Transformer external characteristic scatter diagram-based special transformer user metering abnormality monitoring method

Abstract

The application provides a special transformer user metering abnormality monitoring method based on a transformer external characteristic scatter diagram, adopts a rule of transformer external characteristics to infer a load condition of a user, and uses normal historical data to infer an abnormal threshold value, so as to judge whether metering abnormality exists. Compared with a traditional method, the method has the advantages of less required data, no influence of data synchronization, and good monitoring effect for a user with good power utilization rule.

Description

Technical Field

[0001] This invention belongs to the field of user metering anomaly monitoring technology, and particularly relates to a method for monitoring anomalies in dedicated transformer user metering based on a scatter plot of transformer external characteristics. Background Technology

[0002] With increasingly sophisticated user electricity consumption information collection systems, remote collection of user electricity data can be used to analyze the operation of user electricity metering devices. However, manually analyzing and screening for anomalies in such massive amounts of data is time-consuming. Therefore, efficient utilization of this data is crucial. Against this backdrop, research into methods for monitoring metering anomalies in dedicated transformer users based on data mining is essential for addressing metering faults and electricity theft by dedicated transformer users, as well as for recovering excess electricity consumption. This is of great significance for ensuring the stable and normal operation of power companies.

[0003] There is already a lot of research on methods for monitoring anomalies in user metering, such as the power balance method, the power balance method, and the transformer external characteristic method. However, if the three phases of the system are unbalanced or the power balance data has problems with synchronization and integrity, the effect of such methods in monitoring anomalies is not ideal. Summary of the Invention

[0004] To address the shortcomings and deficiencies of existing solutions, the purpose of this invention is to provide a novel method for monitoring metering anomalies in dedicated transformer users, specifically a method based on scatter plots of transformer external characteristics. Because the load of dedicated transformer users is relatively fixed, the external characteristics of the transformer exhibit certain patterns under different operating conditions. These patterns can be used to determine whether metering anomalies exist. This invention studies a method for monitoring user metering anomalies, thereby addressing metering faults and electricity theft, and recovering excess electricity, which is of great significance for ensuring the normal operation of power companies.

[0005] The specific technical solution adopted is as follows:

[0006] A method for monitoring metering anomalies in dedicated transformer users based on scatter plots of transformer external characteristics, characterized by the following steps:

[0007] Step S1: Select n days of normal data under the same special transformer as the sample population, that is:

[0008] (1)

[0009] Where D is the sample population, The data are the three-phase voltage and three-phase current for day i.

[0010] Step S2: Impute missing values ​​in the sample data using Lagrange interpolation;

[0011] Step S3: Describe the relationship between the transformer output voltage and output current using a scatter plot to form a scatter plot of the transformer external characteristics. Then fit the scatter plot of the transformer external characteristics into a straight line with a certain slope. Draw the daily scatter plots of the transformer external characteristics for phases A, B, and C, and fit the daily scatter plot curves of the external characteristics for phases A, B, and C.

[0012] Step S4: Let the average daily current on day i be... The following formula is used for calculation:

[0013] (5)

[0014] in , , These represent the sampled current values ​​of phases A, B, and C at the nth metering time on the i-th day, respectively.

[0015] Step S5: Use Describe the electricity consumption characteristics of dedicated transformer users on day i. This represents the slope of the fitted straight line in the scatter plot of the transformer's external characteristics for a dedicated transformer user on day i. Let represent the average daily current of the dedicated transformer user on day i; establish matrix Q as the dataset of features of the dedicated transformer user, as follows:

[0016] (6)

[0017] (7)

[0018] Wherein, the Q matrix is ​​a 4×n matrix, where n represents the number of days;

[0019] Step S6: After the matrix Q data has been dedimensionalized, calculate the average distance between the data points in matrix Q. Specifically:

[0020] (8)

[0021] (9)

[0022] Step S7: [The text appears to be incomplete and contains several grammatical errors. A more accurate translation would require the full context.] Defined as Centered on, Calculate the density of all data in matrix Q for a sample region of radius Q. Specifically:

[0023] = (10)

[0024] in,

[0025] Step S8: Form a cluster from the data that meet the conditions of step S7, and calculate the average distance of the sample objects in each cluster. Specifically:

[0026] = (11);

[0027] Step S9: Calculate cluster distance Specifically:

[0028] (12);

[0029] Step S10: Calculate parameters Specifically:

[0030] (13);

