Online verification method and system for automatic verification device of electric energy meter

Abstract

The invention discloses an on-line verification method and system for an automatic verification device of an electric energy meter, and relates to the field of electric power, and the method comprises the steps: carrying out the on-line verification of a preset meter position of the automatic verification device of the electric energy meter to be verified through an installation type intelligent standard electric energy meter, and obtaining first data through data collection; preprocessing the first data to obtain second data; performing feature extraction and principal component extraction based on the second data to obtain a third feature; classifying the third features, marking the sample as a normal or fault type, and obtaining fourth features; by using the fourth feature training model, the state of the device is judged through online verification of the automatic verification device of the electric energy meter; generating a report according to an online verification result, and carrying out manual troubleshooting processing on a fault; according to the method, the efficiency is greatly improved, the real-time performance is good, online verification of the verification device is achieved, possible losses are avoided, and meanwhile the situation that normal verification of the electric energy meter needs to be stopped in a traditional verification method is avoided.

Description

Technical Field

[0001] The present invention relates to the field of electric power, and in particular to an online calibration method and system for an automatic calibration device of an electric energy meter. Background Art

[0002] The automatic calibration device of the electric energy meter uses advanced sensing technology, data processing technology and automatic control technology to realize online calibration of the automatic calibration device of the electric energy meter. The automatic calibration device of the electric energy meter may be affected by environmental factors or the device itself during operation, and the electrical performance may be degraded, resulting in an increase in the calibration error or even exceeding the error limit requirements of the current accuracy level. The reliability of the experimental results and the stability of the device are reduced. The automatic calibration device of the electric energy meter needs to be calibrated regularly according to the national calibration procedures.

[0003] Traditional periodic calibration requires the suspension of the calibration task of the calibration device, and is carried out manually offline on a regular basis, which may result in low calibration efficiency and the possibility of human operation errors. The fault and abnormality detection of the calibration device during operation usually relies on the experience and subjective judgment of the calibration personnel, lacking objectivity and accuracy, which may lead to misjudgment or missed judgment, thus affecting the accurate identification of abnormalities and effective monitoring of the device. Existing research uses manual offline regular calibration, which has low calibration efficiency, low intelligence and automation, and fails to carry out quality monitoring of the calibration device in combination with the operation status of the automatic calibration device of the electric energy meter. Summary of the invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problem solved by the present invention is: how to realize the online verification of the automatic verification device of the electric energy meter.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, an embodiment of the present invention provides an online calibration method for an automatic calibration device of an electric energy meter, comprising:

[0008] Relying on the installed intelligent standard electric energy meter, the predetermined meter position of the automatic calibration device of the electric energy meter to be calibrated is calibrated online, and the first data is obtained by collecting data;

[0009] Preprocessing the first data to obtain second data;

[0010] Perform feature extraction and principal component extraction based on the second data to obtain a third feature;

[0011] Classify the third feature and mark the sample as normal or fault type to obtain the fourth feature;

[0012] The fourth feature training model is used to determine the device status by online verification of the automatic verification device of the electric energy meter;

[0013] Generate a report based on the online verification results and manually troubleshoot the faults.

[0014] As a preferred solution of the online verification method of the electric energy meter automatic verification device, wherein:

[0015] The method of performing online calibration on a predetermined meter position of the automatic calibration device of the calibrated watt-hour meter by relying on the installed intelligent standard watt-hour meter, and obtaining the first data by collecting data includes:

[0016] Select multiple installed intelligent standard electric energy meters, fix them at the predetermined initial meter positions, and set the number of calibration batches;

[0017] The installed intelligent standard watt-hour meter is tested together with the watt-hour meter being tested in the watt-hour meter automatic calibration device at the same time, and the calibration is carried out and the data of the installed intelligent standard watt-hour meter automatic calibration device is collected. After each batch is completed, the installed intelligent standard watt-hour meter continues to be calibrated with the next batch.

