data segments below the moving average value corresponding to the line loss rate sequence to obtain an upward trend indicator corresponding to the target substation.

[0088] Among them, the number of data segments located under the moving average value corresponding to the line loss rate sequence refers to the number of curve segments on the straight line corresponding to the moving average value corresponding to the line loss rate sequence in the curve corresponding to the line loss rate sequence when the line loss rate sequence and the moving average value corresponding to the line loss rate sequence are plotted on the same graph.

[0089] If the line loss rate is less than the moving average value corresponding to the line loss rate sequence, it will rise subsequently, so it is used to reflect the upward trend of the line loss rate.

[0090] S304: Divide the sum of the squares of the line loss rates in the line loss rate sequence that are greater than the moving average value corresponding to the line loss rate sequence by the number of data segments located on the moving average value corresponding to the line loss rate sequence to obtain a downward trend indicator corresponding to the target substation.

[0091] S305. Calculate the change rate of the mean value of each line loss rate in every two adjacent months to obtain the monthly change rate corresponding to the target substation.

[0092] S202. Calculate the difference index corresponding to the target substation based on the daily line loss rate of the target substation during the inspection period.

[0093] Among them, the difference indicators include range, interquartile range, mean difference between the front part of the data and the back part of the data, difference between the mean of the segmented data and the total mean, modulus of the Fourier coefficient difference between the front part of the data and the back part of the data, and fitting slope.

[0094] The first part of the data and the second part of the data are obtained by evenly dividing the line loss rates based on the time sequence. That is, if the line loss rates for n days are obtained, the first part of the data refers to the line loss rates for the first n / 2 days, and the second part of the data refers to the line loss rates for the last 2 / n days.

[0095] Alternatively, as Figure 4 As shown, an implementation of step S202 includes the following steps:

[0096] S401. Calculate the difference between the upper quartile and the lower quartile of a line loss rate sequence composed of line loss rates sorted by time, and obtain the interquartile range corresponding to the target substation.

[0097] It should be noted that quartiles, also known as quartile points, refer to the values ​​at the three dividing points in statistics when all values ​​are arranged from small to large and divided into four equal parts. The quartiles usually refer to the values ​​in the top 25%, called the lower quartile, and the values ​​in the 75% position, called the upper quartile.

[0098] S402: Calculate the mean difference between the first part of the data and the second part of the data.

[0099] Specifically, the mean of the first part of the data and the mean of the second part of the data are calculated, and then the mean of the first part of the data is subtracted from the mean of the second part of the data to obtain the mean difference between the two.

[0100] S403 : Divide the line loss rate sequence into four segments on average, and subtract the mean of all line loss rates from the sum of the mean of the line loss rates in each segment to obtain the difference between the segment data mean and the total mean.

[0101] S404: Calculate the modulus of the difference between the Fourier coefficients of the front part of the data and the rear part of the data.

[0102] Specifically, discrete Fourier transform is performed on the front data and the back data respectively to obtain the Fourier transform coefficient sequence of the front data and the Fourier transform coefficient sequence of the back data, and then the sum of the scores of the difference between the values ​​in the Fourier transform coefficient sequence of the front data and the Fourier transform coefficient sequence of the back data is calculated, and finally the modulus is taken to obtain the modulus of the Fourier coefficient difference between the front data and the back data, which can be specifically expressed as:

[0103]

[0104] Where n represents the total number of days in the observation period, i represents the i-th day; y1 represents the value in the Fourier change coefficient sequence of the first part of the data; y2 represents the value in the Fourier change coefficient sequence of the second part of the data.

[0105] S405: Perform linear fitting on each line loss rate to obtain a fitting slope corresponding to the target substation area.

[0106] S203: Calculate the volatility index corresponding to the target substation based on the daily line loss rate of the target substation during the inspection period.

[0107] The volatility index includes the standard deviation of the overall data (i.e., the standard deviation of all line loss rates), the standard deviation of the first part of the data, the standard deviation of the second part of the data, and the coefficient of variation. The coefficient of variation is equal to the ratio of the standard deviation of the overall data to the mean of all line loss rates.

[0108] S204. Calculate a ratio index corresponding to the target substation based on the daily line loss rate of the target substation during the inspection period.

[0109] Among them, the ratio indicators include the ratio of the mean of the first part of the data to the total mean, the ratio of the mean of the second part of the data to the total mean, and the correlation coefficient.

[0110] The correlation coefficient refers to the correlation coefficient between the mean of the line loss rates of the target substation and the average line loss rate of similar substations. Specifically, substations that belong to the same power supply station as the target substation and have similar electricity consumption types can be selected as similar substations to the target substation. The correlation coefficient between the target substation and its similar substations is calculated based on the average line loss rate of the similar substations.