[0031] Step S11: Let the set of cluster centroids be... Select the largest value calculated in step S7. Corresponding It is added to set C as the first cluster center point, i.e. ,in = All objects that satisfy the sample object and The distance between them is less than The sample objects are classified as cluster 1 and removed from dataset Q;

[0032] Step S12: Using equations (8) to (13), calculate the remaining elements of the dataset. ,Will The largest element as Add it to set C, that is And remove the elements of cluster 2 from the dataset;

[0033] Step S13: Specify that the number of sample objects in each cluster is no less than a given percentage of the dataset Q, and calculate the optimal number of categories k;

[0034] Step S14: Obtain the set of cluster centroids The cluster center is selected from the sample objects in each cluster whose loss function to other sample objects within the cluster is minimized. The loss function is as follows:

[0035] (14)

[0036] in express Size, express In addition to All external data objects to Euclidean distance; to obtain the set of cluster centers. ;

[0037] Step S15: with Using k as the initial cluster centers and k as the optimal number of clusters, the k-medoids algorithm is employed for clustering to obtain matrices of electrical characteristics under different operating conditions. The new set of cluster centers is ;

[0038] Step S16: For each cluster of data after clustering, use... Set a threshold around the center. This ensures that 95% of the data in this category falls within the range;

[0039] Step S17: Based on the obtained threshold set ,Record and As a normal external characteristic indicator of the transformer for users;

[0040] Step S18: For each type of data object, if its distance from the center point of the type is greater than the threshold of the type, it is judged as an outlier and it is considered that there may be a measurement anomaly.

[0041] Furthermore, in step S3, Tilson regression is used to perform linear fitting on the curve, specifically as follows:

[0042] (3)

[0043] (4)

[0044] Where α and β are the model fitting parameters, and ϵ is the model random error. This is the current value. Voltage value

[0045] Let the slope of the fitted line obtained by fitting the data of phase A on day i be . The slope of the fitted line obtained by fitting the B-phase data is... The slope of the fitted line obtained by fitting the C-phase data is... ,remember .

[0046] Furthermore, the specific monitoring method in step S18 is as follows: acquire the three-phase voltage and three-phase current data of the dedicated transformer user, and calculate... , and Calculate the Euclidean distance for each element in the set, and determine... It belongs to the category with the smallest Euclidean distance, when The distance from the center point of this class is greater than according to If the threshold value determined by the set is used, it is judged that there may be a measurement anomaly; otherwise, the measurement is normal.

[0047] Compared with existing technologies, this invention and its preferred embodiment use the patterns exhibited by the external characteristics of the transformer to infer the user's load status and use normal historical data to infer the abnormal threshold, thereby determining whether there is a metering anomaly. Compared with traditional methods, this method requires less data, is not affected by data synchronization issues, and has a good monitoring effect for users with good electricity consumption patterns. Attached Figure Description

[0048] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments:

[0049] Figure 1 This is a flowchart of the method for monitoring metering anomalies of dedicated transformer users based on the scatter plot of transformer external characteristics, according to an embodiment of the present invention. Detailed Implementation

[0050] To make the features and advantages of this patent more apparent and understandable, specific embodiments are provided below for detailed explanation:

[0051] The purpose of this invention is to provide a novel method for monitoring metering anomalies in dedicated transformer users, specifically a method based on scatter plots of transformer external characteristics. Because the load of dedicated transformer users is relatively fixed, the external characteristics of the transformer exhibit certain patterns under different operating conditions. Based on these patterns, it is possible to determine whether metering anomalies exist.

[0052] The key advantage of the method in this embodiment is that it is not affected by data synchronization issues.

[0053] The data source for the method in this embodiment is the three-phase voltage and three-phase current values ​​on the low-voltage side of the special transformer measured by the high-voltage supply and low-voltage metering device. Under normal circumstances, there are a total of 96 metering values ​​per day.

[0054] like Figure 1 As shown, the specific implementation process of the method in this embodiment is as follows:

[0055] (1) Select n days of normal data under the same special transformer as the sample population, that is:

[0056] (1)

[0057] Where D is the sample population, The data are the three-phase voltage and three-phase current for day i.

[0058] (2) Use Lagrange interpolation to fill in the missing values ​​in the sample data.

[0059] For example, suppose the phase A current data for a certain day is... Let the missing value be Take the data before and after m missing values, and denote it as... The corresponding time is denoted as The formula for filling in the gaps is:

[0060] (2)

[0061] Other data are also filled in according to formula (2) to obtain complete three-phase voltage and three-phase current data.