[0018] As a preferred solution of the online verification method of the electric energy meter automatic verification device, wherein:

[0019] The method of performing online calibration on a predetermined meter position of the automatic calibration device of the to-be-calibrated electric energy meter by relying on the installed intelligent standard electric energy meter, and obtaining the first data by collecting data also includes:

[0020] After completing all data collection at the initial position, move the installed standard electric energy meter to the right and fix it at the new position, and repeat the above data collection and verification cycle at the new position until the installed intelligent standard electric energy meter covers all the predetermined meter positions;

[0021] The data collected from all epitopes are summarized to form the first data.

[0022] As a preferred solution of the online verification method of the electric energy meter automatic verification device, wherein:

[0023] The preprocessing of the first data to obtain the second data includes:

[0024] The original data is preprocessed using data cleaning, normalization and transformation methods; the second data is the first data after preprocessing; combined with the second data, the basic error, symmetry, distortion, device output power stability, device measurement repeatability, multi-channel output consistency, variation within the calibration period, and daily timing error of the standard clock tester are calculated for each table.

[0025] As a preferred solution of the online verification method of the electric energy meter automatic verification device, wherein:

[0026] The third feature is classified, and the sample is marked as normal or fault type, and the fourth feature is obtained including:

[0027] The machine learning method is used to classify the samples and label the four types of samples: normal, mechanical failure, electrical failure and reliability failure. The fourth feature is the four types of samples after labeling.

[0028] As a preferred solution of the online verification method of the electric energy meter automatic verification device, wherein:

[0029] The method of using the fourth feature training model to online calibrate the automatic calibration device of the electric energy meter to determine whether the device is qualified and to determine the device status through model analysis includes:

[0030] Calculate the detection performance index of each epitope; use the fourth feature to train the automatic detection model, use the trained detection model to complete the online detection of the automatic detection device, and judge the status of the detection device based on the output result of the model.

[0031] As a preferred solution of the online verification method of the electric energy meter automatic verification device, wherein:

[0032] Generating a verification report based on the online verification results and manually troubleshooting the faults includes:

[0033] A quality report is generated based on the online calibration results. The calibration status is divided into normal, mechanical failure, electrical failure and reliability failure. If the calibration device fails, manual troubleshooting is carried out.

[0034] In a second aspect, an embodiment of the present invention provides an online calibration system for an automatic calibration device of an electric energy meter, comprising:

[0035] The first data acquisition module is used to perform online calibration on a predetermined meter position of the automatic calibration device of the calibrated watt-hour meter by relying on the installed intelligent standard watt-hour meter, and obtain the first data by collecting data;

[0036] A second data acquisition module, used for preprocessing the first data to obtain second data;

[0037] A third feature acquisition module, used for performing feature extraction and principal component extraction based on the second data to obtain a third feature;

[0038] A fourth feature acquisition module is used to classify the third feature, mark the sample as normal or fault type, and obtain a fourth feature;

[0039] A verification module, for determining the status of the device by online verification of the automatic verification device of the electric energy meter using the fourth characteristic training model;

[0040] The report processing module is used to generate reports based on the online verification results and manually troubleshoot faults.

[0041] In a third aspect, an embodiment of the present invention provides a computing device, including:

[0042] Memory and processor;

[0043] The memory is used to store computer executable instructions, and the processor is used to execute the computer executable instructions. When the one or more programs are executed by the one or more processors, the one or more processors implement the online calibration method of the electric energy meter automatic calibration device as described in any embodiment of the present invention.

[0044] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the online calibration method of the automatic calibration device of an electric energy meter.

[0045] The beneficial effects of the present invention are as follows: the present invention can automatically perform online calibration of the automatic calibration device, improve the calibration results and the accuracy of the measurement transmission, the calibration efficiency of the calibration cycle, and at the same time greatly improve the efficiency and save a lot of human resources; the use of machine learning models can significantly reduce the errors caused by human factors; it has good real-time performance and can realize online calibration of the calibration device; combined with online calibration, when a fault occurs in the calibration device, it can be immediately discovered and processed, avoiding possible losses, and at the same time avoiding the need to stop the normal calibration of the electric energy meter in the traditional calibration method. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative labor.