[0111] S103: Input the operating indicators corresponding to the target substation into a pre-trained classification model to obtain the line loss status corresponding to the target substation.

[0112] The classification model is pre-trained using training samples from multiple categories and the line loss status corresponding to each category. The training samples for a category include the operating indicators corresponding to each sample substation belonging to the category. Each category is clustered based on the operating indicators corresponding to each sample substation. The line loss status corresponding to each category is analyzed based on the operating indicators corresponding to each sample substation belonging to the category. Therefore, each category corresponds to a line loss status, meaning that the line loss status is equivalent to the label or category name of the corresponding category. Therefore, the line loss status corresponding to the target substation is recorded as the category of the target substation.

[0113] Since the categories are divided based on multiple operating indicators, it can effectively ensure that the on-site conditions of the substations under a category are the same or similar, so the line loss situation of the target substation is analyzed by the category to which it belongs.

[0114] Specifically, such as Figure 5 As shown, the training method of the classification model includes the following steps:

[0115] S501. Obtain the daily line loss rates of multiple sample substations during an inspection period.

[0116] Optionally, after obtaining the daily line loss rates of multiple sample substations during the inspection period, it is also determined for the sample substations whether there are any negative or null line loss rates among the sample substations. If so, the sample substation is eliminated.

[0117] S502: Calculate the operation index corresponding to each sample substation based on the daily line loss rate of each sample substation during the inspection period.

[0118] It should be noted that the operating indicators corresponding to the sample substation are of the same type as the operating indicators corresponding to the target substation. Therefore, the specific calculation process of the operating indicators corresponding to the sample substation can refer to the calculation process of the operating indicators corresponding to the target substation provided above, and will not be repeated here.

[0119] S503: Clustering each sample substation based on the operation index corresponding to each sample substation to obtain multiple categories.

[0120] Specifically, based on the operating indicators corresponding to each substation, the similarity of each sample substation can be calculated, and then similar sample substations can be grouped into one category.

[0121] Optionally, an existing clustering algorithm may be used to cluster the sample areas.

[0122] Optionally, in another embodiment of the present application, a specific implementation of step S503 includes: first using the density peak algorithm CFSFDP to calculate the operating indicators corresponding to each sample station to obtain multiple initial cluster centers, and then using the clustering algorithm K-means based on the multiple initial cluster centers and the operating indicators corresponding to each sample station to cluster each sample station to obtain multiple categories.

[0123] It should be noted that the K-means clustering algorithm has high requirements for the number of clusters, and the selection of the initial cluster centers directly affects the clustering effect. Improper selection leads to unstable clustering results. To address the above issues, the embodiment of this application first uses the CFSFDP algorithm to determine the number of clusters and the location of the initial cluster centers, and then uses the K-means method for clustering, thereby effectively ensuring the clustering effect.

[0124] S504: For each category, analyze the operating indicators corresponding to each sample substation belonging to the category to obtain the line loss status corresponding to the category and the line loss factor corresponding to the line loss status.

[0125] Among them, the factors affecting line loss mainly include: incorrect user relationship, metering device failure, metering error, unmetered electricity consumption, abnormal electricity consumption, etc. Different influencing factors correspond to different line loss conditions.

[0126] Optionally, the categories whose coefficient of variation corresponding to the cluster center in each category is greater than a preset threshold can be used as an abnormal list, and those that are not greater than the preset threshold can be used as a normal list. Then, the corresponding line loss status can be further determined according to the specific linear situation and line loss factors, for example, normal 1, normal 2, abnormal 1, abnormal 2, etc.

[0127] S505: Use the line loss status corresponding to each category as a label corresponding to each of the sample stations belonging to the category.

[0128] S506: Iteratively train the classification model using the operating indicators and labels corresponding to each sample station area.

[0129] Specifically, the various operating indicators corresponding to each sample substation are connected to form the sample features corresponding to the sample substation. The sample features of the sample substation are then input into the classification model, and the line loss status corresponding to the sample substation is output. The line loss status corresponding to the sample substation is then compared with its corresponding label. If the error between the two is greater than the preset error, the parameters of the classification model are adjusted based on the error, and feedback is used for retraining until the error between the two is no greater than the preset error.

[0130] Optionally, a support vector machine learning algorithm may be used to train the classification model, and of course other algorithms may also be used for training.