[0062] (3) The relationship between the transformer output voltage and output current can be described by a scatter plot, commonly known as the transformer external characteristic scatter plot. The shape of the transformer external characteristic scatter plot is related to the parameters of the transformer itself and also to the nature of the load. The daily load of dedicated transformer users is relatively regular, and their transformer external characteristic scatter plot can be fitted into a straight line with a certain slope. Draw the daily transformer external characteristic scatter plots of phases A, B, and C, and fit the daily external characteristic scatter plot curves of phases A, B, and C. In this embodiment, the Theil-Sen Regression method is used to linearly fit the curve, specifically:

[0063] (3)

[0064] (4)

[0065] Where α and β are the model fitting parameters, and ϵ is the model random error. This is the current value. Voltage value

[0066] Let the slope of the fitted line obtained by fitting the data of phase A on day i be . The slope of the fitted line obtained by fitting the B-phase data is... The slope of the fitted line obtained by fitting the C-phase data is... ,remember .

[0067] (4) Let the average daily current on day i be... It can be calculated using the following formula:

[0068] (5)

[0069] in , , These represent the sampled current values ​​of phases A, B, and C at the nth metering time on the i-th day, respectively.

[0070] (5) Use Describe the electricity consumption characteristics of dedicated transformer users on day i. This represents the slope of the fitted straight line in the scatter plot of the transformer's external characteristics for a dedicated transformer user on day i. Let represent the average daily current of the dedicated transformer user on day i. Establish matrix Q, considering Q as a dataset that describes the characteristics of the dedicated transformer user, as follows:

[0071] (6)

[0072] (7)

[0073] in The Q matrix is ​​a 4×n matrix, where n represents the number of days;

[0074] (6) Calculate the average distance between data points in matrix Q. (Assuming matrix Q has already been dedimensionalized before performing any of the following calculations), specifically:

[0075] (8)

[0076] (9)

[0077] (7) Calculate the density of all data in matrix Q. (Will Defined as Centered on, The sample region (with radius ) is as follows:

[0078] = (10)

[0079] in,

[0080] (8) Data that meet the conditions of step (7) above are grouped into a cluster, and the average distance of sample objects in each cluster is calculated. Specifically:

[0081] = (11)

[0082] (9) Calculate the cluster distance Specifically:

[0083] (12)

[0084] (10) Calculation parameters Specifically:

[0085] (13)

[0086] (11) Let the set of cluster centers be... Select the largest value calculated in step (7). Corresponding It is added to set C as the first cluster center point, i.e. ,in = All objects that satisfy the sample and The distance between them is less than The sample objects can be considered as cluster 1, and they are removed from the dataset Q.

[0087] (12) Using equations (8) to (13), calculate the remaining elements of the dataset. ,Will The largest element as Add it to set C, that is Similarly, remove the elements from cluster 2 from the dataset.

[0088] (13) The number of sample objects in each cluster shall not be less than 5% of the dataset Q (other thresholds may be set as needed). According to the above rules, the optimal number of categories k shall be calculated.

[0089] (14) After the above steps, the set of cluster center points can be obtained. The cluster center is selected from the sample objects in each cluster whose loss function to other sample objects within the same cluster is minimized. The loss function is as follows:

[0090] (14)

[0091] in express Size, express In addition to All external data objects to Euclidean distance.

[0092] (15) After the above steps, the set of cluster centers is obtained.

[0093] (16) with Using k as the initial cluster centers and k as the optimal number of clusters, the k-medoids algorithm is employed for clustering, yielding a matrix that roughly describes the electrical characteristics under different operating conditions. The new set of cluster centers is .

[0094] (17) For each class of data after clustering, use Set an appropriate threshold around the center. This ensures that 95% of the data in this category falls within the range.

[0095] (18) Based on step (16), the threshold set is obtained as follows: .Record and This serves as a normal external characteristic indicator for the user's transformer.

[0096] (19) For each type of data object, if its distance from the center point of that type is greater than the threshold of that type, it is judged as an outlier, and it can be considered that there may be a metering anomaly. Therefore, the specific monitoring method is: to obtain the three-phase voltage and three-phase current data of the dedicated transformer user, and calculate , With the above Calculate the Euclidean distance for each element in the set, and determine... It belongs to the category with the smallest Euclidean distance, when The distance to the center point of this class is greater than the threshold of this class (according to...) If the result is a collection of data, it is considered that there may be a measurement anomaly; otherwise, the measurement is considered normal.