[0047] Figure 1 It is an overall flow chart of the online verification method of the electric energy meter automatic verification device of the present invention;

[0048] Figure 2 It is a data acquisition principle diagram of the online verification method of the electric energy meter automatic verification device of the present invention;

[0049] Figure 3 It is a schematic diagram of the principal component variance contribution rate and the cumulative variance contribution rate in the simulation example of the online verification method of the electric energy meter automatic verification device according to the second embodiment of the present invention;

[0050] Figure 4It is a box plot of principal component data in a simulation example of an online verification method for an automatic verification device of an electric energy meter according to a second embodiment of the present invention;

[0051] Figure 5 It is a box-line diagram of original characteristic data in a simulation example of an online verification method of an automatic verification device for an electric energy meter according to a second embodiment of the present invention;

[0052] Figure 6 It is a confusion matrix of verification results based on data enhancement in a simulation example of an online verification method for an automatic verification device of an electric energy meter described in the second embodiment of the present invention. DETAILED DESCRIPTION

[0053] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0054] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0055] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0056] Example 1

[0057] Reference Figure 1-Figure 2 , which is the first embodiment of the present invention, provides an online calibration method for an automatic calibration device of an electric energy meter, comprising:

[0058] S1: relying on the installed intelligent standard electric energy meter, online calibration is performed on the predetermined meter position of the automatic calibration device of the electric energy meter to be calibrated, and the first data is obtained by collecting data;

[0059] S2: preprocessing the first data to obtain second data;

[0060] S3: Perform feature extraction and principal component extraction based on the second data to obtain a third feature;

[0061] S4: classify the third feature, mark the sample as normal or fault type, and obtain the fourth feature;

[0062] S5: using the fourth feature training model to determine the device status by online verification of the electric energy meter automatic verification device;

[0063] S6: Generate a report based on the online verification results and manually troubleshoot the fault.

[0064] In the embodiment of the present application, in the above step S1, relying on the installed intelligent standard electric energy meter, the predetermined meter position of the automatic calibration device of the electric energy meter to be calibrated is calibrated online, and the first data obtained by collecting data includes:

[0065] like Figure 2 As shown, i installed intelligent standard electric energy meters with higher accuracy than the calibration device are used to perform calibration simultaneously with the electric energy meter under test at meter positions 1 to i. After a batch of electric energy meters are calibrated, i installed intelligent standard electric energy meters continue to be calibrated with the next batch of electric energy meters after caching, and a total of n batches of calibration are performed; then enter the second round of detection, i installed intelligent standard electric energy meters move to the right and are fixed at the meter positions i+1 to 2i, and repeat the calibration steps until the installed intelligent standard electric energy meter completes the data collection for each meter position. Each meter position is collected n times, and the first data includes the current, voltage, and power in each meter position experiment.

[0066] In the embodiment of the present application, preprocessing the first data in the above step S2 to obtain the second data includes:

[0067] The original data is preprocessed using data cleaning, normalization and transformation methods to facilitate subsequent feature extraction; the second data is the preprocessed data of the first data.

[0068] In the embodiment of the present application, the preprocessing includes:

[0069] Data cleaning: dealing with missing, duplicate, erroneous or irrelevant data;

[0070] Data normalization: Z-score normalization: This method normalizes the data based on the mean and standard deviation of the original data. The calculation formula is:

[0071] X'=(X-μ) / σ

[0072] Among them, μ represents the mean of the original data, and σ represents the standard deviation of the original data;

[0073] Data transformation: Data transformation involves simplifying existing features, constructing new features, or changing the distribution of data.

[0074] In another possible implementation, the preprocessing may further include:

[0075] Data filtering: Filter the data using low-pass, high-pass, or band-pass filters to remove noise and retain useful signal features.

[0076] Data interpolation: Interpolate missing data to fill in the gaps in the data and ensure data integrity.

[0077] Data smoothing: Smoothing data through methods such as moving average or weighted moving average to reduce data fluctuations.