[0131] Optionally, in another embodiment of the present application, the classification model includes multiple classification sub-models. Accordingly, a specific implementation of step S506 includes:

[0132] Combine every two categories to obtain multiple combinations, and use the operating indicators and corresponding labels corresponding to each sample area under the two categories in each combination to train a classification sub-model to obtain trained classification sub-models.

[0133] In the embodiment of the present application, the classification model is composed of multiple classification sub-models, and the number of classification sub-models is equal to the number of combinations of each category. Figure 6 As shown, a classification sub-model is trained with the sample data under each combination, and finally the structures of each classification sub-model are integrated to obtain the final result. Specifically, the final prediction result value can be output by majority voting.

[0134] For example, if there are K categories, the data from any two of the K categories are combined, and then a classification sub-model is trained using the data from each combination, thereby generating K(K-1) / 2 classification sub-models. The final output can be obtained by fusing the results of these classification sub-models. Specifically, the prediction results of each classification sub-model can be output as the final prediction result value by majority voting. That is, in the embodiment of the present application, a one-to-one strategy is used for model training.

[0135] Accordingly, a specific implementation of step S103 is as follows: Figure 7 As shown, the following steps are included:

[0136] S701: Output the operating indicators corresponding to the target substation to each pre-trained classification sub-model to obtain the line loss status output by each classification sub-model.

[0137] S702: Determine the line loss state with the largest proportion among the line loss states output by each classification sub-model as the line loss state corresponding to the target substation.

[0138] S104: Output the line loss status corresponding to the target substation and the line loss factor corresponding to the line loss status.

[0139] The line loss factor corresponding to the line loss state is pre-derived based on an analysis of the operating indicators corresponding to each sample substation within the category corresponding to the line loss state. As can be seen from step S504 above, when training the model, the operating indicators corresponding to each sample substation within each category are analyzed to obtain the category's line loss factor and line loss state. Therefore, each category corresponds to a line loss state and a corresponding line loss factor. Finally, the line loss state and the corresponding line loss factor are output to the user to further analyze the cause of the line loss and provide guidance for on-site inspection and verification by maintenance personnel, thereby guiding the power department to identify and resolve abnormal line loss issues in substations.

[0140] The present invention provides a method for detecting abnormal line loss in a substation. The method pre-clustering multiple sample substations based on their corresponding operating indicators yields multiple categories. The method then analyzes the operating indicators of each sample substation within each category to obtain the corresponding line loss status and the corresponding line loss factors. Finally, a classification model is trained using the training samples from the multiple categories and the corresponding line loss status. Therefore, when analyzing a target substation, the daily line loss rate for the target substation during an inspection period is obtained. Based on the daily line loss rate for the target substation during the inspection period, the corresponding operating indicators are calculated. These operating indicators include trend indicators, variance indicators, volatility indicators, and ratio indicators, which can more comprehensively reflect the line loss situation under different on-site conditions. The operating indicators corresponding to the target substation are then input into the pre-trained classification model to obtain the corresponding line loss status of the target substation, thereby determining the category to which the target substation belongs. Finally, the corresponding line loss status and the corresponding line loss factors are output, thereby accurately analyzing the substation's line loss status and its causes.

[0141] Another embodiment of the present application provides a device for detecting abnormal line loss in a transformer area, such as Figure 8 Shown, including:

[0142] The first acquiring unit 801 is configured to acquire the line loss rate of the target substation every day during the inspection period.

[0143] The first calculation unit 802 is configured to calculate an operation index corresponding to the target substation based on the daily line loss rate of the target substation during the inspection period.

[0144] Among them, operating indicators include trend indicators, difference indicators, volatility indicators and ratio indicators.

[0145] The input unit 803 is used to input the operating index corresponding to the target substation into a pre-trained classification model to obtain the line loss status corresponding to the target substation.

[0146] The classification model is pre-trained using training samples from multiple categories and the corresponding line loss status for each category. Training samples for a category include the operating indicators corresponding to each sample substation area within that category. Each category is clustered based on the corresponding operating indicators of each sample substation area. The line loss status for each category is determined by analyzing the corresponding operating indicators of each sample substation area within that category.

[0147] The output unit 804 is used to output the line loss status corresponding to the target substation and the line loss factor corresponding to the line loss status.

[0148] The line loss factors corresponding to the line loss status are obtained in advance by analyzing the operation indicators corresponding to each sample substation belonging to the category corresponding to the line loss status.

[0149] It should be noted that the specific working process of each unit provided in the above embodiments of the present application can refer to the corresponding steps in the above method embodiments, and will not be repeated here.

[0150] Computer storage media includes permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory media such as modulated data signals and carrier waves.