[0097] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0098] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0099] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0100] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0101] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

[0102] This patent is not limited to the above-described preferred embodiment. Anyone can derive other forms of methods for monitoring metering anomalies of dedicated transformer users based on scatter plots of transformer external characteristics under the guidance of this patent. All equivalent changes and modifications made within the scope of this patent application shall fall within the scope of this patent.

Claims

1. A method for monitoring metering anomalies in dedicated transformer users based on scatter plots of transformer external characteristics, characterized in that, Includes the following steps: Step S1: Select n days of historical normal data under the same dedicated transformer as the sample population, that is: (1) Where D is the sample population, The data are the three-phase voltage and three-phase current for day i. Step S2: Impute missing values ​​in the sample data using Lagrange interpolation; Step S3: Describe the relationship between the transformer output voltage and output current using a scatter plot to form a scatter plot of the transformer external characteristics. Then fit the scatter plot of the transformer external characteristics into a straight line with a certain slope. Draw the daily scatter plots of the transformer external characteristics for phases A, B, and C, and fit the daily scatter plot curves of the external characteristics for phases A, B, and C. Step S4: Let the average daily current on day i be... The following formula is used for calculation: (5) in , , These represent the sampled current values ​​of phases A, B, and C at the t-th metering time on the i-th day, respectively. Step S5: Use Describe the electricity consumption characteristics of dedicated transformer users on day i. This represents the slope of the fitted straight line in the scatter plot of the transformer's external characteristics for a dedicated transformer user on day i. Let represent the average daily current of the dedicated transformer user on day i; establish matrix Q as the dataset of features of the dedicated transformer user, as follows: (6) (7) Wherein, the Q matrix is ​​a 4×n matrix, where n represents the number of days; Step S6: After the matrix Q data has been dedimensionalized, calculate the average distance between the data points in matrix Q. Specifically: (8) (9) Step S7: [The text appears to be incomplete and contains several grammatical errors. A more accurate translation would require the full context.] Defined as Centered on, Calculate the density of all data in matrix Q for a sample region of radius Q. Specifically: = (10) in, Step S8: Form a cluster from the data that meet the conditions of step S7, and calculate the average distance of the sample objects in each cluster. Specifically: = (11)； Step S9: Calculate cluster distance Specifically: (12)； Step S10: Calculate parameters Specifically: (13)； Step S11: Let the set of cluster centroids be... Select the largest value calculated in step S7. Corresponding It is added to the set C as the first cluster center point, i.e. ,in = All objects that satisfy the sample object and The distance between them is less than The sample objects are classified as cluster 1 and removed from the dataset Q; Step S12: Using equations (8) to (13), calculate the remaining elements of the dataset. ,Will The largest element as Add to set C, i.e. And remove the elements of cluster 2 from the dataset; Step S13: Specify that the number of sample objects in each cluster is no less than a given percentage of the dataset Q, and calculate the optimal number of categories k; Step S14: Obtain the set of cluster centroids The cluster center is selected from the sample objects in each cluster whose loss function to other sample objects within the cluster is minimized. The loss function is as follows: (14) in express Size, express In addition to All external data objects to Euclidean distance; to obtain the set of cluster centers. ; Step S15: with Using k as the initial cluster centers and k as the optimal number of clusters, the k-medoids algorithm is employed for clustering to obtain matrices of electrical characteristics under different operating conditions. The new set of cluster centers is ; Step S16: For each cluster of data after clustering, use... Set a threshold around the center. This ensures that 95% of the data in this category falls within the range; Step S17: Based on the obtained threshold set ,Record and As a normal external characteristic indicator of the transformer for users; Step S18: Obtain the three-phase voltage and three-phase current data of the special transformer user to be tested, and calculate the feature vector according to the feature vector calculation method in steps S3-S5. , and Calculate the Euclidean distance for each element in the set, and determine... It belongs to the category with the smallest Euclidean distance, when The distance from the center point of this class is greater than according to If the threshold value determined by the set is used, it is judged as a measurement anomaly; otherwise, it is considered a normal measurement.

2. The method for monitoring metering anomalies of dedicated transformer users based on scatter plots of transformer external characteristics according to claim 1, characterized in that: In step S3, the curve is linearly fitted using Tilson regression, specifically as follows: (3) (4) Where α and β are the model fitting parameters, and ϵ is the model random error. This is the current value. Voltage value Let the slope of the fitted line obtained by fitting the data of phase A on day i be . The slope of the fitted line obtained by fitting the B-phase data is... The slope of the fitted line obtained by fitting the C-phase data is... ,remember .

Images