[0078] In the embodiment of the present application, in the above step S3, feature extraction and principal component extraction are performed based on the second data to obtain the third feature including:

[0079] Based on the second data, 17 features are calculated, which are basic error, phase voltage symmetry, line voltage symmetry, current symmetry, phase-to-phase symmetry, line-to-line phase symmetry, A-phase voltage distortion, A-phase current distortion, B-phase voltage distortion, B-phase current distortion, C-phase voltage distortion, C-phase current distortion, device output power stability, device measurement repeatability, determination of multi-channel output consistency, variation within the calibration cycle, and daily timing error of the standard clock tester. For each feature data, five feature values ​​are obtained, namely, maximum value (Max), minimum value (Min), average (E), standard deviation (S), and median (M);

[0080] Specifically, the maximum value (Max), minimum value (Min), average value (E), standard deviation (S), and median value (M) are defined and calculated as follows:

[0081] Maximum value (Max): the maximum value within the collection interval;

[0082] Minimum value (Min): the maximum value within the collection interval;

[0083] Average value(E):

[0084] Where n is the total number of data points, x i is the value of each data point, Σ represents the value of x i The sum of

[0085] Standard deviation (S): expressed as:

[0086]

[0087] Where E is the mean value, x i is the value of each data point, Σ represents the sum;

[0088] Median (M): The median is the value in the middle of a data set of n numbers sorted from smallest to largest. If the number of data n is an odd number, the median is If the data n is an even number, the median is usually the average of the two middle values ​​after sorting.

[0089] Principal Component Analysis (PCA) was used to reduce the dimension of the data and extract m principal components;

[0090] Specifically, when the cumulative variance contribution rate of the current m principal components is not less than a set threshold (usually 85%), the m principal components are used to replace the original high-dimensional feature data.

[0091] The third feature includes the five eigenvalues ​​obtained from the above 17 features respectively, and the result after dimensionality reduction.

[0092] In the embodiment of the present application, the third feature is classified in the above step S4, and the sample is marked as normal or fault type, and the fourth feature is obtained including:

[0093] The machine learning method is used to classify the samples and label the four types of samples: normal, mechanical failure, electrical failure and reliability failure. The fourth feature is the four types of samples after labeling.

[0094] Specifically, the unlabeled data is divided into four categories through clustering algorithm: normal class, mechanical failure, electrical failure and reliability failure. K-means clustering algorithm is a simple and widely used clustering analysis method, which aims to divide n data into k clusters (here k is 4) according to features, so that the member objects in a cluster are more similar than the member objects in other clusters.

[0095] The main process of the K-means clustering algorithm is as follows:

[0096] S401: Determine the number of clusters k. Initially, randomly select k data points as the initial cluster centers of the algorithm, that is, the initial cluster centers a = a1, a2, ... a k ;

[0097] S402: Calculate the sample point x in the data set i The shortest distance to its k cluster centers divides the sample data into the cluster to which the nearest center point belongs;

[0098] In this algorithm, Euclidean distance is used as the similarity between data points.

[0099]

[0100] Among them, D represents the shortest distance between the sample and the cluster center, xi represents the i-th sample, and n is the total number of samples.

[0101] S403: For each cluster, recalculate the cluster center according to the positions of all sample points in the cluster. The new cluster center is the mean position of all its member data points. For each dimension, the coordinates of the new cluster center are the average of the coordinates of all points in the cluster on that dimension. Update the clusters to which all points belong again based on similarity.

[0102] S404: Repeat the above steps S402 and S403, and iterate repeatedly until the cluster center position converges, the clustering criterion function converges, or the maximum number of iterations is reached.

[0103] It should be noted that the sample data is classified and labeled by the K-means algorithm, providing sample labels for training the online fault recognition model.

[0104] In another possible implementation, support vector machine (SVM) classification can also be used:

[0105] Support vector machine is a supervised learning model that can be used for classification and regression tasks. It distinguishes different categories by finding an optimal hyperplane in high-dimensional space.

[0106] Specifically, a support vector machine is selected as the classification algorithm, and an appropriate kernel function (such as linear, polynomial, radial basis function, etc.) is selected.

[0107] Use the labeled training data set to train the SVM model and find the optimal hyperplane.