[0151] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the above description has generally described the components and steps of each example according to their functions. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0152] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present application. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application is not limited to the embodiments shown herein, but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for detecting abnormal line loss in a transformer area, characterized in that: include: Obtain the daily line loss rate of the target area during the inspection period; Based on the daily line loss rate of the target substation during the inspection period, the operating indicators corresponding to the target substation are calculated; wherein the operating indicators include trend indicators, difference indicators, volatility indicators and ratio indicators; the trend indicators include upward trend indicators, downward trend indicators and monthly change rates; the difference indicators include range, interquartile range, mean difference between the front part of the data and the rear part of the data, difference between the mean of the segmented data and the total mean, modulus of the Fourier coefficient difference between the front part of the data and the rear part of the data, and fitting slope; the volatility indicators include the standard deviation of the overall data, the standard deviation of the front part of the data, the standard deviation of the rear part of the data and the coefficient of variation; the ratio indicators include the ratio of the mean of the front part of the data to the total mean, the ratio of the mean of the rear part of the data to the total mean and the correlation coefficient; Inputting the operating indicators corresponding to the target substation into a pre-trained classification model to obtain the line loss status corresponding to the target substation; wherein the classification model is pre-trained using training samples of multiple categories and the line loss status corresponding to each category; the training samples of one category include the operating indicators corresponding to each sample substation belonging to the category; each category is obtained by clustering each sample substation based on the operating indicators corresponding to each sample substation; and the line loss status corresponding to each category is obtained by analyzing the operating indicators corresponding to each sample substation belonging to the category; The line loss status corresponding to the target substation and the line loss factor corresponding to the line loss status are output; wherein the line loss factor corresponding to the line loss status is obtained in advance based on the analysis of the operating indicators corresponding to each of the sample substations belonging to the category corresponding to the line loss status.

2. The method according to claim 1, characterized in that After obtaining the daily line loss rate of the target area during the inspection period, the method further includes: Determining whether any of the line loss rates is a negative value or a null value within the inspection period; wherein, if it is determined whether the line loss rate is not a negative value or a null value within the inspection period, calculating the operating indicator corresponding to the target substation based on the daily line loss rate of the target substation within the inspection period; If it is determined that any of the line loss rates within the inspection period is negative or null, then a feedback is given that the target substation is abnormal.

3. The method according to claim 1, characterized in that The calculating of the operation index corresponding to the target substation based on the daily line loss rate of the target substation during the inspection period includes: Calculate the trend index corresponding to the target area based on the daily line loss rate of the target area during the inspection period; Based on the daily line loss rate of the target substation during the inspection period, the difference index corresponding to the target substation is calculated; wherein the first part of the data and the second part of the data are the line loss rates obtained by evenly dividing the line loss rates in chronological order. Based on the daily line loss rate of the target substation during the inspection period, a volatility index corresponding to the target substation is calculated; wherein the coefficient of variation is equal to the ratio of the standard deviation of the overall data to the mean of all the line loss rates; Based on the daily line loss rate of the target substation during the inspection period, the ratio index corresponding to the target substation is calculated; wherein the correlation coefficient refers to the correlation coefficient between the mean of each line loss rate of the target substation and the average line loss rate of similar substations.

4. The method according to claim 1, wherein The calculating of the trend indicator corresponding to the target area based on the daily line loss rate of the target area during the inspection period includes: Sort the daily line loss rates of the target substation area during the inspection period in chronological order to obtain a line loss rate sequence of the target substation area; For each target line loss rate in the line loss rate sequence, the mean of the moving data corresponding to the target line loss rate is calculated to obtain a moving average value corresponding to the target line loss rate; wherein the target line loss rate is the line loss rate other than the first and last line loss rates in the line loss rate sequence; and the moving data corresponding to one target line loss rate is a maximum number of no more than n line loss rates selected forward from the target line loss rate as a starting point; Calculating the mean of the moving average values ​​corresponding to the target line loss rates to obtain the moving average value corresponding to the line loss rate sequence; The rising trend indicator corresponding to the target station area is obtained by dividing the sum of the squares of the line loss rates in the line loss rate sequence that are smaller than the moving average value corresponding to the line loss rate sequence by the number of data segments located below the moving average value corresponding to the line loss rate sequence; The downward trend indicator corresponding to the target station area is obtained by dividing the sum of the squares of the line loss rates in the line loss rate sequence that are greater than the moving average value corresponding to the line loss rate sequence by the number of data segments located on the moving average value corresponding to the line loss rate sequence; The change rate of the mean value of each line loss rate in every two adjacent months is calculated respectively to obtain the monthly change rate corresponding to the target substation.