[0108] Use the trained SVM model to classify new samples and label them as normal, mechanical failure, electrical failure, or reliability failure.

[0109] In another possible implementation, Random Forest classification can also be used:

[0110] Random forest is an ensemble learning method that combines multiple decision trees for classification.

[0111] Specifically, select Random Forest as the classification algorithm and set the number of decision trees and other parameters.

[0112] Use the labeled training data set to train the random forest model and generate multiple decision trees.

[0113] Use the trained random forest model to classify new samples and label them as normal, mechanical failure, electrical failure, or reliability failure.

[0114] In the embodiment of the present application, the fourth feature training model is used in the above step S5 to determine the device status by online calibrating the automatic calibration device of the electric energy meter, including:

[0115] Calculate the test performance index of each epitope; adopt the big data method, use the fourth feature to train the automatic test model, use the trained test model to complete the online test of the automatic test device, and judge the test status of the test device according to the output result of the model;

[0116] Specifically, according to the verification regulations of JJG597-2005 "AC Electric Energy Meter Verification Device", calculate the basic error, symmetry, distortion, device output power stability, device measurement repeatability, determine the consistency of multi-channel outputs, variation within the verification cycle, and daily timing error of the standard clock tester for each meter position;

[0117] The Multi-class Support Vector Machine (M-SVM) is used to complete the model training of the calibration device:

[0118] A one-vs-one (OvO) strategy is used to achieve multi-classification. The OvO strategy trains an SVM between each pair of categories. For k categories, k classifiers need to be trained. When predicting a new data point, the model provides it to all classifiers. Each classifier outputs a category based on its predicted probability. The final classification result is determined by the voting results of all classifiers. 60% of the samples are used for verification model training, and 40% of the samples are used for verification model verification;

[0119] The Monte Carlo method is used to enhance the samples, and the enhanced samples are used to further train the verification model;

[0120] The trained calibration model is used to complete the status calibration of the calibration device online.

[0121] In the embodiment of the present application, in the above step S6, a report is generated according to the online verification result, and manual troubleshooting of the fault includes:

[0122] A calibration report is generated based on the calibration results. The calibration status can be divided into normal, mechanical failure, electrical failure and reliability failure. If the calibration device fails, manual troubleshooting is carried out.

[0123] Specifically, the verification report should include the following contents:

[0124] Verification time and date: This is the key to determine the date of the verification report. A valid verification report should contain the accurate time and date of the verification.

[0125] Calibration Conditions: This section describes the environmental and equipment conditions when the calibration is performed.

[0126] Verification results: For each verification item, detailed verification results should be recorded. If it is numerical data, the specific value should be attached; if it is a qualitative evaluation, a clear conclusion should be given.

[0127] Verification summary: Based on the verification results and relevant standards, make a comprehensive evaluation to determine whether the electricity meter is qualified.

[0128] Suggestions for handling and improvement: If there are abnormalities in the assay transposition, specific suggestions for handling and improvement should be put forward.

[0129] Calibrators and auditors: In order to ensure the authority of the calibration report, dedicated calibrators and auditors should sign and confirm it.

[0130] In another possible implementation, an automated tool and intelligent analysis technology may be used to automatically generate an inspection report and provide more detailed analysis and suggestions.

[0131] Use automated report generation tools (such as Python's Pandas, Matplotlib and other libraries) to read the test result data and automatically generate a structured test report.

[0132] Use data analysis tools (such as Tableau, Power BI, etc.) to visualize and intelligently analyze the inspection results to identify potential failure modes and trends.

[0133] Based on the analysis results, detailed troubleshooting guides and improvement suggestions are automatically generated to help technicians quickly locate and solve problems.

[0134] Intelligent scheduling and dispatching systems can optimize technicians' scheduling and task allocation to ensure that faults can be handled in a timely manner.

[0135] In another possible implementation, an intelligent scheduling system may be established to automatically assign tasks based on the technicians' skills, available time, and urgency of the verification tasks.

[0136] When a fault alarm is received, the system automatically determines the priority of the fault and assigns the task to the most suitable technician.

[0137] Technicians receive tasks through mobile devices and perform troubleshooting and processing according to the system's guidance.

[0138] The above is a schematic scheme of the online calibration method of the automatic calibration device of the electric energy meter of this embodiment. It should be noted that the technical scheme of the online calibration system of the automatic calibration device of the electric energy meter and the technical scheme of the online calibration method of the automatic calibration device of the electric energy meter above belong to the same concept, and the details of the technical scheme of the online calibration system of the automatic calibration device of the electric energy meter in this embodiment that are not described in detail can all be referred to the description of the technical scheme of the online calibration method of the automatic calibration device of the electric energy meter above.

[0139] The online verification system of the electric energy meter automatic verification device in this embodiment includes:

[0140] The first data acquisition module is used to perform online calibration on a predetermined meter position of the automatic calibration device of the calibrated watt-hour meter by relying on the installed intelligent standard watt-hour meter, and obtain the first data by collecting data;

[0141] A second data acquisition module, used for preprocessing the first data to obtain second data;

[0142] A third feature acquisition module, used for performing feature extraction and principal component extraction based on the second data to obtain a third feature;

[0143] A fourth feature acquisition module is used to classify the third feature, mark the sample as normal or fault type, and obtain a fourth feature;

[0144] A verification module, for determining the status of the device by online verification of the automatic verification device of the electric energy meter using the fourth characteristic training model;

[0145] The report processing module is used to generate reports based on the online verification results and manually troubleshoot faults.

[0146] This embodiment further provides a computing device, which is applicable to the case of an online verification method of an automatic verification device for an electric energy meter, and includes:

[0147] Memory and processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement the online calibration method of the electric energy meter automatic calibration device proposed in the above embodiment.

[0148] This embodiment further provides a storage medium on which a computer program is stored. When the program is executed by a processor, the online calibration method of the automatic calibration device of the electric energy meter proposed in the above embodiment is implemented.

[0149] The storage medium proposed in this embodiment and the online calibration method of the electric energy meter automatic calibration device proposed in the above embodiment belong to the same inventive concept. The technical details not fully described in this embodiment can be referred to the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0150] Example 2

[0151] Reference Figure 2-Figure 6 , which is an embodiment of the present invention, provides an online calibration method for an automatic calibration device of an electric energy meter. In order to verify the beneficial effects of the present invention, a scientific demonstration is carried out through simulation experiments.

[0152] It is assumed that an online calibration is carried out on the automatic calibration device of the electric energy meter in a certain power system metering center.

[0153] 1. Data collection:

[0154] like Figure 2 As shown, the calibration device has 2 calibration units, each calibration unit has 36 meter positions, of which 2 meter positions have 2 installed intelligent standard electric energy meters and the meter being tested are calibrated at the same time, and i installed intelligent standard electric energy meters with higher accuracy level than the calibration device are calibrated synchronously with the meter being tested at meter positions 1 and 2. After a batch of electric energy meters are calibrated, the 2 installed intelligent standard electric energy meters continue to calibrate with the next batch of electric energy meters after caching, and a total of n batches of calibration are performed; then enter the second round of detection, the 2 installed intelligent standard electric energy meters move to the right and are fixed at meter positions 3 and 4, and the calibration steps are repeated until the installed intelligent standard electric energy meters complete the data collection of each meter position. Each meter position is collected n times, and the collected data includes the current, voltage, and power in each meter position experiment.

[0155] 2. Data preprocessing: Use data cleaning, normalization and transformation methods to preprocess the raw data for subsequent feature extraction;

[0156] 3. Feature extraction: Based on the test data collected by the installed intelligent standard electricity meter at the meter position, the features can be calculated, which are basic error, phase voltage symmetry, line voltage symmetry, current symmetry, phase-to-phase phase symmetry, line-to-line phase symmetry, A-phase voltage distortion, A-phase current distortion, B-phase voltage distortion, B-phase current distortion, C-phase voltage distortion, C-phase current distortion, device output power stability, device measurement repeatability. For each feature data, the five characteristic values ​​of maximum value (Max), minimum value (Min), average (E), standard deviation (S), and median (M) are calculated.