5. The method according to claim 1, wherein The calculating of the difference index corresponding to the target area based on the daily line loss rate of the target area during the inspection period includes: Calculating the difference between the maximum and minimum values ​​of each of the line loss rates to obtain the range corresponding to the target area; Calculating the difference between the upper quartile and the lower quartile of the line loss rate sequence composed of the line loss rates sorted in time, to obtain the interquartile range corresponding to the target station area; Calculating the mean difference between the first part of the data and the second part of the data; Divide the line loss rate sequence into four segments on average, and subtract the mean of all line loss rates from the sum of the mean of the line loss rates in each segment to obtain the difference between the mean of the segment data and the total mean; Calculating the modulus of the difference between the Fourier coefficients of the front portion of data and the rear portion of data; Perform linear fitting on each of the line loss rates to obtain a fitting slope corresponding to the target station area.

6. The method according to claim 1, characterized in that The training method of the classification model includes: Obtaining daily line loss rates of the plurality of sample areas during an inspection period; Calculating the operating index corresponding to each sample area based on the line loss rate of each sample area on each day during the inspection period; Clustering each of the sample areas based on the operation indicators corresponding to each of the sample areas to obtain a plurality of categories; For each of the categories, analyzing the operating indicators corresponding to the sample substations belonging to the category, and obtaining the line loss status corresponding to the category and the line loss factor corresponding to the line loss status; respectively taking the line loss status corresponding to each category as a label corresponding to each sample station area belonging to the category; The classification model is iteratively trained using the operating indicators and labels corresponding to each of the sample stations.

7. The method according to claim 6, characterized in that The clustering of the sample areas based on the operation indicators corresponding to the sample areas to obtain a plurality of categories includes: The density peak algorithm CFSFDP is used to calculate the operation index corresponding to each of the sample stations to obtain multiple initial cluster centers; The K-means clustering algorithm is used to cluster the sample areas based on the multiple initial cluster centers and the operating indicators corresponding to the sample areas to obtain the multiple categories.

8. The method according to claim 6, characterized in that The classification model includes a plurality of classification sub-models, and the iterative training of the classification model using the operation indicators and labels corresponding to the sample stations includes: Combining every two of the categories to obtain multiple combinations; The operation indicators and labels corresponding to the sample stations under the two categories in each combination are used to train a classification sub-model to obtain the trained classification sub-models.

9. The method according to claim 8, characterized in that Inputting the operating indicators corresponding to the target substation into a pre-trained classification model to obtain the line loss status corresponding to the target substation includes: Outputting the operating indicators corresponding to the target substation to each of the pre-trained classification sub-models to obtain the line loss status output by each of the classification sub-models; The line loss state with the largest proportion among the line loss states output by each of the classification sub-models is determined as the line loss state corresponding to the target substation.

10. A device for detecting abnormal line loss in a transformer area, characterized in that: include: The first acquisition unit is used to obtain the line loss rate of the target area every day during the inspection period; A first calculation unit is configured to calculate an operation index corresponding to the target substation based on the daily line loss rate of the target substation during an inspection period; wherein the operation index includes a trend index, a difference index, a volatility index, and a ratio index; the trend index includes an upward trend index, a downward trend index, and a monthly rate of change; the difference index includes an extreme value, an interquartile range, a mean difference between the front data and the rear data, a difference between the mean of the segmented data and the overall mean, a modulus of a Fourier coefficient difference between the front data and the rear data, and a fitting slope; the volatility index includes a standard deviation of the overall data, a standard deviation of the front data, a standard deviation of the rear data, and a coefficient of variation; the ratio index includes a ratio of the mean of the front data to the overall mean, a ratio of the mean of the rear data to the overall mean, and a correlation coefficient; An input unit is configured to input the operating indicators corresponding to the target substation into a pre-trained classification model to obtain the line loss status corresponding to the target substation; wherein the classification model is pre-trained using training samples of multiple categories and the line loss status corresponding to each category; the training samples of one category include the operating indicators corresponding to each sample substation belonging to the category; each category is obtained by clustering each sample substation based on the operating indicators corresponding to each sample substation; and the line loss status corresponding to each category is obtained by analyzing the operating indicators corresponding to each sample substation belonging to the category. An output unit is used to output the line loss status corresponding to the target substation and the line loss factor corresponding to the line loss status; wherein the line loss factor corresponding to the line loss status is obtained in advance based on the analysis of the operating indicators corresponding to each of the sample substations belonging to the category corresponding to the line loss status.

Images