[0157] 4. Principal component extraction: Principal component analysis (PCA) is used to reduce the dimension of the data, such as Figure 3 As shown in Figure 1, after PCA transformation, the cumulative contribution of the first 56 principal components in the principal component variance contribution rate and the cumulative variance contribution rate has exceeded 85%. Therefore, the first 56 principal components are selected to characterize the characteristics of the original high-dimensional data.

[0158] 5. Sample labeling: Use machine learning methods to classify samples and complete the labeling of four types of samples: normal, mechanical failure, electrical failure, and reliability failure (recorded as states 1 to 4 respectively):

[0159] Figure 4 (a) Figure 4 (b) Box plots of PCA principal component 1 and principal component 2, respectively. Figure 5 (a) to (f) are box plots of original data features 2, 5, 9, 31, 64, and 66, respectively. Their actual meanings are the minimum basic error, the median basic error, the standard deviation of phase voltage symmetry, the maximum distortion of phase A voltage, the standard deviation of device output power stability, and the maximum measurement repeatability of the device. It can be seen that compared with the original feature data, the median and upper and lower quartile distributions of the principal components under different states are significantly different, showing a good classification effect, proving the effectiveness of PCA in extracting important data components.

[0160] 6. Online calibration: First, calculate the basic error, symmetry, distortion, device output power stability, device measurement repeatability, determine the consistency of multi-channel output, variation within the calibration cycle, and daily timing error of the standard clock tester for each meter position according to the provisions of JJG597-2005 "Verification Procedure for AC Electric Energy Meter Verification Device";

[0161] The labeled samples are used to complete the training of the multi-classification support vector machine M-SVM, and the Monte Carlo method is used for data enhancement:

[0162] (1) Calculate the mean and standard deviation of each column of data for each epitope;

[0163] (2) Generate enhanced samples whose corresponding features satisfy the normal distribution according to the calculated mean and standard deviation;

[0164] (3) The enhanced samples and the original samples are combined and merged to finally obtain an enhanced feature data set for further training of the verification model.

[0165] The trained calibration model is used to complete the online calibration of the calibration device status.

[0166] Figure 6 This is the confusion matrix of the test results after data enhancement. Before data enhancement, the prediction results contained 18 state recognition errors with an error rate of 6.25%. After data enhancement, the prediction results contained 5 state recognition errors with an error rate of 0.35%. Therefore, after data enhancement, the model training effect is better and the data classification accuracy is higher.

[0167] 7. Calibration report: Generate a calibration report based on the calibration results; if the calibration device fails, further manual troubleshooting will be performed.

[0168] In the above verification, a total of 120 groups of data status abnormalities were detected, and 6 meter calibration devices were abnormal, among which 23 groups of meter No. 31 and 17 groups of meter No. 34 were identified as mechanical failures, 21 groups of meter No. 11 and 19 groups of meter No. 48 were identified as electrical failures, 19 groups of meter No. 69 and 21 groups of meter No. 54 were identified as stability failures. After manual verification, the mechanical parts of meter No. 31 were rusted, the port crimping of meter No. 34 was deformed, the phase-to-ground short circuit fault of meter No. 11, the poor contact of the connector of meter No. 48, the aging of the sensor of meter No. 69, and the electromagnetic interference noise of meter No. 54 were found.

[0169] In general, this method provides an efficient and accurate online calibration method for the automatic calibration device of electric energy meters, which reduces the company's operation and maintenance costs and improves the reliability and safety of power equipment.

[0170] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. An online calibration method for an automatic calibration device of an electric energy meter, characterized in that: include: Relying on the installed intelligent standard electric energy meter, the predetermined meter position of the automatic calibration device of the electric energy meter to be calibrated is calibrated online, and the first data is obtained by collecting data; Preprocessing the first data to obtain second data; Perform feature extraction and principal component extraction based on the second data to obtain a third feature; Classify the third feature and mark the sample as normal or fault type to obtain the fourth feature; The fourth feature training model is used to determine the device status by online verification of the automatic verification device of the electric energy meter; Generate a report based on the online verification results and manually troubleshoot the faults.

2. The online calibration method of the electric energy meter automatic calibration device according to claim 1, characterized in that: The method of performing online calibration on a predetermined meter position of the automatic calibration device of the calibrated watt-hour meter by relying on the installed intelligent standard watt-hour meter, and obtaining the first data by collecting data includes: Select multiple installed intelligent standard electric energy meters, fix them at the predetermined initial meter positions, and set the number of calibration batches; The installed intelligent standard watt-hour meter is tested together with the watt-hour meter being tested in the watt-hour meter automatic calibration device at the same time, and the calibration is carried out and the data of the installed intelligent standard watt-hour meter automatic calibration device is collected. After each batch is completed, the installed intelligent standard watt-hour meter continues to be calibrated with the next batch.

3. The online calibration method of the electric energy meter automatic calibration device according to claim 2, characterized in that: The method of performing online calibration on a predetermined meter position of the automatic calibration device of the to-be-calibrated electric energy meter by relying on the installed intelligent standard electric energy meter, and obtaining the first data by collecting data also includes: After completing all data collection at the initial position, move the installed standard electric energy meter to the right and fix it at the new position, and repeat the above data collection and verification cycle at the new position until the installed intelligent standard electric energy meter covers all the predetermined meter positions; The data collected from all epitopes are summarized to form the first data.

4. The online calibration method of the electric energy meter automatic calibration device according to claim 3, characterized in that: The preprocessing of the first data to obtain the second data includes: The original data is preprocessed using data cleaning, normalization and transformation methods; the second data is the first data after preprocessing; combined with the second data, the basic error, symmetry, distortion, device output power stability, device measurement repeatability, multi-channel output consistency, variation within the calibration period, and daily timing error of the standard clock tester are calculated for each table.

5. The online calibration method of the electric energy meter automatic calibration device according to claim 4, characterized in that: The third feature is classified, and the sample is marked as normal or fault type, and the fourth feature is obtained including: The machine learning method is used to classify the samples and label the four types of samples: normal, mechanical failure, electrical failure and reliability failure. The fourth feature is the four types of samples after labeling.

6. The online calibration method of the electric energy meter automatic calibration device according to claim 5, characterized in that: The method of using the fourth feature training model to online calibrate the automatic calibration device of the electric energy meter to determine whether the device is qualified and to determine the device status through model analysis includes: Calculate the detection performance index of each epitope; use the fourth feature to train the automatic detection model, use the trained detection model to complete the online detection of the automatic detection device, and judge the status of the detection device based on the output result of the model.

7. The online calibration method of the electric energy meter automatic calibration device according to claim 6, characterized in that: Generating a verification report based on the online verification results and manually troubleshooting the faults includes: A quality report is generated based on the online calibration results. The calibration status is divided into normal, mechanical failure, electrical failure and reliability failure. If the calibration device fails, manual troubleshooting is carried out.

8. A system using the online calibration method of the electric energy meter automatic calibration device according to any one of claims 1 to 7, characterized in that: include: The first data acquisition module is used to perform online calibration on a predetermined meter position of the automatic calibration device of the calibrated watt-hour meter by relying on the installed intelligent standard watt-hour meter, and obtain the first data by collecting data; A second data acquisition module, used for preprocessing the first data to obtain second data; A third feature acquisition module, used for performing feature extraction and principal component extraction based on the second data to obtain a third feature; A fourth feature acquisition module is used to classify the third feature, mark the sample as normal or fault type, and obtain a fourth feature; A verification module, for determining the status of the device by online verification of the automatic verification device of the electric energy meter using the fourth characteristic training model; The report processing module is used to generate reports based on the online verification results and manually troubleshoot faults.

9. A computing device comprising: Memory and processor; The memory is used to store computer executable instructions, and the processor is used to execute the computer executable instructions. When the computer executable instructions are executed by the processor, the steps of the online calibration method of the electric energy meter automatic calibration device according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the online calibration method of an automatic calibration device for an electric energy meter according to any one of claims 1 to 7.

Images